Advances in Experimental Social Psychology, Volume 37 (Advances in Experimental Social Psychology)

ADVANCES IN

Experimental Social Psychology

VOLUME 37

This Page Intentionally Left Blank

ADVANCES IN

Experimental Social Psychology

EDITED BY

MARK P. ZANNA DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF WATERLOO WATERLOO, ONTARIO, CANADA

VOLUME 37

Elsevier Academic Press 525 B Street, Suite 1900, San Diego, California 92101-4495, USA 84 Theobald’s Road, London WC1X 8RR, UK

This book is printed on acid-free paper. Copyright ß 2005, Elsevier Inc. All Rights Reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (www.copyright.com), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-2005 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-2601/2005 $35.00 Permissions may be sought directly from Elsevier’s Science & Technology Right Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, E-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://elsevier.com), by selecting ‘‘Customer Support’’ and then ‘‘Obtaining Permissions.’’ For all information on all Elsevier Academic Press publications visit our Web site at www.books.elsevier.com ISBN: 0-12-015237-1 PRINTED IN THE UNITED STATES OF AMERICA 05 06 07 08 09 8 7 6 5 4 3 2 1

CONTENTS Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Accuracy in Social Perception: Criticisms, Controversies, Criteria, Components, and Cognitive Processes Lee Jussim I. II. III. IV. V. VI.

What Is Social Perceptual Accuracy? .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Why Is Accuracy Research Valuable? . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Why Addressing Controversies in Accuracy Research Is Important . .. . . . . . .. . What This Review Does Not Address. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Structure of This Review . . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . What Are the Main Conceptual Objections to Accuracy Research? A Critical Evaluation of 50 Years of Criticisms. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . VII. How Can One Establish Criteria for Evaluating the Accuracy of Social Perception? . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . VIII. Must All Research on Accuracy Assess Components?. . . .. . . . . . .. . . . . . .. . . . . . .. . IX. Conclusions .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . References . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. .

2 5 9 11 13 14 36 56 75 85

Over Thirty Years Later: A Contemporary Look At Symbolic Racism David O. Sears and P. J. Henry I. II. III. IV. V. VI. VII. VIII. IX. X.

The Theory of Symbolic Racism. . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Controversies About Symbolic Racism. . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . The Symbolic Racism Belief System . .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Consistency of the Political Effects of Symbolic Racism .. . . . . . .. . . . . . .. . . . . . .. . Origins of Symbolic Racism . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Distinctiveness of Symbolic Racism from Alternative Constructs . . . . .. . . . . . .. . Outgroup Antagonism or White Group Consciousness? .. . . . . . .. . . . . . .. . . . . . .. . Summary . . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Further Questions .. . .. . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Conclusions .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . References . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . v

98 101 106 109 118 123 128 132 135 144 146

vi

CONTENTS

Managing Group Behavior: The Interplay Between Procedural Justice, Sense of Self, and Cooperation David De Cremer and Tom R. Tyler I. II. III. IV. V.

Introduction and Overview. .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . The Relation Between Procedural Fairness and Cooperation . . . . . .. . . . . . .. . . .. . Procedural Fairness Models: Motives and Social Self.. . . . . . .. . . . . . .. . . . . . .. . . .. . Overview of Procedural Fairness Research . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . Social Dilemmas and Procedural Fairness: Trust and Goal-Transformation . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . VI. Overview of Cooperation Research . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . VII. The Self and Justice Research . .. . .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . VIII. Summary and Conclusions . . . . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . Appendix I . .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . Appendix II.. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . References. . .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. .

151 154 158 163 174 180 186 193 206 207 207

So Right It’s Wrong: Groupthink and the Ubiquitous Nature of Polarized Group Decision Making Robert S. Baron I. II. III. IV.

The History of a Fine Idea. .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . The Ubiquity of Groupthink. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . Rethinking Groupthink: The Ubiquity Model . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . Strong Versus Moderate Versions of the Ubiquity Model . . .. . . . . . .. . . . . . .. . . .. . References. . .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. .

219 227 238 245 248

An Integrative Theory of Intergroup Contact Rupert Brown and Miles Hewstone I. II. III. IV.

Introduction . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . The Contact Hypothesis and Its Early Modifications .. . .. . . .. . . . . . .. . . . . . .. . . .. . Theoretical Developments of the Contact Hypothesis.. . . . . . .. . . . . . .. . . . . . .. . . .. . Experimental Studies of the Impact of Stereotype-Disconfirming Information: A Cognitive Analysis of Contact.. . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . V. Experimental Studies of Group Salience. . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . VI. Correlational Studies Using Group Salience as a Moderator Variable . . . . . . .. . VII. Mediators of the Effects of Contact. . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. .

256 258 261 267 270 275 284

CONTENTS VIII. Evidence from Educational Settings . .. . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . IX. Acculturation Orientations and Intergroup Attitudes.. . . .. . . . . . .. . . . . . .. . . . . . .. . X. Toward an Integration . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . References . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. .

vii 309 315 317 331

Says Who?: Epistemic Authority Effects in Social Judgment Arie W. Kruglanski, Amiram Raviv, Daniel Bar-Tal, Alona Raviv, Keren Sharvit, Shmuel Ellis, Ruth Bar, Antonio Pierro, and Lucia Mannetti I. II.

Introduction.. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . The Early Work: Hovland et al.’s (1953) Learning Theory Paradigm of Source Effects.. . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Source Effects from the Dual-Mode Perspective: The Elaboration Likelihood Model . . . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Source Effects in the Heuristic Systematic Model.. . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Source Effects in the Unimodel . . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . Empirical Research in the Epistemic Authority Framework . . . .. . . . . . .. . . . . . .. . Summary and Conclusions .. . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . References . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. . . .. . . . . . .. . . . . . .. . . . . . .. .

348 349 350 357 386 389

Index ........................................................................................... Contents of Other Volumes ...............................................................

393 409

III. IV. V. VI. VII.

346 346

This Page Intentionally Left Blank

CONTRIBUTORS Numbers in parentheses indicate the pages on which authors’ contributions begin.

Ruth Bar (345), Department of Psychology, University of Maryland, College Park, MD 20742 Daniel Bar-Tal (345), Department of Psychology, Tel Aviv University, Tel Aviv 69978, ISRAEL Robert S. Baron (219), Department of Psychology, University of Iowa, Iowa City, IA 52317 Rupert Brown (255), Department of Psychology, University of Sussex, Brighton BN1 9QH, UNITED KINGDOM David De Cremer (151), Department of Social Psychology, Tilburg University, 5000 LE Tilburg, THE NETHERLANDS Shmuel Ellis (345), Department of Psychology, Tel Aviv University, Tel Aviv 69978, ISRAEL P. J. Henry (95), Psychology Department, DePaul University, Chicago, IL 60614 Miles Hewstone (255), Department of Experimental Psychology, University of Oxford, Oxford OX1 3UD, UNITED KINGDOM Lee Jussim (1), Department of Psychology, Rutgers University, Piscataway, NJ 08544 Arie W. Kruglanski (345), Department of Psychology, University of Maryland, College Park, MD 20742 Lucia Mannetti (345), Department of Psychology, University of Maryland, College Park, MD 20742 Antonio Pierro (345), Department of Psychology, University of Maryland, College Park, MD 20742 Alona Raviv (345), Department of Psychology, Tel Aviv University, Tel Aviv 69978, ISRAEL

ix

x

CONTRIBUTORS

Amiram Raviv (345), Department of Psychology, Tel Aviv University, Tel Aviv 69978, ISRAEL David O. Sears (95), Psychology Department, University of California Los Angeles, Los Angeles, CA 90095 Keren Sharvit (345), Department of Psychology, Tel Aviv University, Tel Aviv 69978, ISRAEL Tom R. Tyler (151), Department of Psychology, New York University, New York, NY 10003

ACCURACY IN SOCIAL PERCEPTION: CRITICISMS, CONTROVERSIES, CRITERIA, COMPONENTS, AND COGNITIVE PROCESSES Lee Jussim

This article reviews controversies that have surrounded research on interpersonal accuracy for the last 50 years by providing and justifying a probabilistic conception of accuracy and by addressing the following issues: Whether accuracy can be distinguished from logical coherence; whether accuracy research provides insight into cognitive processes; why accuracy in perception of a behavior or trait must be distinguished from accuracy in explanations for that behavior or trait; how to distinguish self-fulfilling prophecy from accuracy; why distinguishing among several levels of analysis is crucial with respect to evaluating the accuracy of social stereotypes; whether accuracy research helps create or alleviate social problems; how accuracy can sometimes lead to discrimination; similarities and diVerences between establishing scientific construct validity and the accuracy of lay social beliefs; and the nature of some of the contributions to understanding social perception provided by componential and noncomponential approaches to the study of accuracy. What could be a more basic or obvious purpose of social perception research than assessment of the accuracy of people’s perceptions of one another? And what could be simpler? Although both questions are phrased rhetorically, in fact, accuracy was an all-but-dead topic within social psychology for roughly 30 years, from about 1955 to 1985. It turned out that not only was the study of accuracy less simple than it seemed but it was, in fact, widely viewed as a theoretical and political minefield. This chapter reviews the major controversies surrounding accuracy research and resolves those controversies, shows how researchers who have been studying accuracy have resolved those controversies, shows how identical issues have been resolved in other areas of psychology, or partially 1 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 37

Copyright 2005, Elsevier Inc. All rights reserved. 0065-2601/05 $35.00

2

LEE JUSSIM

resolves these issues. In the case of partial resolutions, this review also highlights the remaining unresolved issues and discusses how accuracy can be productively studied despite some residual diYculties.

I. What Is Social Perceptual Accuracy? Social perceptual accuracy, in this review, refers to correspondence between perceivers’ beliefs (expectations, perceptions, judgments, etc.) about one or more target people and what those target people are actually like, independent of those same perceivers’ influence on them. This definition may appear straightforward, but it involves subtle issues and complexities that warrant explication—which is provided next.

A. PROBABILISTIC REALISM This definition of accuracy is based on a set of principles referred to here as probabilistic realism. The main ideas of probabilistic realism, adopted throughout this chapter, are that there is an objective reality out there that, flawed and imperfect though we (both scientists and lay people) may be, we can eventually come to know or understand, at least much of the time. The term ‘‘probabilistic realism’’ is used as a somewhat simpler name for what has been called ‘‘critical realism’’ and ‘‘pancritical rationalism.’’ It is essentially the same approach described by Funder (1995, 1999), which itself was heavily influenced by mainstream psychological approaches to construct validity (e.g., Cook & Campbell, 1979; Cronbach & Meehl, 1955). Most psychological researchers implicitly adopt the idea that they can come to understand much about other people nearly all of the time, at least in their research. This perspective is often implicit in the sense that, although its assumptions are widely adopted (as shall soon be demonstrated), few research articles include sections that make explicit the authors’ assumptions regarding the nature of social reality. The next section, therefore, explains both the ‘‘realism’’ and ‘‘probabilistic’’ aspects of probabilistic realism and then briefly describes what accuracy means in this context. 1. Realism ‘‘Realism’’ refers to the idea that there is an objective reality that is independent of social perception. Few social scientists, except the most radical of social constructivists, deny the existence of such a reality. Whenever

ACCURACY IN SOCIAL PERCEPTION

3

empirical researchers measure any psychological variable, they implicitly assume that variable is real. Even the three-decade exile (roughly 1955– 1985) of accuracy research within social psychology occurred because of heightened interest in bias and recognition of genuine complexities in studying accuracy—not because researchers argued there was no social reality. The idea that objective social reality can influence social perception and constrain the potential for bias is explicit in many theoretical perspectives within psychology (e.g., Allport, 1955; Brophy, 1983; Brunswik, 1952; Festinger, 1957; Funder, 1987, 1995, 1999; Gibson, 1979; Jussim, 1991; Kelly, 1955; Kunda, 1990; McArthur & Baron, 1983; McCauley, Stitt, & Segal, 1980). The ‘‘realism’’ part of probabilistic realism reflects this assumption. 2. Social Reality and Social Beliefs Are Often Inherently Probabilistic The term ‘‘probabilistic’’ is used to capture three diVerent aspects of accuracy. First, it means that most criteria are probabilistic, not definitive. This applies to perceptions both of individuals and of groups. For example, a child with a high IQ is likely to do well in school, but there is no guarantee. Similarly, if the Nepalese are more courageous than other people, this does not necessarily mean that every person from Nepal, but no one from Denmark, will rush into the palace to save the Royal Family from a rampaging prince with a gun. People from Nepal do not need to always be braver than other folks, but if the Nepalese are more courageous than other people, they will act more courageously more of the time than do others. Second, ‘‘probabilistic’’ captures the idea that many social beliefs themselves are inherently probabilistic. The belief that Michael Jordan was the best basketball player of the 1990s does not require believing that his teams would win every game they played. Although people rarely phrase their beliefs in explicitly probabilistic terms, this belief is best interpreted as meaning something like ‘‘all other things being equal, having Jordan on your team will enhance your chance of winning more than having any other player on your team.’’ Thus, many social beliefs, themselves, are probabilistic. Stereotypic beliefs, too, are usually inherently probabilistic (see also Krueger, 1996; McCauley et al., 1980). If John believes that Asians are wealthier than other people, this does not necessarily mean that John expects all Asians to be fabulously rich or denies the existence of a single impoverished Asian. It may only mean that, on average, John believes they are richer than other people. Of course, even absolutist beliefs can be viewed as probabilities. The belief that all Englishmen are dignified can be translated into the belief that 100% of Englishmen are dignified. Not only is social reality probabilistic, and not only are many social beliefs probabilistic, but the relationship between belief and reality is often

4

LEE JUSSIM

probabilistic. For example, if Jorge expects Rose to be late for all of their meetings, and Rose is late only 95% of the time, Jorge is still quite accurate— certainly far more accurate than had he expected Rose to generally be on time. Accuracy is rarely all or none; it is usually a matter of degree.

B. ACCURACY Accuracy is greater the greater the correspondence between perceivers’ beliefs (expectations, perceptions, judgments, etc.) about one or more target people and what those target people are actually like, independent of those same perceivers’ influence on those targets. This definition has three core ideas: correspondence, what people are actually like, and independent of influence. If Bella’s teacher, Mr. Jones, predicts that Bella will receive an A on her next math test, and she does, Jones’s belief corresponds well with the outcome. Similarly, near misses involve closer correspondence than wildly inaccurate misses. If Jones predicted an A for Bella, and she receives a Bþ, Jones is still more accurate than had he predicted that she would receive a C. Accuracy as degree of correspondence between subjective perception and objective reality characterizes most modern and many older approaches to accuracy in perception (both object and social) and memory (e.g., Allport, 1958; Brunswik, 1952; Funder, 1995; Gibson, 1979; Judd & Park, 1993; Jussim, 1991; Kenny, 1994; Koriat, Goldsmith, & Pansky, 2000). ‘‘Independent of influence’’ is also conceptually straightforward. It means that the perceiver cannot have caused the target’s outcome. If a coach predicted that one person will become the team’s best player, and the coach caused that person to become the best player (e.g., by giving that player extra time and attention, being extra supportive and encouraging to that player but not to other players, etc.), this is a self-fulfilling prophecy, not accuracy. For a perceiver’s belief to be accurate, it must correspond to targets’ behavior or attributes without the perceiver having caused them. The phrase ‘‘actually like’’ appears simple but implies several important assumptions made explicit here. Probabilistic realism assumes that people have some characteristics that are independent of any particular perceiver’s judgments of those characteristics (see Funder, 1995, 1999, 2001; Jussim, 1991; McCauley et al., 1980). Second, how to identify what those characteristics are, independent of subjective interpersonal judgment, is in essence, the criterion issue. Because identifying criteria for establishing accuracy has so often been framed as something so diYcult as to threaten the viability of accuracy research altogether, it constitutes one of the major issues around which this review is focused and will be addressed in detail later.

ACCURACY IN SOCIAL PERCEPTION

5

C. WHY ACCURACY IS RARELY ABSOLUTE 1. Is a Social Judgment That Is 99% Correct Completely Wrong? To consider accurate only those lay judgments that are absolutely perfect bullseyes would be to render the concept almost useless, lead to an untenable system of classifying beliefs, and reflect a scientifically embarrassing double standard. For example, a prediction that is 99% correct and one that is 0% correct would both have to be considered ‘‘inaccurate.’’ This is useless at best and dysfunctional and misleading at worst. If, for example, one finds a money manager who can successfully predict stock moves 99% of the time, it would be most dysfunctional to dismiss the predictions as inaccurate. Indeed, one might consider emptying one’s bank accounts and mortgaging the house, because this manager can make one rich very quickly. 2. Does a Theory Need to Be Confirmed 100% of the Time to Be Considered Valid? Furthermore, absolute perfection constitutes a standard that scientists, including but not restricted to social psychologists, do not apply in their own work. Consider a psychological theory whose predictions were confirmed 99% of the time. By the absolutist standard for accuracy, the theory would have to be considered ‘‘inaccurate.’’ This is clearly misleading. Far more common is a situation like the following: Results confirm the hypotheses on two out of three dependent variables in study 1, work only among women in study 2, and then reappear on three out of four dependent variables in study 3 for both men and women. Such a study will most likely get published, and the tone will emphasize confirmation of the author’s theory. Thus, this probabilistic conception of social perceptual accuracy applies a similar probabilistic standard when evaluating the accuracy of lay perceptions that scientists apply when testing their own theories and hypotheses and has been widely adopted in theory and research on accuracy in perception, social perception, and memory (e.g., Brunswik, 1952; Funder, 1995; Jussim, 1991; Kenny, 1994; Koriat et al., 2000; McCauley et al., 1980; Ryan, 2002).

II. Why Is Accuracy Research Valuable? But why even bother? Before even attempting to resolve controversies or navigate minefields, it might be worthwhile to explicitly state why some scientists consider research on accuracy to be important.

6

LEE JUSSIM

A. SOME CONTRIBUTIONS OF ACCURACY RESEARCH Many scientists perform accuracy research because they are interested in understanding the basic social and psychological processes underlying interpersonal interactions, close relationships, motivation and achievement, and intergroup relations (e.g., Ambady & Rosenthal, 1992; Funder, 1995; Ickes, 1997; Judd & Park, 1993; Kenny, 1994; Lee, Jussim, & McCauley, 1995). Accuracy research over the last 20 years has addressed issues such as how and why the length of relationships influences social perceptual accuracy; how people identify and integrate behavioral cues to reach judgments about others’ personal characteristics; whether stereotypes typically exaggerate, underestimate, or accurately describe group diVerences; and whether teacher expectations predict student achievement primarily because of accuracy, bias, or self-fulfilling prophecy (and this incomplete list is provided mainly to provide some sense of the issues addressed by accuracy research, not to comprehensively describe the entire body of work—see, e.g., Archer & Akert, 1977; Dawes, 1979; Goldman & Lewis, 1977; Hall & Carter, 1999; Koriat et al., 2000 for accuracy research on other topics).

B. BALANCE? Another reason that research on accuracy is important is to evaluate the validity of the stark, dark impression of human judgment, perception, and thinking that has emerged from the literature on social cognition and decision making. Social psychology has a long history of emphasizing the evil, immoral, and irrational aspects of human thoughts, feelings, and behaviors (see, e.g., Krueger & Funder, 2004, for a review; or the discussions that appear in almost every social psychology text regarding lack of bystander intervention in emergencies, conformity and obedience, the fundamental attribution error, self-serving biases, and judgment under uncertainty).

1. Social Psychological Scholarship Convinces Undergraduates That People Are Fundamentally Irrational When I teach social psychology as a small seminar, I oVer the following question as an essay on the midterm: ‘‘Does social psychological research indicate that people are fundamentally logical and rational or fundamentally illogical and irrational?’’ The required readings up to the midterm include Myers’s (1999) introductory social psychology text’s chapter on the self,

ACCURACY IN SOCIAL PERCEPTION

7

Aronson’s (1999) Social Animal chapter on self-justification, Merton (1948) on self-fulfilling prophecies, La Piere’s (1934) attitudes-do-not-predictbehavior study, and two chapters each from Cialdini’s (1993) book on social influence and Ross and Nisbett’s (1991) book on the person and the situation. Although not a ‘‘random sample,’’ these are, I think, well-respected, common, and mainstream social psychological writings. Consistently, about three quarters of my students conclude that social psychology indicates that people are fundamentally irrational. Consider the following quotations from three papers: First student, introductory sentence: ‘‘Through taking this class, I have come to the conclusion that people are, and have always been, primarily irrational.’’ Second student, introductory sentence: ‘‘People are not rational beings; rather they are rationalizing beings.’’ Third student, concluding sentence: ‘‘I guess that we are probably irrational and spend our lives trying to convince ourselves that we are rational.’’

Intentionally or not, social psychological scholarship often creates the impression that people are fundamentally irrational. 2. Some Prominent Scientists Have Reached the Same Conclusion This impression is not restricted to undergraduates exposed to social psychology for the first time. Consider the following: ‘‘several decades of experimental research in social psychology have been devoted to demonstrating the depths and patterns of inaccuracy in social perception. . . . This applies . . . to most empirical work in social cognition’’ (Jost & Kruglanski, 2002, pp. 172–173). ‘‘Social perception is a process dominated far more by what the judge brings to it than by what he takes in during it’’ (Gage & Cronbach, 1955, p. 420). ‘‘Our beliefs pervasively color and bias our response to subsequent information, evidence, or argumentation’’ (Lord, Lepper, & Preston, 1984, p. 1231). ‘‘Stereotypes are maladaptive forms of categories because their content does not correspond to what is going on in the environment’’ (Bargh & Chartrand, 1999, p. 467).

And, of course, several highly influential books have been written on error and bias, which, except for an occasional passing acknowledgment that people are not always wrong, primarily focus on how judgment and perception go astray (Gilovich, 1991; Kahneman, Slovic, & Tversky, 1982; Nisbett & Ross, 1980).

8

LEE JUSSIM

3. Is the Glass Really Almost Empty? Evaluating the Validity of Perspectives Emphasizing Human Bias and Irrationality In this context, then, research on accuracy becomes important for either of two reasons. One reason is that many social psychologists may emphasize error and bias because they believe that social judgment really is overwhelmed by foolishness and irrationalities. Those subscribing to this view may not completely deny the possibility of accuracy or rationality, but the quotations above indicate that some highly influential social psychologists seem to conclude that inaccuracy, bias, and irrationality are the norm. To the extent that research within this perspective is interpreted broadly as ‘‘demonstrating the depths and patterns of inaccuracy in social perception’’ (Jost & Kruglanski, 2002), then work on accuracy is essential for evaluating its validity. To reach broad and general conclusions about inaccuracy, one cannot simply study error and bias. If studies are designed just to test for bias, then they will either find bias (if successful) or nothing (if unsuccessful—see Krueger & Funder, 2004, for a more detailed analysis of this point). Either way, many such studies provide little or no information regarding accuracy. The only theoretically viable way to reach broad and generalizable conclusions about inaccuracy is to conduct accuracy research. 4. Is the Glass Really Only Half-Empty? Evaluating the Validity of More Balanced Perspectives Another, not mutually exclusive, possibility is that some social psychological perspectives view people as often accurate and rational but, for a variety of reasons, consider error, bias, and irrationality more interesting and important. This perspective does not mean to create the impression that people are fundamentally irrational and overwhelmingly inaccurate. Proponents would probably agree that a more balanced impression of human social judgment and perception is warranted and desirable. Koriat et al. (2000, p. 522) reached a similar conclusion regarding research on the accuracy of memory: ‘‘Second, while memory may in fact be more fallible and malleable than is assumed by the layman, it seems to us that the interest in memory illusions and false memories, spurred perhaps by real life problems, has led researchers to selectively focus on the dark side of memory, resulting in a somewhat biased picture.’’

Replace ‘‘memory’’ with ‘‘social perception and judgment’’ and this quotation might also aptly describe the current status of research on social cognition. For social scientists subscribing to a more balanced view, work

ACCURACY IN SOCIAL PERCEPTION

9

focusing on accuracy is important to evaluate its validity. A balanced perspective would be largely justified if research typically found people arriving at biased and erroneous judgments to about the same extent that they reach unbiased and accurate judgments. However, such a perspective could also be unjustified in either of two opposing directions. First, perhaps, in general, people’s social beliefs, perceptions, and judgments are indeed dramatically more prone to error and bias than accuracy (consistent with the first perspective); or second, perhaps, in general, people’s beliefs, perceptions, and judgments are much more accurate than they are erroneous or biased. As Brunswik (1952) pointed out over 50 years ago, the only way to justifiably reach broad and general conclusions regarding the viability of the balanced perspective or either of the imbalanced perspectives is to conduct a great deal of research using a wide variety of methodologies that assesses the accuracy of people’s social beliefs, perceptions, expectations, and judgments in a wide variety of contexts—and only then trying to reach broad and general conclusions about the typical accuracy or inaccuracy of social judgments (see also Funder, 1987, 1995; Koriat et al., 2000 for similar points).

III. Why Addressing Controversies in Accuracy Research Is Important The need to resolve theoretical and conceptual issues in the study of social perceptual accuracy is particularly acute. This is because, as shall be demonstrated throughout this chapter, many of the criticisms of accuracy research and even of the concept of social perceptual accuracy have typically not been framed in terms of the types of alternative perspectives and debates that have framed other areas of social psychology (e.g., do attitudes strongly or weakly predict behavior? Is self-enhancement or self-consistency the dominant selfmotive? Does aVect precede cognition, or does cognition precede aVect? Do individuating information or stereotypes dominate person perception?). Instead, many of the criticisms of accuracy research have often been presented in such a manner as to imply or explicitly state that accuracy research is so flawed as to render the entire enterprise of dubious value, or at least not worth the eVort. As a consequence, although theoretical controversies in any area often provide a rich springboard for inspiring research, in the area of accuracy, such controversies have historically tended to restrict and reduce such research by stigmatizing it. Thus, addressing and resolving many of those controversies would seem to be a particularly valuable conceptual and theoretical undertaking. This review, however, does not merely identify those criticisms, diYculties, or

10

LEE JUSSIM

obstacles—it also critically evaluates them. As shall be seen, some (but not all) criticisms are more flawed than that which they criticize, and some (but not all) obstacles are more apparent than real. Some, however, constitute bona fide diYculties in conducting accuracy research. One purpose of this review, therefore, is to discriminate between, on one hand, the unjustified criticisms and pseudo-obstacles, and on the other, justified criticisms and genuine obstacles. When the criticisms are justified and the obstacles real, this review will discuss how they can be overcome. To accomplish this goal, the remainder of this review centers around three major questions, each of which chronicles a set of obstacles to, diYculties in, or criticisms regarding accuracy research. First, what are common theoretical, conceptual, and political criticisms of accuracy research? How valid are these criticisms? When valid, do they threaten the viability of accuracy research and, if so, how can such threats be addressed? Second, is it possible to establish reasonable criteria against which to evaluate the accuracy of social perception? One of the most common criticisms of accuracy research is the claim that it is diYcult to find appropriate criteria with which to assess the validity of social perception. In this review, this claim, which has often been presented as a semirhetorical indictment of accuracy research, is transformed into a scientific question: What criteria can be used to assess the validity of social perceptions and what are their limitations? Third, must all research on accuracy perform a complex analysis of the components of social judgment? Another common criticism of accuracy research is that measures of social perceptions are heavily polluted by artifacts and biases. One set of proposed solutions to this problem involves diYcult and complex statistical procedures to remove or account for those artifacts and biases. Unfortunately, however, these solutions have a long history of discouraging any accuracy research at all. Thus, this review summarizes existing componential approaches and addresses whether accuracy can be meaningfully assessed without using such procedures. These questions are substantively diVerent from one another: some are conceptual or meta-theoretical, others are methodological, and the last set is mathematical/statistical. Nonetheless, they are all similar in that, when they have been addressed in many prior reviews, the answers to them have frequently suggested that the obstacles to accuracy research are so formidable as to render such research of dubious value or not worth the eVort. Thus, one contribution of this review is to critically evaluate the criticisms and objections to accuracy research that have been raised over the last 50 years. In the process of addressing these questions, several substantive theoretical issues are also discussed: Whether accuracy can be distinguished from

ACCURACY IN SOCIAL PERCEPTION

11

logical coherence, whether accuracy research provides insight into cognitive processes, why accuracy in perception of a behavior or trait must be distinguished from accuracy in explanations for that behavior or trait, how to distinguish self-fulfilling prophecy from accuracy, why distinguishing among several levels of analysis is crucial with respect to evaluating the accuracy of social stereotypes, whether accuracy research helps create or alleviate social problems, how accuracy can sometimes lead to discrimination, similarities and diVerences between establishing scientific construct validity and the accuracy of lay social beliefs, and the nature of some of the contributions to understanding social perception provided by componential and non-componential approaches to the study of accuracy.

IV. What This Review Does Not Address A. CORRESPONDENCE VERSUS COHERENCE Accuracy, in this review, involves correspondence between belief and reality, not logical coherence of belief. Logical coherence refers to the internal logical consistency of people’s beliefs and perceptions and is best represented by the vast amount of research on logical flaws and errors in decision making and social judgment (e.g., Gilovich, 1991; Hastie & Dawes, 2001; Kahneman et al., 1982; Nisbett & Ross, 1980). Consider someone who estimates that there is a higher chance of 1000 people dying in California because an earthquake leads to a huge flood than there is a chance of 1000 people dying in the United States because of a flood (Tversky & Kahneman, 1983). These predictions are logically incoherent, because floods in California are a mere subset of all floods in the United States, and floods caused by earthquakes are a subset of all floods. Therefore, it is impossible for a flood caused by an earthquake that kills 1000 people in California to be more probable than a flood that kills 1000 people somewhere in the United States. (In)coherence is, at best, however, only weakly related to accuracy defined as correspondence [Wright & Drinkwater (1997) present a detailed exposition regarding an empirical example of the weak relationship between coherence and accuracy]. Let’s say that the real probabilities of an earthquake causing a flood in California that kills 1000 people sometime next year is 1/10,000 and the probability of a flood killing 1000 people somewhere in the United States next year is 1/1000. Now let’s consider two perceivers generating intuitive probabilities: Social Judge CBI (Coherent but Inaccurate) and Social Judge IBMA (Incoherent but Mostly Accurate). Social Judge CBI

12

LEE JUSSIM

believes the real probabilities are 70% and 90%. CBI is logically coherent, but his estimates are wildly inaccurate. Social Judge IBMA, in contrast, believes that the probabilities are 1/1000 (California) and 1/10,000 (U.S.), respectively. IBMA’s beliefs are logically incoherent, and her incoherence guarantees that her beliefs will not be perfectly accurate, because at least one of two, and perhaps both, incoherent beliefs must be inaccurate. Nonetheless, from the standpoint of accuracy as defined here (correspondence, independent of influence), IBMA’s predictions are far closer to the truth—that is, far more accurate—than are CBI’s. Logical flaws, errors, and biases in the processes of judgment and decision making are extremely important, and nothing in this review is intended to, or should be interpreted as, challenging, undermining, or underestimating them. Such flaws in logic or process, however, do not usually provide much direct information about accuracy, as defined here. One can be coherent and accurate, coherent and inaccurate, incoherent and (largely) accurate, and incoherent and inaccurate. This review focuses exclusively on accuracy and does not address coherence further.

B. MAJOR EMPIRICAL ACCOMPLISHMENTS OF MODERN WORK ON ACCURACY This chapter does not review the major empirical accomplishments of accuracy research. Reviews of, for example, accuracy in personality perception (Funder, 1987, 1995, 1999), accuracy in dyadic interactions (Kenny, 1994; Kenny & Albright, 1987), empathic accuracy (Ickes, 1993, 1997), accuracy based on minimal information (Ambady & Rosenthal, 1992), memory (Koriat et al., 2000), and accuracy in social stereotypes (Eagly & Diekman, 1997; Judd & Park, 1993; Lee et al., 1995; McCauley et al., 1980; Ryan, 1995, 2002) can all be found elsewhere. The purpose of this chapter is to describe how the myriad of conceptual and methodological criticisms and problems with, and obstacles to, conducting accuracy research can be resolved or have already been resolved by the existing scholarship regarding accuracy.

C. THE PHENOMENOLOGY OF ACCURACY This chapter is not about the phenomenology of accuracy. Whether lay people do or do not see accuracy in the same manner as described here is an interesting issue but is beyond the scope of this review. The definition of

ACCURACY IN SOCIAL PERCEPTION

13

accuracy here—probabilistic correspondence between belief and reality, independent of causal influence—is a scientific definition, not a description of or hypothesis about how lay people construe accuracy. D. MORALITY AND LEGALITY Universities cannot reject the application of a particular African American simply because one believes that, on average, African Americans score about one standard deviation lower than do Whites on their Standard Aptitude Tests (SATs). Such a rejection would clearly be immoral and illegal because the average characteristics of a group do not dictate the specific characteristics of any individual member. Nonetheless, the belief that African Americans, on average, score about 100 points lower on each of their SATs is accurate (e.g., Neisser et al., 1996). The claim about this particular applicant’s qualification, however, may be—but is not necessarily—wrong. Whether the claim that this particular applicant has low SAT scores is accurate depends on the applicant’s actual SAT scores. Although accuracy is sometimes intertwined with moral and legal issues, establishing accuracy is clearly a diVerent endeavor than establishing morality or legality. Sometimes, being accurate may be widely seen as oVensive and immoral, especially if one is accurate when referring to others with intellectual, physical, or social limitations. This issue is particularly acute with respect to perceptions of groups. It is usually oVensive to claim that ‘‘they,’’ as a group, have less of some socially desirable attribute or engage less in some socially desirable behavior than do other groups. This chapter primarily addresses issues of accuracy, without regard to whether accuracy might be considered oVensive or immoral by some people or by the legal system. Moral and legal issues are extremely important and will be touched on in a later section addressing political objections to accuracy research and the complex interrelationships of accuracy, bias, and discrimination. Legal and moral issues are, however, generally beyond the scope of this chapter (see, e.g., Dawes, 1988; and Gottfredson, 1994, for discussions of situations in which U.S. courts declared it illegal for employers to use the most valid information available for making decisions about hiring new employees).

V. Structure of This Review The subsequent critical evaluation of objections to accuracy research is divided into three major sections: a review of the major conceptual objections to accuracy research, an evaluation of the methodological criticism

14

LEE JUSSIM

that it is inordinately diYcult to identify criteria for assessing the accuracy of social judgment, and an evaluation of the claim that it is only possible to assess accuracy with statistically sophisticated and diYcult componential analyses.

VI. What Are the Main Conceptual Objections to Accuracy Research? A Critical Evaluation of 50 Years of Criticisms This section critically evaluates some of the most common substantive or theoretical objections to accuracy research (methodological and statistical objections, such as criteria and components, are considered in later sections). This includes claims that research on accuracy is less informative than research on error and bias because accuracy research provides little information about psychological processes; research on accuracy is unnecessary because research on social cognition already shows that social perception is generally inaccurate; research on accuracy is uninformative because it fails to address how perceivers explain others’ behaviors and attributes; accuracy, itself, may not be very meaningful because that which is ‘‘accurately’’ perceived may merely reflect prior self-fulfilling prophecies; and stereotype accuracy research is unnecessary because even a stereotype that fits a group will inaccurately describe most members of that group. It also addresses a political objection to accuracy research: that accuracy research, and especially research on the accuracy of social stereotypes, is immoral because it does little more than justify or reify existing social inequalities.

A. ‘‘COGNITIVE PROCESSES ARE IMPORTANT, ERROR AND BIAS IS IMPORTANT, BUT ACCURACY IS NOT’’ 1. Process Researchers studying cognitive processes have frequently ignored accuracy and have periodically explicitly repudiated it (see, e.g., Fiske & Taylor, 1991; Jones, 1985, 1986; Stangor, 1995). One reason may be that, in its most simplistic form, accuracy research seems to provide no information about process. Knowing only that a perceiver accurately judged some person or group does not tell us very much about that perceiver’s thought processes. This simplistic view of accuracy, however, does not characterize most accuracy research, nearly all of which does examine issues of social and cognitive processes. Brunswik’s (1952) Lens Model, Cronbach’s componential model,

ACCURACY IN SOCIAL PERCEPTION

15

Funder’s (1995, 1999) Realistic Accuracy Model, Judd and Park’s (1993) componential model for studying stereotype accuracy, Ickes (1997) work on empathic accuracy, Jussim’s (1991) Reflection-Construction Model, and Kenny’s (1994) Social Relations Model all address the processes by which people arrive at accurate or inaccurate judgments regarding others (several of these will be discussed in detail later in this chapter). Thus, the idea that research on accuracy is unimportant because it does not address issues of process is false because process issues are central to accuracy research. 2. Error and Bias Error and bias may be considered more important than accuracy by some researchers for several reasons. One may be that errors and biases are widely viewed as revealing psychological processes. This is true (see, e.g., Funder, 1987; Kahneman & Tversky, 1973), but this then becomes the ‘‘process’’ reason for not studying accuracy, which, it has just been argued, is flawed because accuracy research also addresses process. Another possibility is that bad and foolish behavior is inherently attention grabbing, and rational and reasonable behavior rarely is (Aronson, 1999; Krueger & Funder, 2004; Schneider, Hastorf, & Ellsworth, 1979). It is frequently surprising and dismaying to discover that we make systematic errors in certain types of situations. Identifying such errors is a potentially important first step toward correcting them and arriving at better judgments. These reasons for emphasizing error and bias have validity. None, however, justify ignoring or dismissing accuracy research or the evidence of accuracy often available in studies of error and bias. Bad and foolish behavior may be attention grabbing, but presumably, scientists evaluate a phenomenon’s importance by factors such as its frequency, power, and place within some wider theoretical context. For example, scientists typically consider car driving more dangerous than airplane flying, because far more people die per mile traveled in cars than in airplanes. That airplane crashes are far more attention grabbing is seen as an interesting social psychological phenomenon but not as a basis for emphasizing the dangers of flying. Similarly, the greater attention-grabbing power of error and bias does not constitute a scientific basis for considering them more important than accuracy. Furthermore, it is only possible to correct errors if we know what the correct judgment should be and we have means of assessing whether future judgments are correct. Accuracy research is essential if one is interested in correcting systematic errors. Therefore, an interest in improving human judgment does not constitute a basis for considering error and bias research more valuable than accuracy research. Overall, therefore, although these

16

LEE JUSSIM

reasons for emphasizing error and bias have validity, they do not justify deemphasizing accuracy.

B. RESEARCH ON ACCURACY IS NOT NECESSARY BECAUSE THE SOCIAL COGNITION LITERATURE ALREADY SHOWS THAT SOCIAL PERCEPTION AND JUDGMENT ARE DOMINATED BY ERROR AND BIAS Psychological research articles are filled with excellent experimental studies of cognitive processes that researchers interpret as indicating that bias, error, and self-fulfilling prophecy are likely to be common in daily life (e.g., Chen & Bargh, 1997; Fiske & Neuberg, 1990; Fiske & Taylor, 1991; Gilbert, 1995; Jost & Kruglanski, 2002; Kahneman et al., 1982; Nisbett & Ross, 1980; Stangor & McMillan, 1992). Therefore, another reason for dismissing or avoiding accuracy research may be that it is not be seen as necessary. If errors and biases are identified, then, ipso facto, do we not know that people are inaccurate? In fact, however, we do not know this. Conclusions regarding the power and prevalence of error and bias in daily life are only justifiable by research that examines the accuracy of people’s judgments, preferably under realistic conditions. No matter how much researchers believe that the processes discovered in the lab should lead to bias and error in daily life, the only way to find out is by assessing the accuracy of those social perceptions. This occurs for at least two separable reasons: some biases can coexist with some degree of accuracy, and some biases enhance accuracy. For example, consider the ‘‘false’’ consensus eVect—the phenomenon whereby people overestimate the extent to which others agree with them. A simple example shows how this is not necessarily mutually exclusive with accuracy. Consider a simple dichotomous choice, say between candidates from two political parties. Assume that 75% of the people in this district belong to one party and 25% to the other party, that all people prefer the candidate from their own party and the existence of an extreme false consensus eVect in this district—all people believe that their candidate receives the support of the majority. Of the people holding this belief, 75% will be correct (see also, Dawes, 1989). Similar analyses showing that bias can either coexist with or enhance accuracy have been applied to expectancy eVects, social stereotypes, and interpersonal perceptions (Brodt & Ross, 1998; Jussim, 1991; Kenny, 1994; Kenny & Acitelli, 2001; McCauley et al., 1980; Ryan, 1995, 2002). For example, stereotypes sometimes influences person perception judgments—a

ACCURACY IN SOCIAL PERCEPTION

17

pattern routinely interpreted as evidence of bias and source of discrimination (e.g., Aronson, 1999; Bodenhausen, 1988; Borgida, Rudman, & Manteufel, 1995; Darley & Gross, 1983; Fiske & Taylor, 1991; Fiske & Neuberg, 1990; Jones, 1986, 1990; Stangor, 1995). If the following two conditions are met, however, both Bayes’s theorem and regression principles predict that such a bias will enhance the accuracy of those person perception judgments: individuating information is less than perfectly diagnostic, and the stereotype is itself an accurate characterization of the group (Jussim, 1991). Such predictions were later confirmed by research empirically demonstrating that the college students who relied most heavily on stereotypes regarding the types of students who choose to live in diVerent residence halls (hippies, preppies, etc.) also reached the most accurate judgments regarding particular individual residents (Brodt & Ross, 1998). As a consequence, even if experimental laboratory demonstrations of error and bias readily generalized to daily life judgments, perceptions, and decisions—which is not always clear—they do not inexorably lead to the conclusion that those judgments, perceptions, or decisions are inaccurate. If one wishes to make general claims about the extent to which social perception and judgment are dominated by inaccuracies, one cannot rely simply on error and bias research; one needs to assess the accuracy of those judgments.

C. ACCURACY OF EXPLANATIONS: ‘‘JUST BECAUSE SOME BELIEF ABOUT SOME PERSON OR GROUP IS CORRECT DOES NOT TELL US WHY OR HOW THE PERSON OR GROUP GOT THAT WAY’’ One common criticism of accuracy research is that it fails to address explanations—either scientific ones or phenomenological ones—for how individuals or groups developed their characteristics. This criticism of accuracy research is implicit in perspectives indicating that accuracy cannot be studied or is meaningless because social processes and phenomena (e.g., discrimination, poverty) create the diVerences that are perceived (e.g., Claire & Fiske, 1998). It is similarly implicit in the common claim that stereotypes are flawed because they assume group diVerences are biological when they are not (see, e.g., critical reviews of the stereotype concept appearing in Ashmore & Del Boca, 1981; Brigham, 1971; Jussim, McCauley, & Lee, 1995; McCauley et al., 1980). Such perspectives indicate that demonstrating that people accurately perceive group diVerences is meaningless because it provides no information about either how people explain those diVerences or how those diVerences arose.

18

LEE JUSSIM

The claim that ‘‘demonstrating accuracy does not explain how or why those being accurately perceived got that way’’ is absolutely true; however, it also fails to threaten or undermine the viability, importance, or informativeness of accuracy research. This applies to most existing research on accuracy. For example, friends have been found to be more empathically accurate than strangers (Stinson & Ickes, 1992). Empathic accuracy refers to inferring one’s partner’s conscious thoughts and feelings. Consider a Perceiver who believes a Target is feeling hostile. This ‘‘objection’’ focusing on the accuracy of explanations leads to at least four diVerent questions: Is the Perceiver right? What is the Perceiver’s explanation for the Target’s hostility? If the Target is hostile, how did he or she get that way? And why does the Perceiver believe the Target is hostile? Providing an answer to one question provides no information about the others. For example, establishing that the Perceiver is correct (the Target really is feeling hostile) tells us nothing about how the Perceiver explains the Target’s hostility. Maybe the Perceiver is a bigot who thinks that the Target’s ethnicity makes him or her prone to hostility. Maybe the Perceiver thinks the Target was mistreated as a child. Maybe the Perceiver thinks that the Target watches too many old Clint Eastwood movies. Similarly, establishing that the Perceiver is correct tells us nothing about how the Target became hostile. Maybe there are genes for hostility, and the Target has them. Maybe he or she was abused as a child. Maybe the Target has been watching too many old Clint Eastwood movies. One could attempt to establish the validity of the Perceiver’s explanation for the Target’s hostility by comparing it to the ‘‘true’’ reasons for the Target’s hostility, if they could be uncovered. Doing so would probably be a diYcult task, but whole bodies of research have addressed sources of hostile and aggressive behavior (e.g., virtually every social psychological textbook has an entire chapter devoted to explaining aggression), so it would not be impossible. Assessing the validity of the Perceiver’s belief that the Target is hostile is simply a diVerent endeavor than is assessing the validity of the Perceiver’s explanation for the Target’s hostility. That a particular study only focuses on assessing one type of accuracy does not fatally flaw such research—it only means that although considerable information may be provided regarding one type of accuracy (e.g., accuracy in perception of a feeling or trait), no information may be provided about another type of accuracy (e.g., accuracy in the explanation for that feeling or trait). Furthermore, none of this necessarily explains how or why the Perceiver came to believe that the Target is hostile. This is a social and cognitive process question, and process is important, but it is not an accuracy question. The accuracy issue evaporates here, because we are no longer evaluating the

ACCURACY IN SOCIAL PERCEPTION

19

validity of the Perceiver’s judgments, expectations, or beliefs. With respect to evaluating the validity of the Perceiver’s belief that the Target was feeling hostile, it does not matter how the Perceiver explains the Target’s hostility, it does not matter how the Target came to feel hostile, and it does not matter how the Perceiver came to believe that the Target felt hostile. This analysis is equally applicable to evaluating the accuracy of people’s beliefs about groups (stereotypes). No information about how people arrived at their beliefs about group diVerences, how they explain those diVerences, or how those diVerences actually emerged is provided by the now-abundant research showing that people’s beliefs in some diVerences between two groups are often partially or largely accurate (e.g., Brodt & Ross, 1998; Diekman, Eagly, & Kulesa, 2002; Hall & Carter, 1999; Judd, Park, Ryan, Brauer, & Kraus, 1995; Jussim, Eccles, & Madon, 1996; Madon, Jussim, Keiper, Eccles, Smith, & Polumbo, 1998; McCauley & Stitt, 1978; McCauley & Thangavelu, 1991; Ryan, 1996, 2002; Ryan & Bogart, 2001; Swim, 1994). The lack of information about explanations constitutes a limitation to these studies. This limitation, however, does not threaten or undermine what they do show—considerable accuracy in people’s perceptions of diVerences between demographic groups or between individuals from diVerent groups.

D. ACCURACY VS. SELF-FULFILLING PROPHECY: ‘‘IT IS NOT MEANINGFUL TO DISCUSS ‘ACCURACY’ IF WHAT IS BEING ‘ACCURATELY PERCEIVED’ DOES LITTLE MORE THAN REFLECT SELF-FULFILLING PROPHECIES’’ This objection specifies a very particular process by which those being perceived accurately became that way—self-fulfilling prophecies. This review gives it separate consideration because numerous researchers have specifically stated or can be read as implying that accuracy is meaningless because that which is accurately perceived could result from self-fulfilling prophecies; many perspectives on expectancies enthusiastically embrace self-fulfilling prophecies but question accuracy, in part on the grounds that accuracy is confounded with self-fulfilling prophecy; the main type of ‘‘accuracy’’ some theoretical perspectives discuss is the specious form that results from selffulfilling prophecies; and both accuracy and self-fulfilling prophecy involve a belief or expectation corresponding well with targets’ outcomes so that the potential confounding of the two is particularly salient or obvious (Claire & Fiske, 1998; Jones, 1986, 1990; Jost & Banaji, 1994; Olson, Roese, & Zanna, 1996; Snyder, 1984; Snyder & Stukas, 1998; Swann, 1984).

20

LEE JUSSIM

The logic underlying this objection seems to be the following: 1. Self-fulfilling prophecies occur. 2. Therefore, diVerences between targets may reflect eVects of self-fulfilling prophecies. 3. If diVerences that are perceived reflect self-fulfilling prophecies to some unknown degree, attributing ‘‘accuracy’’ to those perceptions is, at best, unjustified, and at worst, reifies diVerences produced through social processes.

There is some truth to this argument. DiVerences that are accurately perceived at some point in time may reflect eVects of prior self-fulfilling prophecies. Furthermore, the confounding of self-fulfilling prophecy and accuracy clearly would be a problem in any situation (e.g., daily life, lab research) in which it was not possible to distinguish these two very diVerent reasons for why a perceiver’s expectations might be confirmed. Simply showing that a perceiver’s belief corresponds well with targets’ actual attributes or behaviors, by itself, cannot distinguish accuracy from self-fulfilling prophecy. Researchers have, however, done far more than simply demonstrate correspondence between perceiver beliefs and targets’ attributes and behaviors. 1. A Wide Array of Methodological and Statistical Techniques Exists for Distinguishing Accuracy from Self-Fulfilling Prophecy One technique is to have people judge targets with whom they do not interact (e.g., by judging them from resumes, college records, photographs, etc.). People cannot create self-fulfilling prophecies among targets with whom they do not interact. Therefore, by ruling out self-fulfilling prophecy, such designs allow for an assessment of the accuracy of person perception judgments (for examples, see Archer & Akert, 1977; Brodt & Ross, 1998; Goldman & Lewis, 1977). Other methods allow for the simultaneous assessment of accuracy and self-fulfilling prophecy. Although a detailed discussion of these methods is beyond the scope of this chapter, the core idea is simple: If perceivers’ expectations are self-fulfilling, they should predict changes in targets’ behavior or accomplishments over time. Theoretical models relying on structural equation techniques have been developed for distinguishing self-fulfilling prophecy from accuracy under naturalistic conditions (e.g., Jussim, 1991; Madon, Guyll, Spoth, Cross, & Hilbert, 2003; Trouilloud, Sarrazin, Martinek, & Guillet, 2002; West & Anderson, 1976; Williams, 1976). Therefore, criticisms along the lines of ‘‘if accuracy is confounded with self-fulfilling prophecy, the meaning of accuracy is clouded’’ are true,

ACCURACY IN SOCIAL PERCEPTION

21

but only in the narrow literal sense of this conditional statement (i.e., if confounded, then clouded meaning). Numerous approaches have been developed that largely or completely unconfound the two. 2. ‘‘Prior Self-Fulfilling Prophecies May Influence That Which is ‘Accurately’ Perceived’’ The self-fulfilling prophecy problem, however, does not go away so easily. Even if any particular study can rule out self-fulfilling prophecy in that particular study, the diVerences that are ‘‘accurately perceived’’ may still have resulted from prior self-fulfilling prophecies that occurred in interactions outside the study (either with other perceivers or with the same perceiver outside the study). Much like the prior criticisms of accuracy research, however, although this claim is valid (prior self-fulfilling prophecies may indeed influence that which is accurately perceived), it does not undermine the value, viability, or interpretation of accuracy. The next sections explain why. 3. The Unconfounding of Impressions and Predictions. I. The Perceiver Is Accurate Even if Self-Fulfilling Prophecies Resulting from Other Perceivers’ Expectations Did Create Target DiVerences Isn’t it vacuous at best and misleading at worst to attribute ‘‘accuracy’’ to a belief that is true only because of prior self-fulfilling prophecies? When this rhetorical question is treated scientifically, rather than rhetorically, the answer is a clear ‘‘no.’’ Understanding why requires understanding the diVerence between impressions and predictions. If target behavior, accomplishments, and so forth predate perceiver beliefs about the target, causality can only flow in one direction: from target behavior to perceiver beliefs. Those perceiver beliefs may indeed become self-fulfilling, but only with respect to future target behaviors. In this chapter perceiver beliefs developed on the basis of prior target behaviors, accomplishments, and so on are referred to as ‘‘impressions’’; and perceiver beliefs that might predict future target behaviors are referred to as ‘‘predictions.’’ The importance of this distinction is illustrated with a composite, concrete example based on the type of data in numerous naturalistic studies of teacher expectations (e.g., Hoge & Butcher, 1984; Jussim et al., 1996; West & Anderson, 1976; Williams, 1976). Consider Miss Smith, a sixth-grade teacher, beginning a new school year. In looking through the records of her new students, she discovers Donna, who received As last year and scored in the top 10% on a statewide standardized achievement test, and Mary, who received Cs and Ds and scored in

22

LEE JUSSIM

the bottom 30% of the same test. Miss Smith concludes that Donna has been a better student, a higher achiever, and knows the material better. Is there anything wrong, inappropriate, unjustified, confusing, confounding, uninterpretable, or vacuous about Miss Smith’s evaluations? Is Miss Smith ‘‘blaming the victim?’’ Is she reifying diVerences between Donna’s and Mary’s ethnic groups? These are rhetorical questions because it is obvious that Miss Smith’s judgment is accurate, even if the diVerence between Donna and Mary resulted, in part, from prior self-fulfilling prophecies. Perhaps Donna’s parents strongly encouraged her participation in intellectual activities, whereas Mary’s parents did not. Perhaps Donna had a great teacher in fifth grade who inspired in her a commitment to educational achievement, and perhaps Mary had a teacher who was insulting and obnoxious and who discouraged her. These and other potential self-fulfilling prophecy explanations do not change the fact that here and now, Donna has a much better record than does Mary. Believing anything else would be starkly incorrect. 4. The Unconfounding of Impressions and Predictions. II. The Perceiver’s Impressions Can Be Accurate Even if Self-Fulfilling Prophecies Resulting from the Same Perceiver’s Expectations Did Create Target DiVerences Again, the key issue here is time. If a swim instructor’s expectations trigger a learning interaction sequence such that those expectations cause the student (Jacques) to become an unusually fast swimmer, those expectations are self-fulfilling (e.g., Trouilloud et al., 2002). But, once Jacques has completed the race in record time, how should the swim instructor perceive him? Would the swim instructor be most accurate if she perceived Jacques as a slow, awkward, incompetent swimmer? Or even as an average swimmer? Again, the answer is obvious. A ‘‘problem’’ arises only when we fail to distinguish between impressions and predictions (keeping in mind that today’s impression can become tomorrow’s prediction). 5. Self-Fulfilling Prophecies Occur but Do Not Invalidate Accuracy Although self-fulfilling prophecies sometimes occur, one cannot assume that diVerences between any two people or groups result from self-fulfilling prophecies. Such a claim requires specific empirical justification (i.e., empirical evidence that self-fulfilling prophecies caused the diVerences among the particular targets being studied). Citation of a handful of dramatic self-fulfilling prophecy studies (e.g., Rosenthal & Jacobson, 1968; Snyder, Tanke, & Berscheid, 1977; Word, Zanna, & Cooper, 1974) does not

ACCURACY IN SOCIAL PERCEPTION

23

constitute adequate justification for a new claim that self-fulfilling prophecies caused diVerences among a new set of targets because so much other research shows that expectancy eVects are far from inevitable, powerful, or permanent (see, e.g., meta-analyses and reviews by Brophy, 1983; Jussim, 1991; McNatt, 2000; Raudenbush, 1984; Rosenthal & Rubin, 1978). Thus, the claim that accuracy cannot be studied because prior self-fulfilling prophecies might influence that which is accurately perceived includes a core falsehood (‘‘accuracy cannot be studied’’) enveloped in a good and valid point (‘‘prior self-fulfilling prophecies might influence that which is accurately perceived’’). Prior self-fulfilling prophecies might have influenced that which is perceived, but it does not mean accuracy cannot be studied.

E. STEREOTYPE ACCURACY AND LEVELS OF ANALYSIS The following represents a distillation of a criticism of the notion of stereotype accuracy that has periodically appeared in the social psychological literature (e.g., Allport, 1955; American Psychological Association, 1991; Fiske, 1998; Hamilton, Shermon, & Ruvolo, 1990; Nelson, 2002; Stangor, 1995): Even if it can be successfully shown that perceivers accurately judge two groups to diVer on some attribute, 1. Perceivers cannot assume that their stereotypes of the group automatically fit all members of the group; 2. Perceivers cannot apply their belief about the group when judging individuals; 3. If perceivers do apply their belief about the group when judging individuals, they are likely to be wrong much of the time because few members perfectly fit the stereotype.

If all stereotypes are known to be largely inaccurate (as this logic suggests), the need to assess their accuracy would be rendered moot. This criticism has some validity, but that validity depends, in part, on what this type of statement means. To the extent that the ‘‘perceivers cannot’’ statements represent moral injunctions, rather than statements about accuracy, they are beyond the scope of a consideration of the accuracy of social beliefs. However, to the extent that ‘‘perceivers cannot’’ means ‘‘they would reach inaccurate judgments if they did,’’ these arguments are a central focus of this chapter. This line of reasoning’s suggestion, however, that all stereotypes are inaccurate because most members of a group fail to fit a stereotype is only partially justified. It is true that most members of a group will fail to

24

LEE JUSSIM

perfectly fit a stereotype. This, however, does not mean that the stereotype is inaccurate. To understand why requires understanding how this reasoning confounds at least three diVerent levels of analysis and how considerably greater conceptual clarity can be brought to understanding stereotype accuracy by clearly distinguishing among these levels of analysis. Table I presents an analytic breakdown of diVerent levels of analysis at which accuracy can be assessed. 1. Stereotypes as Perceptions of Populations The first row in Table I refers to stereotypes—beliefs (or generalizations) about whole populations (typically, but not always, large demographic groups). The level at which one must measure the criterion for assessing the accuracy of beliefs about groups is the population that comprises that group. Claims about the characteristics of New Yorkers (or women or African Americans or librarians) should be compared with the characteristics of a representative sample or the whole population of New Yorkers (or women or African Americans or librarians). It is not possible to evaluate the accuracy of a belief about Asians in general by using as a criterion the characteristics of a particular Asian target. To do so would be equivalent to evaluating the claim that ‘‘Alaska is cold’’ by measuring the temperature at noon on July 4 in Anchorage. Census figures, results from randomly selected samples, and meta-analyses of hundreds of studies have all been justifiably used as criteria against which to compare the accuracy of people’s stereotypes (e.g., Judd et al., 1995; McCauley & Stitt, 1978; Swim, 1994; although, how to measure criteria when assessing accuracy is discussed in detail later in this chapter). The evidence that has slowly accumulated regarding the validity of people’s beliefs about populations is that they are usually, though not always, at least moderately accurate, and that people are about as likely to underestimate as to overestimate real diVerences between population groups (see, e.g., Diekman et al., 2002; McCauley, 1995; Ryan, 2002). Such research, however, cannot and was never intended to evaluate the accuracy of people’s perceptions of individuals from diVerent groups, which requires a level of analysis below that of whole populations. 2. Stereotypes and Person Perception: Accuracy and Bias in Judging DiVerences Between Individuals from DiVerent Groups The second row of Table I presents a second level of analysis for assessing accuracy—that of perceptions of diVerences between individuals belonging to diVerent groups. Perceivers making claims about the diVerences between the Christians and atheists they know personally, or between their own sons

ACCURACY IN SOCIAL PERCEPTION

25

TABLE I Identifying the Appropriate Level of Analysis in Studies of Social Perceptual Accuracy Level of criteria for assessing accuracy of that social belief

Level of analysis

Social belief is a

Population (This level assesses the accuracy of a stereotype about a group). Research examples: Judd et al., 1995; McCauley and Stitt, 1978; Swim, 1994

Stereotype Regarding an Entire Population Examples: 1. An introductory psychology student believes that White Americans are wealthier than African Americans. 2. A high school teacher believes that teenage boys are better at math than are teenage girls.

Small Group (This level assesses the accuracy of beliefs about diVerences between specific individual targets belonging to diVerent groups). Research examples: Brodt and Ross, 1998; Clarke and Campbell, 1955; Madon et al., 1998

Perception of DiVerences Small Groups Between Specific Individual Members of Social Groups Examples: 1. An introductory psychology 1. The wealth (net worth: student sees little diVerence yearly income) of the between the wealth of African African American and American and White students White students in that in his class. student’s introductory 2. A high school teacher psychology class. believes the girls in her 2. Performance in class and class are doing better at on standardized tests of the math than are the boys boys and girls in this in her class. teacher’s class.

Individual (This level assesses accuracy in perceptions of individuals, not in perceptions of population or small group diVerences). Research examples: Funder, 1987; Jussim, 1989; Kenny, 1994

Person Perception Examples: 1. An introductory psychology student believes that Mary Anne is wealthier than Rashid who is wealthier than Lois. 2. A high school teacher believes that John is doing better at math than Bonita who is doing better than Lou.

Population

1. Income of White Americans and African Americans in a nationally representative sample or in the U.S. Census. 2. Meta-analyses of hundreds of studies assessing sex diVerences in teenagers’ math performance.

Individuals 1. Mary Anne’s, Rashid’s, and Lois’s wealth.

2. John’s, Bonita’s, and Lou’s performance on math tests.

26

LEE JUSSIM

and daughters, or between the police in two diVerent towns, are all making claims about diVerences between small groups. They are not necessarily making claims about whole populations. A soccer coach might believe that the girls on her team play better than do the boys on her team without necessarily believing that girls are typically better soccer players than are boys. This level of analysis addresses the role of stereotypes in causing systematic inaccuracy in perceivers’ judgments about individuals they know personally. Such claims occur at a diVerent, smaller, level of analysis than do claims about diVerences between whole populations. Assessing the accuracy of the perceived diVerence at this level of analysis must be accomplished by comparing the perceived mean diVerence between targets from diVering groups to the actual mean diVerence. A handful of studies have addressed accuracy at this level of analysis. For example, Clarke and Campbell (1955) had students (African American and White) predict each other’s performance on an upcoming test. White students’ perceptions were a mix of accuracy and bias (slightly underestimating African American achievement), whereas the African-American students predicted the achievement of the African–American students quite accurately (they did not report results for each group’s predictions of White students). Other research has assessed the accuracy of teachers’ perceptions of diVerences between students from diVerent demographic groups in their classes and the accuracy of college students’ perceptions of individuals belonging to diVerent college residence halls, and has found a similar pattern of high accuracy and small bias (Brodt & Ross, 1998; Jussim et al., 1996; Madon et al., 1998). 3. Person Perception The third row of Table I presents a third level of analysis—the individual target.1 At this level of analysis, most stereotype accuracy questions disappear. Accuracy in the perception of diVerences between individuals belonging to diVerent groups can no longer be assessed. Without some comparison of diVerences in perceptions of groups (large or small), only accuracy in the judgment of individual targets can be assessed. For example, many studies examine how well teachers’ perceptions of their individual students’ performance corresponds with those students’ actual performance (Jussim & Harber, in press). How teachers arrives at 1

This corresponds to what is frequently called the dyadic level of analysis (e.g., Jussim, 1991; Kenny, 1994), because there is one perceiver and one target. However, the discussion of all three levels of analysis here has assumed a single perceiver, and that what is varying is the number and nature of the targets (population, small group, individual). Therefore, ‘‘person perception,’’ is discussed here as occurring at the individual level of analysis.

ACCURACY IN SOCIAL PERCEPTION

27

their perceptions—whether through the use of stereotypes, shoe size, or acute insight—is not relevant at this level of analysis to evaluating the accuracy of their judgments of individual students. No matter how they arrived at their evaluation of a student, that evaluation either is or is not well-justified by that student’s actual performance. Similarly, each of several perceivers might judge the personality attributes of each of several new acquaintances. Accuracy is determined by comparing the judgments to some criteria for measuring those acquaintances’ personalities. Accuracy at this individual level of analysis is frequently at least moderately high (e.g., Funder, 1987; Jussim, 1991; Kenny, 1994). Although stereotypes might influence those judgments, the role of such influence in accuracy can rarely be determined at this level of analysis. Instead, to determine whether stereotypes increase or reduce accuracy in judgments of individuals, researchers will usually need to group targets (by gender, ethnicity, or whatever stereotype one is studying) and then compare perceived group diVerences to actual group diVerences (i.e., at the intermediate level of analysis shown in Table I). For example, a business owner’s evaluations of employees might correlate .6 with those employees’ overall performance, indicating moderately high accuracy. Such accuracy tells us nothing, however, about whether the owner exaggerates diVerences between males’ and females’ job performance. Few of the large number of studies comprising the stereotypes and person perception literature were designed to assess accuracy. Instead, most were intended to assess psychological processes involved in stereotyping. Thus, although they do not address accuracy, this limitation does not constitute a threat to the main purposes of such studies. It is nonetheless important to highlight their inability to address accuracy, because so much of the discussion of this literature states or implies that stereotypes typically lead to inaccurate judgments of individuals (American Psychological Association, 1991; Darley & Fazio, 1980; Fiske, 1998; Fiske & Neuberg, 1990; Jones, 1986, 1990; Jost & Kruglanski, 2002). One reason for this common misinterpretation of this literature is that it often yields results showing that, even when targets have identical personal characteristics, accomplishments, or behaviors, people view targets diVerently, depending on whether they belong to one group or another. If two people (or people in two experimental conditions) hold diVerent views of targets with identical personal characteristics, doesn’t this imply that that at least one of them must be wrong? Not necessarily. First, few such studies included targets with real attributes against which the accuracy of perceiver judgments could be evaluated. In the absence of targets with real behaviors and attributes, it is often impossible to evaluate the accuracy of perceiver judgments. For example, there rarely are

28

LEE JUSSIM

standards against which to compare perceivers’ behavioral predictions or trait inferences regarding fictitious targets invented by researchers. Second, many researchers seem to assume that when targets’ behavior is identical, perceivers reaching diVerent conclusions regarding targets from diVerent groups are inherently inaccurate. This assumption is true sometimes, but not always. For example, if two students get 8 of 10 math problems correct on a test, perceiving those students as correctly solving 8 of 10 problems is the only correct perception. In this situation, if two individuals or groups diVer in their estimation of the number of problems correctly solved, one or both must be inaccurate. However, if the judgment task involves inferring their general level of intelligence or predicting their future performance, it might not be accurate for the inferences and predictions to be identical. For example, if one student was in an enriched math class and the other in a remedial math class, the most accurate inferences and prediction might not be identical, despite the students’ identical performances on this one test. Most likely, the students were placed in these diVerent classes because they had dramatically diVerent records of past math achievement. Even if the perceiver was not aware of those diVerences, they would typically be more accurate predicting that the student in the enriched class would perform more highly in the future than would the student in the remedial class (as compared to predicting their performances to be identical). Last, in the rare cases in which experimental studies of the role of stereotypes in person perception include clear criteria for evaluating accuracy, almost none include no label or no stereotype control groups. It is possible that, although label A may increase accuracy relative to label B, any label increases accuracy compared to no label. To reach conclusions about whether relying on a stereotype or having a label to guide perception increases or decreases the accuracy of perceivers’ judgments, of course, such control groups would be necessary. Although many perspectives suggest that reliance on stereotypes is both inaccurate and immoral (e.g., American Psychological Association, 1991; Devine, 1995; Fiske, 1998; Fiske & Neuberg, 1991; Jones, 1986; Stangor, 1995), the only study providing data capable of empirically comparing the accuracy of person perception judgments in the presence versus absence of a label (Cohen, 1981, experiment 2) found that having a stereotype label before exposure to a target increased accuracy. 4. Stereotype Accuracy and Level of Analysis: Conclusion Claims suggesting that stereotypes are inaccurate because they do not apply to all individual members of a group (Allport, 1958; American Psychological Association, 1991; Fiske, 1998; Hamilton et al., 1990; Nelson,

ACCURACY IN SOCIAL PERCEPTION

29

2002; Stangor, 1995) are both true and false. The claim that stereotypes cannot possibly apply to all individual members of a group is completely true. The suggestion that this renders stereotypes inaccurate is, however, unjustified because it confounds two levels of analysis (population and individual). A claim about a population cannot be evaluated against the characteristics of an individual, or even against subsets of individuals. Consistency between the level of the perception and the level of the criterion must be maintained when assessing accuracy by comparing beliefs about populations (stereotypes) to characteristics of those population groups, beliefs about diVerences between small groups of individuals to the actual diVerences between those small groups of individuals, and beliefs about an individual to the characteristics of that individual. 5. The One Exception: Absolutist Stereotypes Absolutist stereotypes—beliefs that all members of a group have some attribute—will indeed almost always be false, because there are almost always wide variations among individuals. A single exception invalidates an absolutist belief. Just as a belief that the temperature in all locations in Alaska is always below freezing will be disconfirmed by a single reading of 33 degrees Fahrenheit in Juneau on July 15 at 1 p.m., a belief that all Germans are eYcient will be disconfirmed by discovery of a single ineYcient German. The vast accumulated empirical evidence on stereotypes, however, has yet to report a single person who holds absolutist stereotypes. Instead, the evidence indicates that most stereotypes are quantitative and probabilistic, not absolute (e.g., Judd et al., 1995; McCauley & Stitt, 1978; Swim, 1994). Probabilistic stereotypes, which permit many exceptions and wide variability, can only be evaluated by comparison to population-level criteria. People who hold absolutist stereotypes undoubtedly exist and probably comprise significant portions of extremist groups such as the Ku Klux Klan and neo-Nazis. Nonetheless, such people are atypical of the participants in most scientific research on stereotypes.

F. POLITICAL OBJECTIONS Politics seems to lead to both explicit and implicit objections to accuracy research—especially research on the accuracy of stereotypes. First, people’s political stances may lead them to be more likely to raise scientific concerns about research that they perceive as opposing their political positions than about research that supports their political stances (e.g., Lord, Ross, &

30

LEE JUSSIM

Lepper, 1979). This will almost never be stated quite so explicitly. No researcher will ever state ‘‘I am a liberal’’ (or ‘‘I am a conservative’’) or ‘‘Because I find this research politically oVensive, I am going to work extra hard to come up with intellectual arguments against it.’’ Furthermore, people may often not even be aware of how their politics influences their reactions to research. Politics, however, seems to be a major basis for considering stereo type accuracy research to be a ‘‘problem’’ (e.g., Fiske, 1998; Stangor, 1995) while at the same time not considering research on, for example, automatic stereotyping, ingroup favoritism, ambivalent sexism, or subtle racism to be problems. These areas have their own substantial degree of theoretical and empirical controversy and complexity, so it would be diYcult to make the case that stereotype accuracy research is more problematic than other areas on purely scientific grounds. Thus, political issues often remain implicit undercurrents underlying objections to accuracy research, rather than explicit statements. Occasionally, however, political objections to accuracy research have been explicitly articulated.

1. What Is the Political Criticism of Accuracy Research? The main political objection is that accuracy research, at best, does not help anyone, and at worst, it helps exacerbate social inequalities. Specifically, accuracy research has been accused of doing little more than reifying or reflecting socially constructed injustices or assisting bigots and oppressors in their nasty goals (e.g., Claire & Fiske, 1998; Fiske, 1998; Jost & Banaji, 1994; Stangor, 1995). Before evaluating the validity of this criticism, why accuracy motivates this type of politically based criticism is briefly discussed.

2. Why Does Accuracy Arouse Political Motivations? Accuracy runs against the grain of many social scientists’ concern for helping alleviate social inequalities and injustices. Demonstrating that some aspects of many people’s sex stereotypes are accurate (Diekman et al., 2002; Hall & Carter, 1999; Swim, 1994) or that some aspects of many people’s racial stereotypes are accurate (McCauley & Stitt, 1978; Ryan, 1996, 2002; Ryan & Bogart, 2001) does little to alleviate or explain injustices associated with sexism or racism. Worse yet, demonstrating social perceptual accuracy can be viewed not merely as documenting high acumen in perceiving individual and group diVerences but as reifying those diVerences.

ACCURACY IN SOCIAL PERCEPTION

31

To say that a belief that some child is unintelligent or is awkward on the basketball court or is socially inept is ‘‘accurate’’ has a feel of ‘‘blaming the victim.’’ If the belief is accurate, then we cannot point to perceivers’ errors, biases, misconceptions, egocentrism, or ethnocentrism as explanations for target diYculties. The unintelligent, unathletic, or socially awkward target, in these cases, really is flawed in some way. It is especially distasteful to suggest that some negative beliefs regarding large demographic groups (i.e., stereotypes) are valid. People who publicly declare that two groups diVer in some societally valued attribute (intelligence, motivation, propensity for alcoholism or crime, morality, etc.) run the risk of being accused of being an ‘‘-ist’’ (racist, sexist, classist, etc.) or, at minimum, of holding beliefs that do little more than justify existing status and hierarchy arrangements (e.g., Jost & Banaji, 1994; Sidanius & Pratto, 1999). In contrast, an emphasis on bias and error implies a benevolent and egalitarian concern with injustice. Such an emphasis implies that the emphasizer believes that so-called ‘‘real’’ diVerences between groups do not result from any actual attributes of members of those groups (their cultures, their religions, their histories, their economic conditions, their geography, their practices, their politics, their genetic predispositions)—they result solely or primarily from the oppressive eVects of others’ misbegotten beliefs. In addition, an emphasis on cognitive errors and biases provides a clear villain—the person committing the error or bias. It also points to a relatively straightforward way to ameliorate some social inequities—change the source of the bias (people’s beliefs, expectations, processing, motivations, etc.). Whereas belief change is not necessarily an easy thing to accomplish, it is much easier than changing the mean level of competence among a large portion of the population. In contrast, if the belief about group diVerences is accurate, then changing the situation may require work on a grand scale. To make groups more equal, we must upgrade one group’s competence (we could also downgrade the higher-status-groups’ competence, which may appeal to some people with deeply felt resentments, but which most scientists would agree is not the optimal solution). Doing so may require years of education, training programs, mentoring, and the like. This is both much more labor intensive and typically not under the purview of most social psychological activities. Furthermore, a series of studies on interracial stereotypes in the United States have converged on the conclusion that many Whites have adopted an ‘‘egalitarian’’ ideology that, as part and parcel of its emphasis on intergroup equality, actively denies diVerences between groups (Judd et al., 1995; Ryan, 1996, 2002; Wolsko, Park, Judd, & Wittenbrink, 2000). This denial-ofdiVerence ideology is not shared by many African Americans, who have adopted a ‘‘multicultural’’ ideology that involves acknowledging both the

32

LEE JUSSIM

positive and negative diVerences between groups (see also Lee et al., 1995; Ryan, 2002 for discussions of the similarities between multiculturalism and stereotype accuracy). Nor is this denial-of-diVerence ideology shared by many researchers outside the United States—despite active programs of research on stereotypes and prejudice in Australia, Canada, and Europe, the most vigorous critiques of the concept of accuracy, especially when applied to stereotypes, have been produced by American researchers (e.g., Claire & Fiske, 1998; Fiske, 1998; Jones, 1985, 1986; Stangor, 1995). In contrast, for example, an Australian team, in a broad and sweeping review of research on stereotypes, characterized the social cognition emphasis on stereotype inaccuracy as ‘‘mythic’’ (Oakes, Haslam, & Turner, 1994). Regardless, researchers (whether White or not, and whether in the United States or not) who share the type of denial-of-diVerence egalitarian ideology uncovered in the research by Ryan and Judd and their colleagues may be oVended by research on stereotype accuracy because it implies the potential existence of bona fide diVerences between groups. This may also help explain the relatively greater political appeal to many social psychologists and other social scientists of error, bias, and process compared to accuracy, especially with respect to stereotypes. 3. Are the Political Criticisms Justified? Does accuracy research blame the victim and justify and exacerbate inequalities, as suggested by the politically motivated criticisms? Social and political judgments have the potential to influence all sorts of research, especially in the social sciences (Herrnstein & Murray, 1994; Jacoby & Glauberman, 1995; Redding, 2001). In this sense, political analyses of accuracy research are justified, but no more than they are in other areas of research. Accuracy research is neither more nor less subject to such influences than other social science research. The idea that accuracy research somehow reifies, creates, or justifies social inequalities and injustices is simply false—at least if the criterion is what researchers studying accuracy have actually written. I know of no articles that provide or review evidence of accuracy that state, imply, or even hint that accurately perceived diVerences between individuals or groups reflect immutable, permanent, or essential features of those people (see, e.g., Brophy, 1983; Eagly & Diekman, 1997; Funder, 1987, 1995; Ickes, 1993, 1997; Judd & Park, 1993; Jussim, 1991; Kenny, 1994; Lee et al., 1995; McCauley et al., 1980; Ryan, 1995, 2002). Regardless, as in most social science fields, strong steps can be taken to considerably reduce political bias from entering into consideration of issues involving accuracy (discussed later in

ACCURACY IN SOCIAL PERCEPTION

33

this chapter; see also Funder, 1987, 1995; Judd & Park, 1993; Kenny, 1994; Ryan, 2002). Whether it is scientifically and politically more appropriate to deny and ignore group diVerences or to understand and acknowledge such diVerences is clearly a contested and controversial issue. For both scientific reasons (discussed throughout this chapter) and political reasons (discussed next), accuracy research involves understanding and acknowledging such diVerences. 4. Accuracy Research As A Tool in the Alleviation of Social Problems Even if the sole purpose of research were to alleviate social problems, wouldn’t we want to know whether people’s beliefs about groups and their individual members (i.e., stereotypes) are accurate? We should want to know for several reasons. First, if we think we are curing a social disease by changing people’s inaccurate, biasing, or misbegotten beliefs about groups, our eVorts will be sorely misplaced if their perceptions of groups and individuals are already accurate. Second, if some beliefs are widely inaccurate and some are reasonably accurate (as is likely the case), only research directly and empirically assessing the accuracy of stereotypes could possibly tell us which beliefs and whose beliefs need to be changed through social interventions. Furthermore, we need to be able to assess and understand accuracy to improve the quality of people’s judgments and evaluations. Only after determining that some people hold highly inaccurate beliefs would it be reasonable to begin work on changing those beliefs. Work on changing inaccurate beliefs itself would only be useful if it were conducted after we knew how to lead people to arrive at more accurate judgments. Of course, there will be no way to assess our success at leading people to adopt more accurate beliefs unless we have techniques to assess accuracy. By understanding what leads people astray and what leads them to accurate judgments, we will be much more capable of harnessing those factors that lead to accurate judgments and, therefore, reduce social problems resulting from inaccurate beliefs. 5. Accuracy Can Lead To Discrimination Although most of the political criticisms of accuracy and accuracy research are not justified, social perceptual accuracy can indeed lead to unequal distribution of valued resources to diVerent groups and to discrimination. The perspective taken in this chapter is that it is more valuable to understand the social and psychological processes that might lead accuracy to cause discrimination than to simply deny the existence of

34

LEE JUSSIM

accuracy or to suppress accuracy research. At least two classes of such situations are distinguishable: situations in which accurately perceived group diVerences lead to justifiable diVerences in group outcomes, and situations in which accurately perceived group diVerences lead to unjustified outcomes for at least some individuals. Each class is discussed next. a. Accurately Perceived Bona Fide DiVerences Between Individuals from DiVerent Groups Can Lead to Group DiVerences in Outcomes. People with low incomes, for example, will typically not be eligible to receive as large a mortgage from a bank as are people with higher incomes. If bank loan oYcers accurately judge others’ wealth and income, they will oVer larger mortgages to wealthier people. Regardless of whether one considers this evidence of discrimination against poor people, demographic groups often diVer in their wealth. On average, in the United States, Asians, Whites, Latinos, Native Americans, and African Americans rank from highest to lowest, respectively, in yearly household income (U.S. Census, 2002). If bank loan oYcers are good at judging wealth, and even if they are completely unprejudiced, Asians will be more likely to be approved for larger loans than will be African Americans. Similarly, many middle-class jobs require a college education. In the United States, a higher proportion of Whites than Latinos complete college. Therefore, even if race and ethnicity–based discrimination evaporated, employers who accurately evaluate applicants’ level of education would provide proportionately more of those jobs to Whites. Accurately perceived diVerences in skills, competencies, accomplishments, and so forth between individuals belonging to diVerent groups can sometimes produce mean group diVerences in valued societal outcomes (jobs, college admissions, etc.). b. Accuracy Leads to Bona Fide Discrimination. It is also possible, however, that accurately perceived group diVerences can simultaneously lead perceivers to be as accurate as possible in their judgment of individuals and to unfairly discriminate against many members of certain groups. Because this issue is politically charged, a nonsocial example is used first to illustrate the basic principles. A U.S. penny is almost, but not quite, perfectly balanced—when flipped in the air, it will come down as heads slightly more frequently than tails. If one wished to successfully predict the outcome of a coin toss, one would maximize one’s accuracy by predicting heads 100% of the time. Such a person could be justifiably characterized as having an ‘‘extreme heads bias,’’ but the predictions would also be as correct as possible. Such a situation characterizes U.S. car insurance policy rates. Young people, and especially young males, are legally charged far more for car insurance than are middle-aged adults. This is because young people,

ACCURACY IN SOCIAL PERCEPTION

35

and especially young males, are actuarially more likely to have accidents. Although undoubtedly true at the group level, this policy clearly discriminates against individual young adults who drive safely. A structurally similar situation, but one in which the discrimination is illegal in the United States, is one in which a bank uses race/ethnicity as a basis for making loans. Because credit history and past income are not 100% diagnostic of future ability to pay, however, in an actuarial (or Bayes’s theorem) sense, it might be more accurate to predict that an Asian with a mixed credit history will be more likely to pay oV a loan in the future than will a Native American with an equally mixed credit history. Nonetheless, instituting a policy that denied loans to Native Americans with credit histories comparable to those of Asian Americans who received loans would clearly constitute illegal discrimination. The potential for bias to enhance accuracy and then to produce discrimination is probably one of the least well understood (or perhaps merely least discussed) processes and set of interlocking phenomena in the social sciences. Bias is frequently seen as implying inaccuracy; accuracy is frequently assumed to be antithetical to bias and discrimination. Such a view is, however, not only seriously flawed, but may account for some of the resistance to and criticisms of accuracy, and especially stereotype accuracy, within social psychology. To the extent that researchers view accuracy as conflicting with their egalitarian goals of alleviating social problems, it is understandable that they may be tempted to denigrate accuracy research. But understanding that bias sometimes enhances accuracy, that accurately perceived group diVerences in inputs may lead to group diVerences in outcomes, or that biased and discriminatory predictions may be as accurate as possible under some circumstances seems crucial with respect to the scientific analysis of intergroup relations and social policy. The goals of accuracy and social justice sometimes conflict. It seems unlikely that deciding how to reconcile those goals will be best accomplished by merely denying accuracy. It seems more likely that a thoughtful understanding of the ways in which accuracy, bias, and discrimination are intertwined, and not necessarily mutually exclusive, could contribute to informed and durable policy decisions regarding issues such as civil rights legislation, antidiscrimination laws, and aYrmative action. Social scientists and lawmakers will be in a much stronger intellectual position to weigh the relative costs and benefits of accuracy and egalitarianism when making policy if they have a more complete understanding of those costs and benefits than if they merely deny lay accuracy or attempt to suppress research on accuracy.

36

LEE JUSSIM

6. Conceptual Objections to Accuracy Research: Conclusions Within social psychology, accuracy research has had a turbulent and controversial history, which probably helps explain why such little research on accuracy was performed from about 1955 to 1985. This, in turn, may help explain the minimal discussion of accuracy in most major reviews of social cognition and social perception phenomena, such as stereotypes, expectancies, schemas, categorization, attribution, and person perception (see, e.g., the 1985, 1996, and 1998 editions of the Handbook of Social Psychology, and the last 30 years of Annual Review chapters on these topics). This section has reviewed many common conceptual objections to accuracy research. It has suggested that many of these objections have considerable validity, in the sense that the criticisms may be true and in that they often raise interesting and important questions. Accuracy in perceptions of an attribute really is diVerent than accuracy in explanations for that attribute. Prior self-fulfilling prophecies might explain some diVerences between targets that are accurately perceived. Despite whatever validity they might have, however, this critical evaluation of the common criticisms leads to the conclusion that none constituted serious threats to the value or interpretation of existing research on social perceptual accuracy. One important and common objection to accuracy research has not yet been discussed—the so-called criterion ‘‘problem.’’ The selection and identification of appropriate criteria for evaluating accuracy has been frequently characterized as a problem suYciently severe to call the viability of accuracy research into question (e.g., Fiske, 1998; Jones, 1985, 1990; Snyder & Stukas, 1998; Stangor, 1995). Next, therefore, this review will critically evaluate whether such issues are more complex or problematic than those facing most areas of psychological research.

VII. How Can One Establish Criteria for Evaluating the Accuracy of Social Perception? Only gods know Absolute Truth. The goal of accuracy research is not to establish Absolute Truths. Instead, the criteria against which social beliefs should be evaluated are similar to the criteria against which psychological hypotheses are evaluated—truths with a small ‘‘t.’’

ACCURACY IN SOCIAL PERCEPTION

37

A. TRUTH WITH A SMALL ‘‘T’’ Truth with a small ‘‘t’’ is evidence, preferably from a variety of sources, indicating that some belief is valid (true, warranted, justified, etc.). In short, the solution to the problem of partially fuzzy, intangible constructs is to establish their validity through multiple methods and approaches (e.g., Campbell & Stanley, 1963; Cook & Campbell, 1979; Cronbach & Meehl, 1955). Truth with a small ‘‘t’’ is typically established via rigorous methodological variations of the ‘‘duck’’ test (Block, 1993; Funder, 1995). If it looks, walks, acts, and sounds like a duck, although the possibility remains open that it really is an antelope, sport utility vehicle, alien visitation, or bacterial growth, it is most likely a duck. A person is, of course, more likely to confuse a duck with a goose or a cormorant than with the Statue of Liberty or a praying mantis. Even if that goose is believed to be a duck, however, probabilistic realism further assumes that it would eventually be possible to receive information that corrects this faulty view. The same analysis applies to all the fuzzy attributes studied by psychologists. Does Voter A have a positive attitude toward the current president of the United States? We might ask A whether she supports the current president, voted for him in the last election, likes him, would vote for him in an election held today, and belongs to the same political party as the president. If Voter A says she does support and like the president, voted for him last time, would vote for him again, and belongs to his political party, classifying her as having a positive attitude toward the president does not seem unreasonable. A similar analysis could be applied to mental and emotional states, self-esteem, personality, expectations, schemas, intelligence, stereotypes, prejudice, motivation, and indeed, the whole pantheon of constructs studied by psychologists. The myriad of papers on these and similar hypothetical, intangible constructs are written as if the authors believe these constructs are real. If the constructs filling the journals and texts are ‘‘real enough’’ for psychologists to study in their scientific research, why are they not ‘‘real enough’’ to constitute criteria for evaluating the accuracy of social perception, social beliefs, and social judgments? B. CONSTRUCT VALIDITY AND SOCIAL PERCEPTUAL ACCURACY There are only two logically coherent possibilities with respect to the issues of construct validity and criteria for social perceptual accuracy. One possibility is that the constructs routinely studied by psychologists have no

38

LEE JUSSIM

reality, so that the articles relying on and reporting them are meaningless and it makes no sense to use meaningless criteria in research on accuracy. This perspective, though logically coherent, is far too extreme and has few proponents [except perhaps some radical social constructivist work that characterizes mainstream social science research as, in essence, story-telling (e.g., Gergen, 1985)]. A second possibility is that the constructs routinely studied by psychologists are real enough, so that the articles relying on and reporting them are meaningful. It would, therefore, make sense to use those same criteria when studying accuracy. This is the perspective advocated here. Of course, even this perspective permits the possibility of there actually being no good criteria for certain types of social perceptions, judgments, or beliefs. If scientific research has yet to develop a valid method for measuring a construct that constitutes the content of some perception, judgment or belief, then it would actually and indeed literally be the case that there is no valid measure against which to assess the accuracy of that perception, judgment, or belief. Given the wide variety of measures available for assessing all sorts of psychological constructs, however, such situations are likely to be the exception, rather than the rule. 1. A Logically Incoherent Possibility: The Constructs Psychologists Study Are Real, But There Are No Good Criteria for Assessing Accuracy Articles that take for granted the validity of numerous social psychological constructs, such as attributions, perceptions, stereotypes, identity, attitudes, and so on, sometimes claim that establishing criteria for assessing accuracy is problematic (e.g., Fiske, 1998; Jones, 1985; Stangor, 1995). In general, it is logically incoherent to assume that the constructs studied by psychologists are real but to characterize accuracy criteria as problematic. One cannot claim that attitudes, personality, behavior, emotions, and so forth can be studied and then claim that we have no criteria against which to evaluate people’s beliefs about others’ attitudes, personalities, behaviors, or emotions. 2. Self-Fulfilling Prophecies, Accuracy, and the Criterion ‘‘Problem’’ Double Standard An extraordinary double standard regarding criteria issues has emerged within the social cognition tradition. Many researchers have embraced and emphasized the power of self-fulfilling prophecies as a mechanism by which social beliefs, and especially stereotypes, create objective social realities (e.g., Claire & Fiske, 1998; Fiske & Taylor, 1991; Jones, 1986, 1990; Jost &

ACCURACY IN SOCIAL PERCEPTION

39

Kruglanski, 2002; Miller & Turnbull, 1986; Snyder, 1984, 1992). Many of these same researchers, however, have expressed doubts or concerns about the viability of identifying objective scientific criteria for establishing accuracy (see also Fiske, 1998; Kruglanski, 1989; Jones, 1985; Snyder & Stukas, 1998). This is peculiarly ironic because, although the processes by which a perceiver’s beliefs become true are diVerent for accuracy and self-fulfilling prophecy, the criteria for establishing their trueness must be identical. To establish either accuracy or self-fulfilling prophecy, one must show that a belief corresponds to some criterion. There is a diVerence only in the process—in how the correspondence comes about—but the criteria for establishing whether any particular belief is true must be just as good (or bad) as the same criteria, regardless of whether they are used in self-fulfilling prophecy research or in accuracy research. Researchers are logically coherent if they accept the criteria used to establish correspondence between perception and reality in both selffulfilling prophecy and accuracy research, a position that is well-justified. A position suggesting that ‘‘both accuracy research and self-fulfilling prophecy research should be dismissed’’ because of criteria problems is internally consistent and logically coherent. This position is so extreme, however, that no articles advocate it. It would be, however, logically incoherent to accept the criteria used to establish correspondence between perception and reality in the service of self-fulfilling prophecy research while rejecting the same criteria when used to establish correspondence between perception and reality in the service of accuracy research.

C. WHAT IS AND IS NOT INCLUDED IN THIS SECTION ON CRITERIA This next section identifies and describes four broad classes of criteria: objective criteria, behavior, agreement with others, and agreement with targets. Each type of criterion is defined and distinguished from other types of criteria. When a certain type of criterion has been controversial, those controversies are discussed and critically evaluated. Finally, the viability of each type of criterion is also critically evaluated. This next section, however, only addresses how to identify plausible criteria. It does not address how best to use those criteria for estimating accuracy (this is addressed in a subsequent section on componential and other approaches for assessing accuracy). The term ‘‘criteria’’ here is not used in the general sense of ‘‘how to assess accuracy.’’ Even the limited aspects of ‘‘how to assess accuracy’’ that are addressed in this review involve the issues raised throughout this entire

40

LEE JUSSIM

chapter (not any one section) and cannot possibly reflect all the issues and details involved in assessing accuracy in all the ways and contexts in which it has been assessed (see e.g., Ambady & Rosenthal, 1992; Brunswik, 1952; Cronbach, 1955; Funder, 1995; Ickes, 1993, 1997; Judd & Park, 1993; Jussim, 1991; Kenny, 1994; Kruglanski, 1989; McCauley et al., 1980 for such details). Thus, the term ‘‘criteria’’ is used quite narrowly herein to refer to the measures, variables, or constructs that might be used as standards against which to evaluate the validity of social perceptions, judgments, and beliefs. As shall be seen, however, even the issues surrounding criteria defined narrowly are both complex and controversial. D. OBJECTIVE CRITERIA Criteria are objective when that which is being judged is assessed in a standardized manner that is independent of the perceiver’s judgment. Many interesting and important social phenomena involve simple, clear, objective criteria. A tennis player wins or loses the point. A salesman makes or loses the sale. An applicant is either accepted into or rejected from a college. There is nothing the least bit diYcult or ‘‘problematic’’ about the criteria for predictions regarding these events. Ninety-nine times out of 100, or more, there will be no controversy on whether the ball was in or out. In the event that one player hits a clear and obvious winner, we all agree that it was a winner, but the criterion for evaluating whether or not it was a winner is not our agreement. We agree because we all saw the ball clearly land in and because the second player did not even hit the ball. Agreement is a result of accuracy, not a criterion for establishing accuracy. Even John McEnroe, perhaps the most argumentative professional tennis player of all time, at his most argumentative only argued a few calls per match—and a tennis match includes hundreds of shots. All sorts of outcomes occur independent of individual perceivers’ beliefs, predictions, or expectations (e.g., election outcomes, returns on investments, the achievements and accomplishments of our friends and acquaintances, etc.). These come up all the time in daily life, and the criterion issue is rarely a ‘‘problem.’’ 1. One-Shot Hit or Miss: The Prototypical Case It is perhaps easiest to understand the elegant simplicity in identifying and using objective criteria by starting with a class of cases, referred to here as ‘‘one-shot hit or miss.’’ This means there is a single perceiver making a single judgment about an objective single attribute or behavior of a single target.

ACCURACY IN SOCIAL PERCEPTION

41

Although finding out whether a single person makes an accurate or inaccurate single prediction is not usually particularly psychologically interesting, by virtue of eliminating a slew of methodological, conceptual, and statistical issues irrelevant to the criterion issue, the one-shot hit or miss situation with objective criteria, though oversimplified, is particularly informative. Despite many social psychologists’ emphasis on pervasive ambiguity in social reality (e.g., Gilbert, 1995; Jones, 1990; Ross & Nisbett, 1991), many aspects of human social behavior, including but not restricted to sports, relationships, investments, academics, and politics, are objective. If I predict that my friend’s marriage won’t last 2 years, the criterion is similarly clear and objective—they either do or do not get divorced within 2 years. If Margaret predicted that the Republicans would win a majority of Senate seats in 2002, she was right. If, in early 2000, Bob predicted that the stock market was in the midst of a speculative bubble that was destined to crash and put all his money into money markets and bonds, he clearly made the right decision. The processes by which Bob, Margaret, and I arrived at our beliefs—our propensities to be self-serving, expectancy-confirming, reliant on faulty heuristics, ideologically motivated, or dispositionist in our attributions— are all irrelevant to evaluating whether we are right or wrong. The one-shot hit or miss situation makes it clear that objective criteria exist for many personally and socially significant beliefs, expectations, and predictions, and accuracy can be assessed independent of process. 2. Overall Levels of Accuracy One-shot hit or miss situations, however, reveal few, if any, important psychological principles and rarely have much bearing on any psychological theory or hypothesis. Usually, therefore, research on accuracy requires investigating many people making one or more judgments or predictions about one—and sometimes many—other person(s). Although statistical procedures for assessing accuracy sometimes become more diYcult in such situations, the criterion issue is no more diYcult than in the one-shot hit or miss situation. If a criterion is appropriate for assessing the accuracy of some prediction or judgment in the one-shot hit or miss situation, it is equally and identically appropriate for assessing accuracy when there are multiple judges, judgments, or targets. Dollars in real estate sales, numbers of college acceptances and rejections, and length of a romantic relationship are all exactly identical in their appropriateness as criteria (for judgments of real estate sales success, predictions of college admissions success, and length of relationship predictions, respectively), regardless of

42

LEE JUSSIM

whether one person judges one other person or 1000 people judge 1000 other people. For example, Archer and Akert (1977) developed a test of nonverbal sensitivity that involved assessing people’s ability to accurately identify objective aspects of people, their experiences, or their relationships. In two of their test questions, people were exposed to two women playing with a baby and talking and two men who had just finished a one-on-one basketball game. Participants’ task was to correctly identify the mother of the baby and the winner of the game. In the actual test, there is a series of such vignettes and the more questions answered correctly, the more accurately people are at judging others on this test. This is simple and objective; there is no criterion problem here. The criterion issue is also a nonproblem in much of the deception– detection literature. That is, one question psychologists have asked is ‘‘How good are people at detecting when others lie?’’ For example, Ekman and Friesen (1969, 1974) had targets view either a pleasant film or a horrible gory one. All had to inform perceivers that they had just seen a lovely, pleasant film. The research question was how often people can tell when others are liars. Again, there is no criterion problem. In studies of some aspects of stereotypes, one can also compare people’s perceptions to objective data. In one classic study, McCauley and Stitt (1978) compared people’s beliefs about diVerences between African Americans and other Americans to U.S. Census data (e.g., in size of family and likelihood of completing high school). In others, people’s beliefs about sex diVerences (e.g., in aggression) were compared to the results from meta-analyses assessing sex diVerences (Hall & Carter, 1999; Swim, 1994). Although even the U.S. Census and meta-analyses do not necessarily reveal Absolute Truth, they do provide strong sources of evidence. With this type of data, one can establish overall levels of accuracy, error, and bias. Do people err on the side of believing what others say? Then Ekman and Friesen’s studies should have yielded results showing that people underestimate deception (it did, but their research also typically showed that people do better than chance at accurately detecting deception). Do people exaggerate diVerences between demographic groups? Then McCauley and Stitt (1978) and Swim (1994) should have found people consistently overestimating the diVerences between diVerent groups (they did not—people tended to be accurate and, when wrong, to underestimate real diVerences). These are important and interesting accuracy questions, and the availability of clear, objective criteria in these cases is not problematic.

ACCURACY IN SOCIAL PERCEPTION

43

3. Independent and Standardized but Not Universally Persuasive Objective Criteria Not all people may agree that certain objective criteria are good ones. Such agreement might be irrelevant regarding a target’s height or marital status, but it becomes much more relevant when estimating a target’s extraversion or intelligence via a personality questionnaire or standardized IQ test. Is the personality questionnaire a good one? Is it reliable? Valid? IQ tests, in particular, have a long and controversial history (e.g., Gould, 1978; Herrnstein & Murray, 1994; Neisser, Boodoo, Bouchard, Boykin, Brody, Ceci, Halpern, Loehlin, PerloV, Sternberg, & Urbina, 1996), as has the measurement of enduring personality traits (see Funder, 2001, for a review). To the extent that some people do not find such tests credible, they are likely to discredit or dismiss research on accuracy using such criteria. Thus, use of objective but controversial criteria can be viewed as boiling down to agreement (if one agrees with the criteria, the study assesses accuracy; if one does not agree with the criteria, it does not—see Kruglanski, 1989), and socially and politically, this is probably how things work. People who do not accept a study’s criteria most likely will not accept its conclusions (whether on accuracy or any other social science topic). Often, however, the reverse may happen: People who do not like a study’s conclusions will come up with arguments against the appropriateness of its criteria (Lord, Ross, & Lepper, 1979). Indeed, people are often extremely good at engaging in the intellectual contortions necessary to maintain a cherished viewpoint even in the face of clear evidence that they are wrong (Aronson, 1999; Festinger, Riecken, & Schachter, 1956). The fact that some people do not like or respect some criterion, of course, does not mean that it cannot be used. It only means that one needs to make a credible case for the appropriateness of that criteria in one’s study. For example, standardized test scores have often been used as criteria in studies of the extent to which teacher expectations for students were selffulfilling prophecies or accurate (see edited volumes by Brophy, 1998; Dusek, 1985; see reviews by Brophy, 1983; Brophy & Good, 1974; Jussim, 1986; Spitz, 1999). If teachers are good at identifying which students are smarter, then their expectations should predict but not influence students’ standardized test scores. If teacher expectations do predict standardized test scores without causing them, any real or imagined weaknesses in standardized tests would seem largely irrelevant to interpreting this result. Use of imperfect criteria will generally lead researchers to underestimate people’s accuracy. If so, in sharp contrast to attempts to dismiss the viability of accuracy research because of imperfect criteria, this means that people are

44

LEE JUSSIM

probably even more accurate than indicated by the evidence emerging from any particular study using imperfect criteria. This is obviously the case when using a criterion measure with less than perfect reliability. Unreliability in measurement artificially lowers correlations between judgments and criteria, so that the greater the unreliability, the higher the actual accuracy. Consider a case in which people’s perceptions of others’ extraversion correlate .3 with a self-report extraversion scale. If there is less-than-perfect reliability in measurement of both perception and criterion, then the true correlation will be higher than .3. If, for example, the reliability of both scales was .9, the true correlation between perception and criterion would be .33; if the reliability of both scales was .75, the true correlation between perception and criterion would be .4 (Carmines & Zeller, 1979). This would often be equally true with a scale of imperfect or partial validity. Perhaps a particular psychological measure only captures a partial aspect of some attribute. People’s judgments, perceptions, and so forth, if they reflect more aspects of the attribute than the measure being used, may be more accurate than indicated by the correlation between judgment and criterion. Consider an assessment of people’s accuracy in judging how big a target is. The criterion, however, is height. Height, of course, is only one aspect of size (the others being width and depth). If people use height, width, and depth to judge size, the correlation of their judgments with the (imperfect, partial) criterion will be too low. That means their accuracy could be even higher than indicated by that criterion. Similarly, consider the use of an IQ test as criterion for evaluating the accuracy of judgments of intelligence. If a particular IQ test primarily taps verbal intelligence and people’s judgments include, in addition to estimations of verbal intelligence, creativity, social skill, political savvy, wit, and common sense, if people’s judgments closely correspond to actual intelligence defined in this multifaceted way, the correlation between judgments and criterion will be artificially low. People might be more accurate than indicated by the correlation (correspondence) between judgment and criterion. An identical analysis applies to using any measure, such as a standardized personality inventory, as a criterion for assessing accuracy in judgments of personality traits. Of course, just because such mismatches may lower the observed correspondence between perception and criterion, one cannot infer high accuracy from low correlations, but the argument that any particular criterion in any particular study is ‘‘bad’’ (unreliable, low validity) means that whatever evidence on accuracy is obtained probably often represents a lower bound on accuracy. This, in turn, means that people may actually be more accurate than indicated by the empirical evidence of accuracy obtained in that study. Just as imperfect criteria do not preclude the possibility of

ACCURACY IN SOCIAL PERCEPTION

45

objective, social scientific research on all sorts of topics (aggression, identity, achievement, social class, etc.), nor do they preclude the possibility of accuracy research. Some criteria are subject to debate and controversy. The extent to which nearly all measures of social science constructs have been validated or well established has limitations, thereby inflicting imperfections on any research using any measure of any construct. Researchers should almost always carefully evaluate the flaws and limitations of their research, including their use of measures. However, this issue is no more a problem for accuracy research than for any other type of research.

E. BEHAVIOR 1. What Is Behavior? In the broadest sense, anything a person (or any organism) does is behavior. In this sense, answers on a standardized test constitute behavior. Similarly, agreement among judges usually constitutes agreement concerning either targets’ behavior or inferences about their attributes based on their behavior (Kruglanski, 1989). Even completely unobservable thinking, feeling, or daydreaming can be considered behavior. Indeed, transmission or inhibition of neurotransmitters can be considered behavior at the level of the individual cell. None of this is what is meant by behavior here. For the social perception contexts addressed in this chapter, behavior is considered to be observable action, not unobservable thought processes, cell electrochemical processes, or underlying attributes. Counting how frequently Natasha smiled is related to but diVerent from determining whether she is thinking happy thoughts or how sociable she is. Smiles are visible. Similarly, measuring how much time Jorge spends on some task is related to but diVerent from evaluating Jorge’s motivation. Again, time on task is an observable behavior, not an underlying attribute. Similarly, ‘‘answers on a standardized test’’ is not what most social scientists mean when they use the term ‘‘behavior.’’ Such tests are generally designed to measure some attribute that has relevance beyond the testing situation. Intelligence test scores should predict school outcomes, work outcomes, and many life outcomes. Scores on a personality scale measuring hostility should predict levels of angry and aggressive behavior in all sorts of interpersonal contexts. Although standardized test scores could be considered one class of behavior, it is both possible and useful to distinguish between standardized tests and overt actions.

46

LEE JUSSIM

2. Does Behavior Boil Down to Agreement? Some have argued that it does (e.g., Kruglanski, 1989), and for good reasons. At least in most research contexts, a common way to measure human behavior is by having independent judges observe and code the behavior. However, there is a diVerence between the need for people to observe and record behavior and the need for people to interpret behavior. Behavior itself is observable; personality, beliefs, attitudes, motivation, competence, intelligence, and so forth may be inferred from behavior, but they are not directly observable. Observers recording nonverbal behaviors (e.g., smiling, speech dysfluencies, etc.), time spent on some task (e.g., proportion of time spent talking in a dyadic conversation; time spent trying to solve an impossible anagram before giving up; time it takes to walk down a hallway, etc.), task choice (how many advanced math courses does a person take? Do they apply to college? Number of parties attended, etc.) are all recording objective aspects of behavior. Around the fringes, there may be room for interpretation—Was Fred smiling, or was the corner of his mouth just twitching a bit? And it is very useful to have multiple observers record behavior in order to resolve these fringe or ambiguous situations. When that happens, judgments of behavior do boil down to agreement. Nonetheless, when behavior rather than underlying attribute is being recorded, such fringe or ambiguous situations are generally likely to be the exception rather than the rule. Thus, there is an important distinction between using observers to record what a person does and using observers to interpret and evaluate the meaning of what a person does. Behavior is what a person does. Many of the classic programs of research in social psychology have studied phenomena with behavioral criteria, such as research on the attitude–behavior relationship (Eagly & Chaiken, 1993), self-eYcacy (Bandura, 1977), behavioral confirmation (Snyder, 1984), obedience and conformity (Asch, 1955; Milgram, 1974), and helping in an emergency (e.g., Latane & Darley, 1970). I know of no critical work that takes these areas to task for their use of behavioral criteria as somehow ‘‘problematic’’ or even for ‘‘merely studying agreement.’’ The rationale for believing that behavior can be identified when one studies such classic social psychology topics as these, but not when one studies accuracy, has never been explicitly stated in any prior article criticizing or rejecting accuracy research or arguing that accuracy boils down to nothing more than agreement (e.g., Claire & Fiske, 1998; Fiske, 1998; Jones, 1986, 1990; Kruglanski, 1989; Stangor, 1995). It remains a challenge, therefore, for scientists arguing that it is somehow problematic to use behavior as a criterion for studying accuracy to explain why the use of behavioral criteria is not equally problematic in other areas of psychology.

ACCURACY IN SOCIAL PERCEPTION

47

F. AGREEMENT WITH OTHER PERCEIVERS Agreement, at first glance, appears to be a poor criterion for accuracy. Agreeing that the world is flat does not make it flat. Thus, agreement, like other criteria, is almost always imperfect. But agreement cannot be so readily dismissed. If both of us are accurate, we must also agree. If a baseball player hits a home run, and we both saw it, we will both agree that he hit a home run. If Hanna is a brilliant student, and we both recognize her brilliance, we will both agree that she is brilliant. Accuracy causes agreement, although agreement does not necessarily imply accuracy. Furthermore, if we disagree, at least one of us must be wrong (at least if we are talking about the same thing).2 This analysis indicates that accuracy typically increases agreement. In probabilistic or correlational terms, as accuracy among multiple perceivers increases, so should their agreement most of the time. In fact, this is one of the keys to understanding why agreement among multiple perceivers is often a good, if imperfect, index of accuracy. Correlations work two ways: If A is positively correlated with B, then it is equally appropriate to claim that as A goes up so does B as it is to claim that as B goes up so does A. If accuracy and agreement among multiple perceivers will typically be positively correlated, then, as their agreement goes up, so will accuracy most of the time. Thus, agreement is an imperfect but probabilistic indicator of accuracy. In fact, however, even ‘‘agreement’’ itself is a multifaceted construct. Agreement with whom? Experts? Other people? The targets themselves? In fact, all of these types of agreement may, under diVerent circumstances, constitute better or worse criteria for accuracy. The next section, therefore, describes and reviews the types of agreement criteria most commonly used in accuracy research and identifies the major potential limitations of each. 1. Agreement with Experts Perceivers’ judgments or expectations can be compared to those of experts. For example, personality or psychopathology judgments could be compared with those of professional experts, such as clinical psychologists or psychiatrists; ratings of students’ aggressiveness or academic motivation 2

We do need to be careful in considering what constitutes ‘‘the same thing.’’ You and I could disagree about how pleasant, extroverted, or intelligent Fred is. If Fred acts diVerently with you from with me, then we both could be right (Swann, 1984). So if we are talking about Fred in general, both of us may be partially right and partially wrong, because perhaps there is no ‘‘Fred in general’’—there is only Fred interacting diVerently with diVerent people.

48

LEE JUSSIM

might be compared with those of the students’ teachers. For some types of research, nonprofessionals could also be considered ‘‘experts’’ in the sense that they have some sort of unique access to knowledge about the target. Thus, spouses, romantically involved couples, close friends, and coworkers could be used as experts with whom to compare perceivers’ judgments of targets. 2. Agreement with Experts’ Models Sometimes, experts, such as statisticians, decision-making theorists, or psychologists, have developed formal models for what constitutes the most appropriate way to arrive at a judgment. Accounting for regression to the mean, appropriate use of base rates, and use of consensus, consistency, and distinctiveness in arriving at attributions all constitute formal models of the appropriateness of social judgment (see, e.g., Dawes, 1979; Kahneman et al., 1982; Kelley, 1967; Nisbett & Ross, 1980). Thus, judgments or decisions based on such models may be used as criteria against which to evaluate the accuracy of laypeople’s judgments or decisions. 3. Agreement with Independent Judges If a social interaction is recorded, for example, by videotape, transcription, and so on, or if it can be monitored (e.g., through intercoms, computers, or one-way mirrors), people other than the perceiver can be asked to evaluate the target. These people are referred to here as ‘‘independent’’ judges because they were not involved in the interaction between perceiver and target; thus, they are independent of the perceiver. Often in such situations, these independent judges may only be exposed to the targets’ responses to further minimize any eVects of the perceiver’s verbal or nonverbal behavior on the independent judges (e.g., Goldman & Lewis, 1977; Word et al., 1974). Thus, the perceptions of independent judges may be used as criteria against which to evaluate the accuracy of the perceiver’s person perception judgments. 4. Agreement with Nonindependent Judges Sometimes the judgments of people who do interact with the target, either simultaneously while interacting with the perceiver or in other contexts, may be used as criteria for evaluating the perceiver’s accuracy. In such cases, all perceivers’ judgments of targets essentially constitute the criteria for one another. Indeed, Kenny’s (1994) social relations model, which addresses numerous aspects of agreement, accuracy, and social perception, requires

ACCURACY IN SOCIAL PERCEPTION

49

multiple people to interact with and rate one another. Nonindependent judges may be strangers, acquaintances, or experts. 5. Limitations to Agreement with Other Perceivers The perspective of probabilistic realism indicates that nearly all criteria will be imperfect to some degree. Agreement is no exception. The fundamental limitation to use of agreement is, of course, that everyone can be wrong. a. Even Experts May Be Wrong. Indeed, sometimes, expert judgments or predictions may be little better than those of lay people (Cronbach, 1955). Psychiatrists, for example, predicted that hardly any teachers would give 450 volts worth of shock to learners providing wrong answers on a test in Milgram’s (1974) studies of obedience (even though half or more often did). In addition, experts may be subject to their own biases. Happy romantic couples, for example, may see each other in an overly positive, almost idealized way (Murray, Holmes, & GriYn, 1996). The doctors and staV at psychiatric hospitals at least sometimes misinterpret patient boredom or hostility as psychopathology (Rosenhan, 1973). Teachers’ expectations sometimes color their interpretations of students’ performance (Williams, 1976). Thus, expert judgments are susceptible to a variety of potential imperfections. They are probably best used as a criterion for accuracy when prior research has provided strong evidence for their validity. b. Experts’ Models May Be Contradictory. Expert models often have a formal logical rigor or statistical/mathematical basis that gives them immediate credibility. Nonetheless, they may not always be correct. This will be demonstrated below using two examples in which diVerent expert models make opposing predictions (meaning that both cannot possibly be right): one involving the stock market and the other involving stereotypes. i. EYcient markets theory versus regression to the mean. EYcient markets theory (the dominant view of the stock market among economists) states, in essence, that because whatever is known is already factored into stock prices, it is impossible to consistently obtain returns that beat the overall market (Malkiel, 1999). The stock market, according to this view, is essentially a random walk (the future is inherently unpredictable) with an upward overall trajectory. The practical implication is that most people who do not plan to spend their investment dollars any time soon will receive the greatest return by buying a stock index fund that reflects the entire stock market and holding through (unpredictable) ups and downs. One will most likely get oneself into trouble (reduce one’s overall return) if one tries to ‘‘time’’ the market (buy low, sell high) or select individual stocks to beat the market averages.

50

LEE JUSSIM

EYcient markets theory, however, can be viewed as clashing with one of the main claims of statistical experts. The statistical idea of regression to the average means that extreme values are likely to return over time to the overall mean. This could be interpreted to mean that if stock valuations are unusually high, one is more likely to lose money and less likely to make money (going forward) than usual; and if they are unusually low, one is more likely to make money and less likely to lose money (going forward) than usual. If one enters the market when valuations are unusually low, therefore, and exits when they are unusually high, one should be able to beat the overall market (by receiving the unusually large gains that follow low valuations and avoiding the severe losses following high valuations). ii. Bayesian inference versus stereotypes ‘‘biasing’’ person perception. A similar contradiction occurs in expert models regarding use of stereotypes. Many social psychological theories and perspectives on stereotypes, much current cultural discourse, and many laws and judicial decisions state or implicitly assume that people are acting rationally and appropriately only when they judge others solely and entirely on the basis of those others’ personal characteristics, rather than their group memberships. This view argues or implies that people act in a biased, prejudiced, or irrational manner when they allow their stereotypes to influence their judgments of individuals (e.g., Borgida et al., 1995; Brewer, 1988a; Darley & Fazio, 1980; Dunning & Sherman, 1997; Fiske & Neuberg, 1990; Jones, 1986, 1990; Myers, 1999; Nelson, 2002). According to widely accepted principles of probability (e.g., Bayes’s theorem, see e.g., Kahneman & Tversky, 1973), regression (e.g., Jussim, 1991), and philosophy of science (e.g., Krueger & Funder, 2004; Meehl, 1990), however, base rates, prior beliefs, and expectations should often influence people’s interpretations of new evidence (see McCauley et al., 1980 for an analysis focusing specifically on stereotypes). These principles suggest that if perceivers are not sure how tall a person is, they will, on average, be right more often if they estimate any particular male to be a few inches taller than any particular female than if they estimate them to be exactly equal (i.e., use, rather than discard, their sex stereotype regarding height). Even if perceivers find out that both a male and female target are avid basketball players (identical individuating information), the best guess, according to these principles, is still that the male is taller than the female. Only if the perceivers find out that both are 6 feet tall, which is perfectly diagnostic of height, should they not apply their belief that men are usually taller than women. This set of alternative expert models has created the following amusing and ironic state of aVairs within the stereotypes literature. Specifically, people are accused of irrationality if they ignore their stereotypes when judging individual targets because they are supposed to use base rates

ACCURACY IN SOCIAL PERCEPTION

51

when judging others under uncertainty (e.g., Kahneman & Tversky, 1973; Locksley, Borgida, Brekke, & Hepburn, 1980); and more typically, they are accused of error and bias if they use their stereotypes when judging individual targets (e.g., American Psychological Association, 1991; Brewer, 1988a; Darley & Gross, 1983; Dunning & Sherman, 1997; Fiske, 1998; Fiske & Neuberg, 1990; Jones, 1986, 1990; Nelson, 2002). These and other discussions of the role of stereotypes in person perception assume that people are acting in an irrational, biased, or prejudicial manner if they perceive mean diVerences (on whatever attribute or behavior is being judged) between individuals belonging to diVerent social groups, holding individuating information constant (Bodenhausen, 1988; Borgida et al., 1995; Darley & Fazio, 1980; Hamilton et al., 1990; Miller & Turnbull, 1986). Relying exclusively on individuating information and ignoring stereotypes/base rates is often held up almost as a moral ideal to which people should strive. It cannot be normatively appropriate to both use and ignore prior expectations. Thus, the expert assumption that people should judge others solely on the basis of individuating information, which has emerged from much of the stereotype literature, conflicts with the claim that prior expectations and beliefs should influence judgments (in the absence of perfectly diagnostic individuating information) that have emerged from the cognitive judgment and decision-making literature. In general, explicitly stating one’s model of rationality is typical of the judgment and decision-making literature (e.g., Kahneman et al., 1982) but atypical of the stereotyping literature. The few exceptions in the stereotype literature (e.g., Krueger, Hasman, & Acevedo, 2003; Locksley et al., 1980; McCauley et al., 1980) use Bayes’s theorem as their model of rationality. Obviously, by adopting the same model that underlies much of the judgment and decision-making literature, these perspectives avoid the conflict between expert models concerning stereotypes. Apparently, however, there are social and political risks in scientists relying on Bayes’s theorem when attempting to understand accuracy, error, and bias in judgments of individuals. When McCauley et al. (1995) argued that a Bayesian analysis indicates that reliance on accurate beliefs about groups should increase, not decrease, the accuracy of person perception judgments in the absence of perfectly diagnostic individuating information, they were accused of ‘‘disagreeing with civil rights law’’ (Fiske, 1998, p. 381). Such an accusation confounds moral and legal issues with accuracy; if valid, represents another situation in which it may be illegal to be accurate; and highlights the role of politics in some criticisms of stereotype accuracy research. Regardless, discussions emphasizing the power of stereotypes to bias judgments (e.g., Aronson, 1999; Borgida et al., 1995; Darley & Fazio, 1980;

52

LEE JUSSIM

Devine, 1995; Dunning & Sherman, 1997; Fiske, 1998; Fiske & Neuberg, 1990; Fiske & Taylor, 1991; Jones, 1986, 1990; Jost & Banaji, 1994; Stangor, 1995; Stangor & McMillan, 1992) have rarely grappled with the implications of Bayes’s theorem and regression principles for understanding the role of stereotypes in enhancing accuracy in social perception. Considerably more conceptual clarity could be brought to research on stereotypes and person perception if researchers wishing to reach conclusions about error and bias explicitly stated and justified how they define rationality, both conceptually and operationally within their empirical research, and what type of pattern of person perception judgments would constitute reasonable, justified, appropriate, or accurate judgments under the circumstances being studied. In the absence of such explicit justifications, it will remain too easy for researchers to continue to interpret diametrically opposed judgments—those relying on stereotypes and those ignoring stereotypes—as both reflecting faulty and erroneous judgment processes.

G. AGREEMENT WITH THE TARGET Agreement with targets’ self-descriptions can and have been used as a criterion for assessing accuracy. Two broad types of targets’ self-descriptions are distinguished here. The term ‘‘self-reports of behavior’’ is used to refer to the actions targets claim they have engaged in. Examples might be how many glasses of alcohol they consumed yesterday, how much time they spend exercising each week, how often they argue with their spouse, and how much sleep they get each night. In contrast, the term ‘‘self-perceptions’’ is used to include targets’ attitudes, beliefs, feelings, and evaluations of themselves, their characteristics, and their accomplishments. Self-perceptions might include self-esteem, self-perceptions regarding personality traits (independent, assertive, extraverted, etc.), feelings, political positions, self-evaluations of academic or athletic performance, and so on. Selfperceptions, therefore, generally involve unobservable, underlying attributes of some type, whereas self-reports of behavior involve overt, observable actions. 1. Agreement with Targets’ Self-Reports of Behavior Targets’ self-reports of behavior can be used as a criterion against which to evaluate perceivers’ judgments. Social reality typically constrains bias (e.g., Jussim, 1991; Kunda, 1990), so that self-reports regarding specific and objective behaviors may be less likely to be tainted by self-serving,

ACCURACY IN SOCIAL PERCEPTION

53

defensive, or impression management biases than are self-reports regarding vague or ambiguous attributes. To get concrete, a fraternity member’s response to ‘‘How many alcoholic drinks have you had in the last week?’’ may be less likely to be biased than is his response to questions such as ‘‘How often did you get drunk last week?’’ Although bias may emerge regarding even the most objective of behaviors, there is a lot more room for interpretation in the ‘‘drunk’’ question (‘‘well, I did drink two sixes of beer, but I never really got drunk’’) than in the ‘‘drinks’’ question. Similarly, responses to ‘‘How much money did you donate to charity last month?’’ are more likely to be constrained by reality than are responses to ‘‘How generous are you?’’ Again, bias is always possible, but even a highly biased person is not likely to interpret buying a pizza as donating to charity, although people may vary a great deal on whether they consider sharing the pizza with a friend as a hallmark of great generosity. 2. Agreement with Targets’ Self-Perceptions Despite their imperfections, targets’ own self-perceptions regarding underlying or ambiguous attributes, such as personality characteristics or dispositions, can often be used as a criterion (see, e.g., Judd & Park, 1993; Judd et al., 1995; Ryan, 1995). As usual with probabilistic realism, the issue is not whether such a criterion is perfect, because no criteria are perfect. The question is whether the specific self-perceptions being used as a criterion are likely to reflect what the target is like. Indeed, there are theoretical reasons for expecting self-perceptions to be good criteria, at least sometimes. People have access to much more information about some of their experiences, inner states, relationships, and so on than do outsiders. Indeed, at least some research has shown that people’s memory for both randomly sampled and particularly striking thoughts and actions, up to 5 months after they occurred, is excellent (e.g., Barclay & DeCooke, 1988; Brewer, 1988b). Because bias can exist side by side with accuracy (e.g., Jussim, 1991), even when biases taint self-perceptions, they do not necessarily eliminate their validity (see, e.g., Block, 1965; Funder, 1995). If so, they can be used in the same manner as any imperfect criteria. Self-perceptions of academic ability, for example, usually correspond highly with indicators of academic achievement, such as grades and standardized test scores (e.g., Eccles & Wigfield, 1985; Felson, 1984). Attributes that are more readily observable, such as sociability or extraversion, tend to generate high agreement between self-reports and observers’ ratings (e.g., Funder, 1995; Kenny, 1994).

54

LEE JUSSIM

Furthermore, considerable validity evidence has been obtained regarding numerous self-report scales of personality characteristics and political attitudes (Robinson, Shaver, & Wrightsman, 1991, 1999). More highly validated self-perceptions are, obviously, better criteria than less well validated self-perceptions. If well-validated by prior research, however, self-report scales cannot (or at least should not) be summarily dismissed as ‘‘merely’’ self-report, at least not without refuting the preexisting validity evidence for the scale. In the absence of clear validity evidence, however, self-reports must be interpreted with caution and due regard to the topic. Self-reports of height, broccoli-eating, and interest in pornography fall on a continuum of least to most likely to be contaminated by social desirability biases. 3. Limitations to Agreement with Targets’ Self-Reports and Self-Perceptions Targets, of course, are imperfect themselves. Although they may have unique access to certain types of information (personal experiences, feelings, etc.), many people are subject to both motivated and unmotivated errors and biases (Kunda, 1990; Nisbett & Ross, 1980). Sometimes, people do not have access to their own cognitive processes or to some of their implicit beliefs and attitudes (Nisbett & Wilson, 1977; Wilson, Lindsey, & Schooler, 2000). Memory is imperfect and potentially subject to all sorts of biases (e.g., Koriat et al., 2000; Stangor & McMillan, 1992), so that self-reports may often be imperfect records of behavior. Researchers, however, have also developed a slew of methods for improving the accuracy of self-reports (e.g., event sampling, daily diary methods, etc.). Although even these types of methods may not completely eliminate error and bias, they have become widely used because of their demonstrated validity for many types of self-reports (e.g., Hedges, Jandorf, & Stone, 1985; Ra¨ikko¨nen, Matthews, Flory, Owens, & Gump, 1999). The problem of motivated biases may often (though not always) be greater when using self-perceptions (rather than self-reports of behavior) as a criterion, in part because of the common tendency for most people to view themselves in self-serving ways (Myers, 1999; Taylor & Brown, 1988). In addition, many people may either lie outright or slant their responses in such a manner as to present themselves in as socially desirable, moral, and competent a manner as possible (Paulhus, 1991, 2002). Although this means that the overall average level for some self-perception may often be too favorable, this may not prevent self-perceptions from being a good correlational criterion for assessing the accuracy of perceivers’ beliefs.

ACCURACY IN SOCIAL PERCEPTION

55

For example, Bertha may think she is a great athlete and Nyesha may think she is a good athlete. Both may be overestimates (Bertha may only be pretty good and Nyesha may be pretty average), but if their degree of selfinflation is similar, it may be true that Bertha is more athletic than Nyesha. So a coach who views Bertha as more athletic than Nyesha would be correct (and the coach’s views would correlate well with Bertha and Nyesha’s selfperceptions). Thus, even biased target self-perceptions may, under many circumstances, constitute good, if imperfect, criteria for assessing the accuracy of perceivers. In practice, although a default rejection of self-perceptions as criteria is not justified, researchers considering the use of particular selfperceptions need to thoughtfully consider their limitations and develop a convincing case that, in the context under study, they are likely to be good criteria.

H. CRITERIA: CONCLUSIONS Probabilistic realism occasionally provides a gold standard for establishing accuracy. A real estate salesman either did or did not sell the house. Except in rare cases, that is not a matter of opinion. The absence of such a gold standard, however, does not justify the conclusion that the criteria available to assess accuracy are so fundamentally flawed or their interpretation so ultimately ambiguous as to cloud the meaning of accuracy research. Establishing that a social belief or perception is accurate is much like establishing validity in social science research. The strongest and clearest evidence regarding accuracy comes from research that typically uses multiple measures and methods to establish the accuracy of social perception (e.g., Cronbach, 1955; Funder, 1987, 1995; Kenny, 1994). This does not mean that research examining accuracy using a single method or criterion is uninformative. Such research, however, may be less informative than research using multiple methods or criteria (except when that single criterion itself has previously been validated using multiple criteria!). Despite scientists’ inability to achieve absolute 100% certainty, probabilistic realism, truth with a small ‘‘t,’’ and the principles of construct validity provide clear ways of identifying criteria against which researchers can assess the accuracy of social perception. Once one has a good set of criteria, assessing accuracy would appear to be simple—just evaluate how well the perceptions correspond with the criteria. In fact, however, this turns out to be considerably more complex than it seems at first glance, and the next section explains why.

56

LEE JUSSIM

VIII. Must All Research on Accuracy Assess Components? A stopped clock is right twice each day. That does not make it a good clock. What does this have to do with social perception? More than it seems. Consider the following example: A wife successfully predicts that her husband will come home at least half an hour later than he said he would come home. It looks like this wife knows her husband pretty well, doesn’t it? Not necessarily. Maybe she always predicts that her husband will be late. Maybe she always predicts everyone will be late. It could even be worse than that: Maybe she always predicts all people will always do bad things and have bad attributes. Even though she might have happened to have been right that one time, she could not necessarily be considered a particularly astute judge of her husband. One could think of her prediction as stemming from several sources or components: her overall tendency to think well (or poorly) of people, her overall tendency to predict that people will be late (over and above her general tendency to think well/poorly of people), her overall tendency to think that her husband will be late (over and above her general tendency to think well of people and that people will be late), and her specific tendency to predict that her husband will come home later than he claimed this particular time (over and above her tendency to think well of people; to predict that they, in general, will be late; and to think her husband will run late). Each component of her prediction can be accurate to some degree, and each contributes to her overall likelihood of being accurate (across lots of judgments or predictions).

A. COMPONENTIAL APPROACHES TO THE STUDY OF ACCURACY This type of thinking inspired Cronbach’s (1955) classic review and at least two other more recent perspectives (Judd & Park, 1993; Kenny, 1994), all of which identified several processes contributing to social perception and that argued that accuracy needs to be separated into diVerent components reflecting these diVerent processes. Next, therefore, a brief overview of each of these three componential approaches to the study of accuracy is presented.

1. Cronbach’s Components Cronbach’s (1955) analysis involves computing an accuracy score for each perceiver judging a set of targets on a set of traits. Thus, there is a separate score for each perceiver. In this situation, both social judgment and social

ACCURACY IN SOCIAL PERCEPTION

57

reality can be divided according to eVects for targets and traits (Kenny, 1994): Social Judgment = ConstantðelevationÞ þ Target main effect (differential elevation) þ Trait main effect (stereotype accuracy) þ Target  Trait interaction ðdifferential accuracyÞ:

ð1Þ

The constant (which Cronbach called ‘‘elevation’’) is simply the grand mean of all judgments, the target main eVect (which Cronbach called ‘‘differential elevation’’) refers to mean diVerences in how each target is judged (over all traits), the trait main eVect (which Cronbach called ‘‘stereotype accuracy’’) refers to mean diVerences in how each trait is judged (over all targets),3 and the target  trait interaction (which Cronbach called ‘‘diVerential accuracy’’) refers to how each target is rated on each trait after subtracting out the other three eVects. Social judgments are then compared to criteria, which are broken down into the same sets of components. Accuracy, then, refers to how well each component of the judgment corresponds to the identical component on the criterion. If, for example, a student tends to see other people through rosecolored glasses, her elevation score will be higher than the elevation score on the criterion. If she sees Professor Smith as a better teacher than Professor Jones and the criteria also indicate that Professor Smith is a better teacher than Professor Jones, her diVerential elevation accuracy will be high (even if seeing the world through her rose-colored glasses leads her to overestimate both of them). If, when rating a group of professors, she sees absent-mindedness as more common than is organization, when in fact few professors are absent-minded but many are well-organized, her stereotype accuracy will be low. And if she rates Jones as more organized than Smith (after subtracting

3

Cronbach (1955) probably referred to the trait main eVect as ‘‘stereotype accuracy’’ because it refers to how perceivers seem to view ‘‘people in general’’ on this particular trait, which could be considered a type of stereotype. This was unfortunate, however, because Cronbach’s term is potentially confusing. It means something quite diVerent from what most people usually think of as social stereotypes regarding, for example, race, class, sex, and so on. ‘‘Trait accuracy’’ would, in my opinion, have been a clearer label. In this chapter therefore, I use the term ‘‘stereotype accuracy’’ to refer to two very diVerent phenomena, depending on the context. With one exception, I always use the term ‘‘stereotype accuracy’’ to refer to the validity of people’s beliefs about the attributes and behaviors of groups and their individual members. The one exception occurs in sections (such as this one) discussing Cronbach’s components. In the context of discussions of Cronbach’s components, the term ‘‘stereotype accuracy’’ refers to the validity of people’s perceptions of the level of a trait amont the particular targets being rated. This use of the term is explained in more detail in the text discussing the meaning of each of Cronbach’s components.

58

LEE JUSSIM

out elevation, diVerential elevation, and stereotype accuracy), and Jones really is more organized than Smith, her diVerential accuracy will be high [see Kenny (1994, pp. 117–121) for a fully explicated mathematical example of Cronbach’s components involving three perceivers, three targets, and three traits].

2. Kenny’s Social Relations Model The Social Relations Model (SRM) is a componential model that is related to, but diVers from, Cronbach’s (1955) in several important ways (see, e.g., Kenny, 1994; Kenny & Albright, 1987, for detailed discussions of similarities and diVerences). First, it is intended to be a broad and general model for assessing many diVerent aspects of social perception, of which accuracy is only one. Second, research using SRM typically focuses on perceptions regarding one trait at a time, rather than the multiplicity of traits addressed by Cronbach’s components. Third, however, it also typically focuses on several perceivers, rather than the one perceiver (at a time) that was the focus of Cronbach’s analysis. Thus, SRM research might perform one analysis to find out how accurately Dave, Charles, and Bella perceive each other’s intelligence and another to find out how accurately they perceive each other’s friendliness. Whereas Cronbach partitioned the judgment into target, trait, and target  trait components, SRM partitions judgment into target, perceiver, and target  perceiver components. Thus, the SRM equation is: Each Social Judgment ¼ Constant (elevation) þ Perceiver main effect (perceiver effect) þ Target main effect (target effect) þ Perceiver  Target interaction (relationship effect). ð2Þ The equation can be computed separately for each judgment provided by each perceiver, as shown by this example (from Kenny, 1994, p. 130): AI’s judgment of Bob ¼ Constant þ AI’s perceiver effect þ Bob’s target effect þ AI’s relationship with Bob.

ð2aÞ

The constant is the grand mean of all judgments, the perceiver main eVect is the mean diVerence between perceivers (over all targets), the target main eVect is the mean diVerence between targets (over all perceivers), and the perceiver  target interaction constitutes the unique manner in which a

ACCURACY IN SOCIAL PERCEPTION

59

particular perceiver judges a particular target after subtracting out the constant, the perceiver main eVect, and the target main eVect. In contrast to Cronbach’s (1955) components, however, SRM has a substantially diVerent equation for the criterion: Criterion/Social Reality/Behavior ¼ Constant þ Partner effect þ Actor effect þ Partner  Actor Interaction ð3Þ This equation can be computed separately for each actor (also from Kenny, 1994, p. 130): Bob’s behavior with AI ¼ Constant þ Bob’s Actor Effect þ AI’s Partner effect þ Bob’s relationship effect with AI ð3aÞ The constant is the mean level of the criterion; the actor eVect refers to mean diVerences between actors (averaging over all partners); the partner eVect refers to mean diVerences in how all actors, on average, behave with each partner; and the partner  actor interaction refers to the unique manner in which a particular actor behaves with a particular partner after subtracting out the constant, partner eVect, and actor eVect. To assess accuracy, each component of the judgment is compared to the corresponding component on the criterion. Overall, perceivers may see a high level of intelligence in their partners. If, on average, the actors really are pretty smart, elevation accuracy will be high (the constants will be similar). If, on average, AI sees his partners as much smarter than they really are, AI’s perceiver accuracy will not be high (comparison of AI’s perceiver eVect [Eq. 2a] to AI’s partner eVect [Eq. 3a]). If, on average, people see Bob as smarter than the others, and he really is, Bob’s target accuracy (called ‘‘generalized accuracy’’ in SRM) will be high (comparison of Bob’s target eVect, i.e., how he is usually perceived [Eq. 2a], with Bob’s actor eVect, i.e., what he is usually like [Eq. 3a]). If, on average, AI sees Bob as particularly smart (after subtracting out the constant, AI’s perceiver eVect, and Bob’s target eVect from the judgment) and Bob really does act particularly smart when interacting with AI (e.g., as might occur if AI’s perception of Bob is self-fulfilling), then AI’s dyadic accuracy will be high4 (comparison of the relationship eVects in Eqs. 2a and 3a).

4 Although I do not consider self-fulfilling beliefs to be accurate, they could be considered a form of accuracy within SRM.

60

LEE JUSSIM

3. Judd and Park’s Full Accuracy Design For Research On Stereotypes Judd and Park (1993) developed the first componential model focusing on explaining sources of accuracy and inaccuracy in social stereotypes. Their approach was primarily designed to analyze accuracy and bias in at least two groups’ perceptions of one another’s attributes. Thus, their analysis incorporates four main components of judgments regarding groups: elevation, perceiver group, target group, and attributes and all possible two-way and three-way combinations of these components: Judgment ¼ constant þ perceiver group effect(pge) þ target group effect(tge) þ attribute effect(ae) þ ðpge  tgeÞ þ ðpge  aeÞ þ ðtge  aeÞ þ ðpge  tge  aeÞ:

ð4Þ

Because Judd and Park’s (1993) componential model is not identical to those of Cronbach (1955) and Kenny (1994), they are discussed here. Because they are so similar, however, and because there are so many more terms than in Cronbach’s or Kenny’s approaches, only a brief simplification of their main ideas is presented. The constant is, once again, the mean level of judgments. The perceiver group eVect refers to mean diVerences in the judgments of diVerent groups of perceivers, the target group eVect refers to the mean diVerence in how diVerent groups of targets are judged, and the attribute eVect refers to mean diVerences between attribute classes. The various two- and three-way interaction terms represent unique combinations of perceiver group, target group, and attribute, over and above the constant and the three main eVects. These components are then compared to groups’ scores on the criterion. If one group of perceivers generally tends to overestimate targets’ score on the criterion, there will be some perceiver elevation inaccuracy. If one group of targets is consistently overestimated, there will be some target elevation inaccuracy. If one class of attributes is consistently overestimated, there will be an attribute elevation inaccuracy. And so on. Do people see other groups (in general) in a stereotypical manner? The attribute elevation eVect represents an overall tendency to over- or underestimate the prevalence of a particular type or class of attributes across the target groups. If attributes are chosen such that they are stereotypic for one group and counterstereotypic for the other, the attribute elevation eVect becomes a ‘‘stereotypicality’’ eVect—the tendency to view groups as more or less stereotypic than they appear on the criterion. A general tendency to overestimate stereotypical attributes and underestimate counterstereotypical

ACCURACY IN SOCIAL PERCEPTION

61

attributes represents a general tendency (across target groups) for the stereotype to exaggerate real diVerences. A general tendency to underestimate stereotypical attributes and overestimate counterstereotypical attributes represents a general tendency (across target groups) for the stereotype to underestimate real diVerences. For example, let’s say perceivers are asked to rate men and women on the traits aggressive, strong, and mathematical. High attribute elevation means that people perceive a greater diVerence between men and women, across these traits, than really exists. This approach can also be used to assess valence inaccuracies (is a group seen more positively or negatively than they deserve?) and dispersion inaccuracies (is a group seen as more or less variable than they really are?). Like the other componential approaches, Judd and Park’s (1993) full accuracy design was modeled after an analysis of variance—but with three ANOVA factors (perceiver group, target group, attributes and all twoway and three-way interactions) rather than the two of Cronbach and Kenny’s SRM. Although a full discussion of those factors is beyond the scope of this chapter, the three-way combination is particularly important to the study of stereotype inaccuracy because it tests for ingroup bias. The Subject Group  Target Group  Attribute factor tests whether stereotype exaggeration or underestimation (the attribute eVect) is more likely to occur when people from group A judge people from group B and when people from group B judge people from group A than when people from group A judge people from group A and people from group B judge people from group B. If, for example, stereotype exaggeration only occurs when people judge groups other than their own, one would have an ingroup bias eVect.

4. Components as a Means to Measure and Eliminate Noise In Judgments Componential approaches are particularly useful for assessing and eliminating noise (judgment factors irrelevant to accuracy) from judgments. For example, consider a case in which both judgment and criteria are measured on a subjective scale, as might be the case when judging personality traits and the criteria are target self-reports. There could be a discrepancy between judgment and criteria not because perceivers are inaccurate but because they use the judgment scale in a diVerent manner than do targets. Such subjective diVerences (or response biases) can be measured and eliminated when using a componential approach. Typically they manifest as one or more of the various elevation components in the diVerent models (see, e.g., Kenny, 1994,

62

LEE JUSSIM

for a detailed discussion of how componential approaches can assess and remove response biases).

5. Must Components Be Assessed In All Accuracy Research? Ever since Cronbach’s (1955) review, researchers have emphasized the importance of assessing components, sometimes going as far as to claim or convey that components must be assessed to address accuracy questions (see, e.g., Hastie & Rasinski, 1988; Jones, 1990; Judd & Park, 1993; Kenny & Albright, 1987). Whether all accuracy research must assess components, however, depends on the precise meaning of this claim. Next, therefore, moderate and strong forms of this claim are considered. a. The Moderate Form: Should Researchers Understand Existing Componential Approaches? The moderate form of this claim suggests that accuracy researchers should understand existing componential approaches to have better insights into the meaning of the results obtained from studies that do not explicitly assess components. This is undoubtedly true. It certainly behooves researchers interested in accuracy research to have more, rather than fewer, insights into the potential sources of social perception and the processes leading to accurate or inaccurate judgments and, especially, of the limitations and potential artifacts that influence whatever index of accuracy they do use. b. The Strong Form: Process Versus Accuracy (Again). The strong form of the claim is that all accuracy researchers must explicitly perform componential analyses because otherwise their research will be meaningless or uninterpretable. A variety of logical, theoretical, and empirical considerations indicate that this strong claim is overstated. First, componential approaches provide one class of explanations for how a person arrived at an accurate or inaccurate judgment. Indeed, Cronbach (1955) titled his article ‘‘Processes AVecting Scores on ‘Understanding of Others’ and ‘Assumed Similarity’’’ (emphasis mine). Why? Because components provide information about the processes of judgment. How do I arrive at my prediction that Mike Piazza will hit a grand slam? Do I always predict that he will hit home runs, or am I a particularly astute judge of Mike’s hitting? Why does Louise think African Americans are less likely to complete high school than they really are? Does she underestimate every group’s likelihood of completing high school? Does everyone, including African Americans, underestimate African Americans’ likelihood of completing high school? Or is Louise ethnocentric, underestimating only African Americans’ success and not her own group’s success at completing high school? Componential analyses provide answers to these types of questions

ACCURACY IN SOCIAL PERCEPTION

63

by addressing the processes by which people arrive at accurate or inaccurate social judgments. However, process is irrelevant with respect to establishing the degree of (in)accuracy of some perception. If I say ‘‘Mike is going to hit a home run’’ and he does, this particular prediction is right. End of discussion regarding my degree of accuracy. Although a single judgment is rarely psychologically interesting, the principle is just as true for multiple judgments. If I say that Mike Piazza is the best home-run-hitting catcher of the modern era and that he hits more home runs than any other catcher, again, I am right. With respect to understanding how I arrived at that prediction, it would be valuable to estimate my elevation, stereotype accuracy, diVerential elevation, and diVerential accuracy (if you like Cronbach); or, if you prefer SRM, my elevation, my perceiver eVect, Mike’s target eVect, and our interaction eVect. But if one wants to determine whether my prediction is accurate, the only thing we need to do is figure out whether he hit the ball over the outfield fence in fair territory (in the case of the single judgment), and whether he hit more home runs during his career than any other catcher (in the case of multiple judgments). If non-Armenians commonly believe that Armenians are public parasites burdening the financial community with their constant need for charity far more than other groups, and Armenians actually make fewer demands on public charity than other groups (LaPiere, 1936), then non-Armenians overestimate the financial burden created by Armenians. Again, period, the end, with respect to establishing the inaccuracy of this belief. Interpreting that inaccuracy is another matter. As Ryan, Park, and Judd (1996) have pointed out, in the absence of their full accuracy componential design, we cannot conclude, as did LaPiere (1936), that this means those holding nasty beliefs about Armenians are necessarily anti-Armenian bigots. Perhaps they overestimate every group’s need for charity, including that of their own. In that case, they are not ethnocentric at all. People with nasty beliefs about all groups including their own may be mean and evil, but they are not ethnocentric. Judd and Park’s (1993) full accuracy design would be extremely useful for providing some insight into why they overestimate Armenians’ request for charity, but it is completely irrelevant with respect to establishing whether anyone overestimates Armenians’ requests for charity. That question can only be answered by comparing estimates of Armenian need for charity to some criteria. c. No Reification of Components. It is extremely tempting to reify componential approaches to accuracy. First, they are statistically rigorous and complex, which gives them a particularly scientific aura. Second, they capture important, fundamental aspects of social perceptual judgment processes. Third, they successfully identify sources of bias or noise in judgments

64

LEE JUSSIM

that few researchers usually mean by accuracy. Thus, it is tempting to view components as concrete fixtures on the social perceptual landscape. If they are there, then should not accuracy researchers have to assess them? Such absolutist positions regarding components (‘‘Cronbach’s [or SRM] components must always be assessed’’ or ‘‘accuracy can only be viewed componentially,’’ etc.) are overstated for several reasons. First, there is no one right way to divide up components of social perception, as should be clear from this brief review of Cronbach’s, Kenny’s, and Judd and Park’s componential approaches. They have important similarities, but, obviously, there are also diVerences between all three. Such diVerences are made salient when the three approaches appear side-by-side, as they do in Table II. If there were any single ‘‘right’’ set of components that ‘‘must’’ be examined, if components were actually hard and fast fixtures in the social perception landscape, there could not possibly be three diVerent breakdowns of components unless one breakdown is ‘‘right’’ and the other two are ‘‘wrong’’ or unless each was woefully incomplete. If all were partially right, but incomplete in that they failed to address components identified by other researchers, then a full componential model would need to assess all the components identified by all models. Such a model is presented at the bottom of Table II. If components are ‘‘real’’ and ‘‘must’’ be assessed, then the only complete way to do it would be to assess the more than 50 components identified in this model. Such a model has never been recommended even by advocates of componential approaches and is not being recommended here. Indeed, it is so extreme as to border on absurd—but such an absurd model might be required if all components ‘‘must’’ be assessed. The situation, however, is far more complex than even this hypothetical combined componential model suggests. There is a potentially infinite number of ways in which social perception could be broken down into components (see also Kruglanski, 1989). Attributes could be further broken down into a variety of types or subclasses (e.g., positive vs. negative; explanations vs. descriptions vs. predictions; behaviors vs. traits, and so on). Similarly, both perceiver and target groups could be broken down not only by ingroup and outgroup but by any of the infinite ways of identifying groups (culture, demographic characteristics, memberships in organizations, professional expertise, etc.). This is not meant to suggest, however, that existing componential approaches are purely subjective and arbitrary and, therefore, can be ignored, but the choice of components will depend entirely on the types of processes one would be interested in studying and the types of response biases one would like to assess or eliminate. DiVerent componential breakdowns serve

ACCURACY IN SOCIAL PERCEPTION

65

TABLE II Componential Approaches to Social Judgment Cronbach’s (1955) Components: Judgment of a Person’s Trait ¼ Constant (elevation) þ Target main eVect (diVerential elevation) þ Trait main eVect (stereotype accuracy) þ Target  Trait interaction (diVerential accuracy). Social Relations Model Components (e.g., Kenny, 1994): Judgment of a Person’s Trait ¼ Constant (elevation) þ Perceiver main eVect (perceiver eVect) þ Target main eVect (target eVect) þ Perceiver  Target interaction (relationship eVect). Judd and Park’s (1993) Components for Research on Stereotype Accuracy: Judgment of a Group’s Traits ¼ Constant þ Perceiver Group eVect (pge) þ Target Group eVect (tge) þ Attribute (stereotypic vs. counter-stereotypic) eVect (ae) þ (pge  tge) þ (pge  ae) þ (tge  ae) þ (pge  tge  ae). Hypothetical Componential Approach Combining Cronbach’s, Kenny’s, and Judd and Park’s Approaches: Judgement of a Person’s Trait ¼ Constant þ Perceiver Group eVect þTarget Group eVect þ attribute (stereotypic vs. counter-stereotypic) main eVect þ Target main eVect þ Perceiver main eVect þ individual trait eVect þ 15 two-way interactions þ all 20 three-way interactions þ all 11 four-way interactions þ all 4 five-way interactions þ the six-way interactiona a

This model assumes all main eVect terms are independent, which may not be the case. For example, the target group main eVect may not be independent of the target main eVect and the attribute main eVect may not be independent of the individual trait eVect. In such a situation, there might be fewer interactions than displayed in this model. There would, however, still be literally dozens of such interactions. No researcher has ever advocated this model, including me.

diVerent purposes and provide insights into diVerent aspects and processes of social perception. Thus, understanding existing componential approaches would seem crucial to anyone studying accuracy to gain insights into how best to interpret their own or anyone else’s data addressing the degree of (in)accuracy in social perception.

66

LEE JUSSIM

d. Componential Models May Be Most Important When the Criteria are Self-Reports and Self-Perceptions. Although Cronbach’s componential approach never generated much empirical research, Kenny’s and Judd and Park’s have, and much of that research has used target individuals’ or target groups’ self-perceptions as criteria (see, e.g., reviews by Kenny, 1994; Ryan, 1995). When the criteria are self-perceptions, especially self-perceptions regarding traits, attitudes, or dispositions, rather than behaviors or other objective characteristics, using a componential analysis is highly likely to be the best (clearest, most informative) method of analysis. Self-perceptions of traits, for example, typically have no objective referent. How extraverted is someone who rates themselves ‘‘5’’ on a 1–7 scale with endpoints labeled ‘‘not at all’’ and ‘‘very’’? It is hard to say because each choice is subjective, in that each rater imputes their own meaning to each scale point (e.g., Biernat, 1995). Such diVerences in subjective meanings cloud the assessment of accuracy. Componential approaches, however, are particularly good at identifying diVerences in subjective meaning and removing them from estimates of accuracy (this is often captured by the various elevation components). Thus, it is probably a good idea to use a componential approach, if possible, almost any time one uses self-perceptions as criteria.

B. NONCOMPONENTIAL APPROACHES TO ASSESSING ACCURACY It has just been argued that componential processes are not necessary for the assessment of accuracy. This argument, however, does not rest solely on a critical evaluation of the claim that ‘‘all accuracy research must perform a complex componential analysis.’’ Instead, much of the best evidence for the idea that componential analyses are not strictly necessary comes from the many noncomponential approaches to the study of accuracy that have made important and enduring contributions to understanding social perception. The next section, therefore, briefly reviews three of the most influential noncomponential approaches. The term ‘‘noncomponential’’ here is potentially misleading because it unfortunately implies that one can completely ignore the components issue. Even ‘‘noncomponential’’ approaches can themselves be considered to assess subsets of components in the various componential models, as shall be made explicit in the discussion that follows. The term ‘‘noncomponential’’ is used, therefore, to refer to approaches that assess accuracy without an explicit and intentional assessment of components. When, from a componential standpoint, such approaches only assess a subset of components in one or more of the componential models, this is pointed out explicitly below. What

ACCURACY IN SOCIAL PERCEPTION

67

may be lost by not performing a full componential analysis is also explicitly discussed. 1. Correlational Approaches Most noncomponential approaches to assessing accuracy or processes underlying accurate and inaccurate social perceptions use correlations (typically Pearson’s, occasionally regression coeYcients) to assess the extent to which judgments correspond to criteria. In general, when judgments concern a single attribute, correlations between judgments and criteria capture Cronbach’s (1955, p. 191) diVerential accuracy correlation, which he described as ‘‘sensitivity to individual diVerences. . . . These are the only processes included in present measures of social perception which depend on J’s [perceivers’] sensitivity to the particular O [target].’’5 The simplest and most typical form of correlation in accuracy research is that between a set of perceivers’ judgments or predictions regarding a single trait or attribute of a set of targets and a criterion measure of that attribute or trait. For example, teachers’ predictions regarding students’ achievement may be compared with grades or standardized test scores; interviewers may evaluate a set of interviewees’ skills, which are then compared with work samples; or perceivers may estimate the percentage of people belonging to various demographic groups that complete college, which are then compared to Census figures. Such correlations automatically remove the elevation and stereotype accuracy components from correspondence between perceivers’ judgments and the criterion. (This is because these correlations reduce data to deviations from the mean.) Thus, a simple correlation (between judgment and criterion) goes a long way toward eliminating many of the biases, artifacts, and problems in assessing accuracy first identified by Cronbach. Of course, the correlation coeYcient is not perfect. First, it removes or avoids but does not directly assess Cronbach’s (1955) elevation and stereotype accuracy components. Because correlations remove average diVerences between judgments and criterion, they cannot assess any consistent tendency to over- or underestimate targets (elevation, a` la Cronbach). If it were important to assess those components to address some research question, one could not use the correlation to do so.

5

Cronbach was, in addition to diVerential accuracy correlation, referring to diVerential elevation correlation in this quotation, which refers to the correlation between the perceivers’ judgment of a target averaging over all attributes and the targets’ average score on the criteria representing those attributes. This, however, is irrelevant when there is only a single attribute being judged.

68

LEE JUSSIM

Second, correlations equate the variability of judgments and criterion. Therefore, they cannot assess whether perceivers consistently over- or underestimate target variability. Because mean and variability diVerences between judgments and criteria probably often reflect response bias or scaling discrepancies between perceiver and criterion, these limitations to correlations do not greatly undermine their utility in assessing accuracy. The term ‘‘scaling discrepancies’’ is used here to refer to the idea that people may use scale points in a manner diVerently than used in the criterion. This would obviously be true if, for example, judgment and criterion are assessed in diVerent metrics (e.g., subjective rating scale and percentages, respectively). People, however, still might use the numbers in some scale diVerently than is used for the criterion, even if they are supposedly on the same scale. For example, let’s say Alfred estimates three people’s IQ scores as 40, 50, and 60, when they are really 115, 120, and 125. Although it is possible that Alfred believes all three of these fairly intelligent people are neurologically impaired, it is more likely that Alfred does not fully understand how IQ scores are scaled. He dramatically underestimates people’s IQ in absolute terms, but his estimates are also overly sensitive to actual variations in IQ (Alfred’s judgments go up 10 IQ points for every 5-point increase in actual IQ). Given his subjective IQ scale, however, the correlation between his judgments and actual IQ would be perfect (1.0), because mean diVerences in judgment criteria are irrelevant to computation of the correlation, the correlation coeYcient is computed after statistically equating the variability in judgment and criterion, and his judgments move in (diVerently scaled) lockstep with targets’ actual IQ. Thus, the correlation coeYcient would yield a conclusion that Alfred is an excellent judge of people’s intelligence. Is the conclusion justified? As long as one keeps in mind that what this really means is that ‘‘Alfred is very good at detecting diVerences in people’s intelligence, but this does not tell us anything about either how Alfred uses the IQ scale or about whether he consistently over- or underestimates people’s intelligence,’’ it is perfectly justified.6 2. Construct Validity and Correlational Approaches To Accuracy The section on criteria argued that assessing accuracy was much like assessing the validity of many social science constructs. This is also important here, because the correlation coeYcient is so frequently used to 6 In empirical studies, one also needs to be concerned with the level of analysis and nonindependence issues that arise when, for example, each perceiver judges several targets. This, however, is a design and statistical issue that is beyond the scope of this chapter (but see Kenny, Kashy, & Bolger, 1998, for a detailed discussion of this issue).

ACCURACY IN SOCIAL PERCEPTION

69

establish the validity of some measure that it is often referred to as the ‘‘validity’’ or ‘‘validity coeYcient’’ of some measure (e.g., Campbell & Stanley, 1963; Cook & Campbell, 1979; Dawes, 1979). In much the same manner, correlations can be used to assess the accuracy of social perception. Establishing accuracy, however, is somewhat more complex than establishing construct validity because social perception, judgment, and expectations are themselves constructs. Thus, all the issues involved in establishing construct validity arise not just when measuring targets’ attributes but when measuring perceivers’ expectations, perceptions, judgments, and beliefs about others. Accuracy, therefore, is not usually best reflected by correlations between observable measures (e.g., a measure of perceiver expectations and a measure of target extravertedness). Accuracy will often best be reflected by correlations between the underlying constructs representing the social perception and the criteria (this can often be accomplished either by disattenuating correlations for unreliability or by using LISREL-type models; see, e.g., Bollen, 1989; Carmines & Zeller, 1979). This should not, however, be misinterpreted to mean that all accuracy research must necessarily assess relationships between underlying constructs rather than observed measures. Sometimes it may just not be possible to do so, as, for example, when there is only a single item assessing either the perception or the criterion. This, of course, does not constitute any sort of immovable obstacle to or fatal flaw in accuracy research. Correlations among such measures, however, will tend to underestimate accuracy to the extent that the observed measures only imperfectly reflect the underlying attributes or expectations. Therefore, people may be more accurate than indicated by research that only assesses correlations between observed measures of expectations and observed measures of criteria. This brief delving into construct validity, correlations, and unmeasured variables was necessary to lay the foundation for understanding three of the main noncomponential approaches for assessing accuracy in social perception. All three are fundamentally based on the correlation between perception and criteria. 3. Brunswik’s Lens Model Brunswik (1952) suggested that accurate perception of reality (both object and social) involves the use of cues probabilistically related to objective reality. He metaphorically called his approach the Lens Model to capture the idea that objective reality is never observed directly. Instead, cues related to objective reality must be observed and interpreted as relevant to some judgment; that is, objective cues are seen through the ‘‘lens’’ of subjective perception and judgment.

70

LEE JUSSIM

Fig. 1. Brunswik’s Lens Model. Single-headed arrows represent causes. The double-headed arrow represents the correlation. This figure depicts a general example of how the Lens Model might be used. It can be used for assessment of accuracy in perception of physical attributes (e.g., size, distance) as well as of psychological attributes. Furthermore, the four cues shown here are simple concrete examples of potential manifestations of some attribute; the Lens Model is not restricted to these types of cues. In Lens Model research, what is called accuracy here is usually called ‘‘achievement.’’

This does not necessarily mean that perception is a purely subjective phenomenon unrelated to objective reality. Indeed, one of the main purposes of the Lens Model is to provide a mechanism not only for assessing people’s degree of accuracy but also for understanding sources of both accuracy and inaccuracy in their judgments. The Lens Model has been used to study accuracy, inaccuracy, and bias in a wide variety of business, educational, criminal, and interpersonal contexts (e.g., Bernieri & Gillis, 2001; Craik & Appleyard, 1980; Reynolds & GiVord, 2001; Schmitt, Noe, & Gottschalk, 1986). Figure 1 presents a simplified but general version of the Lens Model. It captures several main ideas. First, the circled ‘‘Psychological Attribute’’ represents some sort of psychological construct that cannot be directly observed (self-esteem, extraversion, intelligence, etc.). The Cues, shown in the middle of the figure, are directly observable or measurable phenomena. The arrows pointing from the Psychological Attribute to the Cues are labeled ‘‘Validity,’’ because they represent the extent to which the underlying attribute manifests itself in the observable Cues. The rightmost circle represents perceivers’ judgments (or perceptions). The arrows going from the Cues to Judgments represent the extent to which the observable cues influence perceivers’ judgments (labeled Cue Utilization). The Lens Model characterizes social perception as a two-step process: observable manifestation of psychological attributes and perceiver use of

ACCURACY IN SOCIAL PERCEPTION

71

observable cues to arrive at judgments. Accuracy, therefore, is captured by the correlation between the psychological attribute and the judgments (the long, double-headed arrow in Fig. 1). Correlations assess how well the judgments correspond to the attributes—that is, accuracy. The Lens Model is a noncomponential, correlational model for assessing both degree of accuracy and processes of social perception. Identifying cue validity and cue utilization focuses on a very diVerent set of the processes than is typically the focus in componential models. As such, it provides diVerent (not better or worse) types of insights into processes of social perception than do componential models. 4. The Realistic Accuracy Model Funder’s (1995, 1999) Realistic Accuracy Model (RAM) draws on the set of fundamental assumptions described under ‘‘probabilistic realism’’ to create a model that could be viewed as an extension and elaboration of Brunswik’s Lens Model. Some of the main ideas of RAM are depicted in Fig. 2. As with the Lens Model, overall accuracy is typically assessed by the correlation of the underlying attribute with the perceiver’s judgment (represented by the large, curved, double arrow on the bottom of Fig. 2). Four steps needed for perceivers to arrive at an accurate judgment are displayed in between the underlying attribute and the judgment. First, the underlying attribute needs to create some sort of observable evidence relevant to that attribute (the cues, in the Lens Model). Dishonesty, for example, is not likely to be displayed in a large lecture hall (except perhaps during test time). Interest in the class is more likely than honesty

Fig. 2. Funder’s Realistic Accuracy Model. The single-headed arrows here represent steps in a sequence, not causal eVects. The double-headed arrow represents the correlation between targets’ attributes and perceivers’ judgments.

72

LEE JUSSIM

to be displayed, for example, through high attendance levels, keeping up with class assignments, or class participation. Second, the evidence needs to be at least hypothetically available to the perceiver. Whether or not a student has completed all assigned readings may rarely be directly observable to the class’s teacher. Attendance and participation, however, would be considerably more available. Third, the perceiver has to detect that evidence. In a large lecture hall, which students participate is pretty obvious. However, detecting precisely which students do and do not attend regularly, out of a swarming mass of hundreds of students in the lecture hall, is considerably more diYcult. Fourth, the perceiver has to actually use the detected evidence/cues (and weight them appropriately) for arriving at an accurate judgment. If lecture hall teachers cannot remember which students regularly participate, it would be diYcult to use participation as a basis for judgment regarding interest in the class. Although this is the heart of Funder’s (1995) RAM, applying the model might be considerably more complex. People have many underlying characteristics. Funder (1995) focused primarily on personality traits, but RAM seems applicable to all sorts of unobservable personal characteristics, including, for example, emotions, attitudes, motivation, and so on. Furthermore, one attribute may create many cues (as is most obviously captured in the Lens Model), and some cues may reflect multiple attributes. Thus, one type of complexity involves the sheer number of possible interrelationships among attributes, cues, and judgments. Like several of the componential models, RAM also considers how the perceiver, target, attribute, and evidence relate to accuracy (Funder [1995] referred to these as Judge, Target, Trait, and Information, respectively, but I am sticking with the terminology I have used throughout this chapter). This is not to identify components, however. Instead, the purpose is to analytically identify how specific combinations of perceiver, target, attribute, and evidence might combine to influence accuracy. For example, some perceivers may be particularly good (poor) at evaluating certain types of traits (e.g., clinical psychologists might be better than most of the rest of us at evaluating others’ mental health). Some perceivers might be particularly good (poor) at judging particular targets (e.g., close friends might be better judges of each other than are strangers). Some perceivers might be especially good (poor) at using or obtaining certain types of evidence (e.g., some people may be better than others at picking up on targets’ emotion-revealing nonverbal cues). Funder (1995, 1999) has discussed at considerable length how the various combinations of perceiver, target, attributes, and evidence may combine to influence accuracy. Like the Lens Model, RAM assumes that relationships between underlying attributes, cues, and judgments are probabilistic. Like the Lens Model,

ACCURACY IN SOCIAL PERCEPTION

73

overall accuracy is typically assessed via correlations, although discrepancy scores (between judgment and criterion) can be used, too (see Funder, 1987, 1995). RAM is particularly good at explaining why accuracy in person perception may often be low. For the judgment to closely correspond to the criterion, that criterion needs to clearly manifest itself in ways that could be and, in fact, are detected by the perceiver, and then the perceiver needs to use that detected information (as well as not use information that is not relevant to the judgment). A breakdown at any step will dramatically undermine accuracy. Furthermore, by focusing on combinations of perceiver, target, attributes, and evidence, RAM is also particularly good at highlighting processes that may enhance or undermine accuracy. 5. Dawes’s Improper Linear Models Dawes (1979) made a very interesting discovery. In reviewing his own and others’ research on decision-making, he found that people tend to be very good at identifying the evidence or cues that are relevant to making some prediction but are not very good at combining or integrating those cues. Thus, their overall predictive accuracy tends to be quite low. Note, however, that this is not because people are completely inept. They are good at one part of the prediction task (identifying criteria for making a prediction) but poor at another part (putting those criteria together). One of Dawes’s (1979) examples involved admissions to graduate school in psychology. The criteria typically used for making admissions decisions seem appropriate: GRE scores (general intellectual ability), GPA (achievement at academic tasks over an extended period), and letters of recommendation (what experts in the field who are highly familiar with the applicant have to say about him or her). Nonetheless, Dawes (1979) found that the correlation of graduate admissions committee evaluations with later success in graduate school is typically quite low (.19). If people were completely inept, they would not even use appropriate criteria; that is, the criteria they do use would not predict success in graduate school. However, if they are good at identifying the appropriate criteria, but use them poorly, then the raw criteria themselves should do a much better job at predicting graduate success. This was indeed the case—the overall (multiple) correlation of the criteria themselves with graduate success was about .4. What to do? It is unreasonable to expect admissions committees to compute complex statistical formulas in their heads or to create a formal statistical score for each applicant. Dawes provided an elegantly simple and even amusing answer. Identify the criteria, weight them all equally, and add. For example, GRE, GPA, and letters of recommendation might each be transformed into standardized scores (obviously, letters would need to first be placed on some sort of rating scale). Priscilla, with good GREs, a high

74

LEE JUSSIM

GPA, and excellent letters of recommendation might receive z-scores of 1, 2, and 2, respectively, for a total score of 5. George, with high GREs, a good GPA, and good letters, might receive z-scores of 2, 1, and 1, for a total of 4. Priscilla would be ranked higher than George. This is diVerent from a formal statistical model primarily because the weights for each predictor have been chosen in a less than optimal manner (many statistical prediction techniques, such as regression, identify how to weight the criteria in such a manner as to maximize their overall predictive validity). Dawes’s (1979) analysis showed, however, that this ‘‘suboptimal’’ weighting of criteria predicts outcomes nearly as well as do formal statistical models. In the graduate admissions example, Dawes’ improper linear model correlated .38 with future success in graduate school. Dawes (1979) went on to show that a simple, improper linear model performed similarly well in predicting all sorts of outcomes, including choice of bullet-type for a police department and a bank’s predictions regarding companies likely to go bankrupt.7 Dawes’ improper linear model is fundamentally diVerent than the Lens Model and RAM. The Lens Model and RAM were specifically designed to assess degree of accuracy and processes underlying social perception. That is, they were meant to describe aspects of the social perception process. In contrast, Dawes’ model is primarily prescriptive (it suggests how people should go about making decisions and arriving at predictions). Nonetheless, it has been included here for two reasons. First, Dawes’s (1979) conclusion that people are good at selecting criteria, but not good at using them, is descriptive. In RAM terms, it indicates that people often are good at detecting available and relevant cues, but that they often do not use them well (in Lens Model terms, their cue utilization would be poor). Second, although Dawes did document that people were, on their own, not very good at arriving at accurate predictions, he also showed that the accuracy of their predictions could easily be improved. Identify the criteria, weight them equally, then add!

C. COMPONENTIAL AND NONCOMPONENTIAL APPROACHES TO THE STUDY OF ACCURACY Componential approaches to the study of accuracy have provided several major contributions to the study of social perception. As a research tool for assessing the processes of social perception, they demonstrate again how

7 Dawes’s results were actually even more striking. The equal weight improper linear model generally (i.e., not just in the graduate admissions example) predicted outcomes better than the split sample cross-validated regression weights.

ACCURACY IN SOCIAL PERCEPTION

75

accuracy research typically provides the type of double contribution that few, if any, experimental studies focusing exclusively on error, bias, and process can provide. Componential approaches address accuracy, inaccuracy, and process. They are particularly useful at identifying some of the reasons a perceiver is a particularly good or poor social judge and also can demonstrate that some perceiver might be accurate in some ways, yet inaccurate in others, even when judging the same target. Last, they are especially good at measuring and eliminating response biases that are largely or completely irrelevant to the assessment of accuracy. It is also clear, however, that componential approaches have not been the only way to productively study accuracy. Three noncomponential approaches, Brunswik’s (1952) Lens Model, Funder’s (1995) RAM, and Dawes’s (1979) improper linear models were also discussed. All rely on the correlation as the primary means for establishing correspondence between social judgment and criteria, and all have provided important insights into both the (in)accuracy of social perception and basic social perceptual process. Other approaches relying on Bayes’s theorem (e.g., Krueger, 1996; Krueger et al., 2003; McCauley & Stitt, 1978; McCauley et al., 1980) and structural equation modeling (e.g., Jussim, 1991; West & Anderson, 1976; Williams, 1976) have also contributed to understanding processes leading to accurate and inaccurate social perceptions, especially with regard to interpersonal expectations and social stereotypes. It is clear, therefore, that although componential approaches are valuable and important, one need not need perform any particular componential analysis to assess accuracy. Indeed, in the absence of data permitting a full componential analysis, performing a simple correlation between belief and criterion, though imperfect, would typically provide far more information than would doing nothing. Even in such a situation, however, it behooves the researcher performing the correlation to understand componential approaches to appropriately interpret just what that correlation does and does not show.

IX. Conclusions This review has documented many of the conceptual, methodological, and statistical advances in the psychological study of social perceptual accuracy, with a particular focus on addressing many of the criticisms of accuracy research that have been raised over the last 50 years. In doing so, the conclusion has been repeatedly reached that, although there are genuine diYculties, ambiguities, or obstacles to accuracy research, they are no greater than in most other areas of social psychological research. Indeed, this

76

LEE JUSSIM

review has shown that the conceptual, theoretical, methodological, and empirical bases of accuracy research are identical to those of much other research in psychology. By taking stock of the many conceptual and methodological advances in accuracy research that have been made over the last 20 years, this review has attempted to provide a road map for further accuracy research. Although it is impossible for such a map to represent every twist and turn in the road, it has attempted to provide some valuable signposts providing direction in the pursuit of accuracy research. These signposts show how to circumvent or avoid many known obstacles, indicate that the road is clear in many directions that were once believed to be filled with obstacles, and have identified many alternative routes available when known obstacles arise.

A. NONE OF THE COMMON CRITICISMS OF ACCURACY RESEARCH THREATENS ITS VIABILITY This review has contested the viability of some of the common theoretical, conceptual, and methodological criticisms of accuracy research. Some criticisms were seen to be manifestly false; others were seen to be logically incoherent. Yet others were seen to be partially valid but not to constitute serious threats to accuracy research. This review also showed that even completely valid criticisms of accuracy research could readily be addressed by a clear understanding of the nature of, and limitations to, the insights into social perception that accuracy research can provide.

B. THE CRITERIA FOR ACCURACY ARE IDENTICAL TO CRITERIA USED IN PSYCHOLOGICAL, SOCIAL, AND BEHAVIORAL SCIENCE RESEARCH The crucial issue of identifying criteria for establishing accuracy was also explored. This review contested arguments criticizing accuracy research for alleged diYculties in identifying appropriate criteria for validating social beliefs. Such criticism would seem to logically require a similar criticism of nearly all research in psychology (with the possible exception of biological research). The logic of establishing accuracy of social perception overlaps almost completely with the (far less controversial) logic of establishing construct validity in the social sciences.

ACCURACY IN SOCIAL PERCEPTION

77

C. BOTH COMPONENTIAL AND NONCOMPONENTIAL APPROACHES TO ACCURACY ARE VALUABLE Furthermore, several major approaches to the assessment of accuracy and processes leading to accurate and inaccurate social judgments were reviewed. This included reviews of the three main componential approaches to accuracy and three of the major noncomponential approaches to the study of accuracy. It was concluded that both componential and noncomponential approaches have made major and significant contributions to understanding social perception and accuracy.

D. PROCESS RESEARCH ON ERRORS, FLAWS, AND BIASES IS NOT MUTUALLY EXCLUSIVE WITH ACCURACY Despite the often-demonstrated existence of a slew of logical flaws and systematic biases in lay judgment and social perception, such as the fundamental attribution error, false consensus, over-reliance on imperfect heuristics, self-serving biases, and so forth, people’s perceptions of one another are surprisingly (though rarely perfectly) accurate. Despite being most wellknown for their self-fulfilling eVects, teacher expectations, for example, typically predict student achievement more because they are accurate than because of expectancy eVects (e.g., Brophy & Good, 1974; Brophy, 1983; Eccles & Wigfield, 1985; Jussim, 1991; West & Anderson, 1976). People are also often surprisingly good at judging one another’s personalities (e.g., Funder, 1987, 1999; Funder & West, 1993; Paunonen, 1991). People in close relationships are often better than strangers at inferring each other’s thoughts and feelings (Ickes, 1993, 1997; Stinson & Ickes, 1992). Indeed, people even show some accuracy under very poor or minimal conditions in arriving at initial impressions among strangers (Kenny, 1994), for example, and in predictions based on very small samples of behavior (Ambady & Rosenthal, 1992).

E. BELIEFS ABOUT SOCIAL GROUPS (STEREOTYPES) OFTEN SHOW MODERATE TO HIGH ACCURACY Perhaps most surprisingly, considerable research has accumulated showing that people’s perceptions of groups and individual targets from diVering race, sex, and social class backgrounds are often quite (though

78

LEE JUSSIM

not perfectly) accurate (Diekman et al., 2002; Hall & Carter, 1999; Judd et al., 1995; Jussim et al., 1996; Lee et al., 1995; Krueger, 1996; Madon et al., 1998; McCauley & Stitt, 1978; Ryan, 1996, 2002; Ryan & Bogart, 2001; Swim, 1994; Williams, 1976). Because, however, accuracy is not mutually exclusive with, and sometimes is even enhanced by, bias, this claim does not refute or deny the existence of prejudice or discrimination. Indeed, many of the same studies demonstrating moderate to high accuracy have also demonstrated that stereotypes do sometimes lead to systematic biases in judgments of groups and individuals.

F. UNDERSTANDING ACCURACY IS ESSENTIAL FOR UNDERSTANDING AND ALLEVIATING SOME SOCIAL PROBLEMS Research on accuracy would appear essential in the application of psychology to the remediation of many social problems, especially those involving intergroup relations. Beliefs about groups have for decades been presented as either a cause of social inequalities or as a reflection of prejudice and existing dominance hierarchies, or both (e.g., Allport, 1958; American Psychological Association, 1991; Fiske & Taylor, 1991; Jost & Banaji, 1994; LaPiere, 1936; Pickering, 2001; Sidanius & Pratto, 1999). If such beliefs are largely erroneous, and if they cause injustices, as many such perspectives imply, then the solution to many group-based social problems should focus almost entirely on correcting the beliefs and perceptions of those holding such beliefs. If the problem is erroneous perception, then correcting the perceptions should eliminate, or at least greatly alleviate, that problem. To the extent, however, that some beliefs about group diVerences are accurate because groups really do diVer on some societally valued attributes, then correcting erroneous perceptions would be a misguided endeavor, or at best a woefully incomplete solution. In this situation, eVorts need to be focused on redressing the underlying inequalities, which will probably involve devoting resources to improving the station of lower status or unjustifiably stigmatized groups. Accuracy research, therefore, would appear to be crucial with respect to identifying how to devote resources to the alleviation of social problems—with relatively more resources being directed to correcting perceiver beliefs when those are shown to be inaccurate but relatively more resources being directed to improving the station of low status groups when perceiver beliefs are shown to be accurate.

ACCURACY IN SOCIAL PERCEPTION

79

G. AN INVITATION TO STUDY ACCURACY There has been a great blossoming of accuracy research over the last 20 years (Ambady & Rosenthal, 1992; Funder, 1987, 1995, 1999; Ickes, 1993, 1997; Kenny, 1994; Kenny & Albright, 1987; Judd & Park, 1993; Jussim, 1989, 1991; Jussim et al., 1996; Lee et al., 1995; Oakes et al., 1994; Ryan, 2002; Swim, 1994). Now that the major obstacles to accuracy research have been cleared away and that new and innovative ways to address accuracy are being continuously developed, if anything, the pace of research on accuracy should accelerate over the next 20 years. Nonetheless, specific future developments are diYcult to anticipate. Twenty years ago, I would never have predicted that people could be as accurate as they apparently are on the basis of minimal information (Ambady & Rosenthal, 1992) or that stereotypes would show as much accuracy as they consistently seem to show (Brodt & Ross, 1998; Hall & Carter, 1999; Lee et al., 1995; McCauley & Stitt, 1978; McCauley, 1995; Ryan, 2002; Swim, 1994; Wolsko et al., 2000). Nonetheless, many classic issues and recent developments in social psychology naturally raise a host of interesting and important questions about accuracy. 1. Moderators: The Person, the Situation, and the Relationship Social and personality psychology have a long tradition of examining the role of individual diVerences, situational forces, or both in causing or moderating all sorts of social phenomena. This tradition naturally lends itself to the study of accuracy. Hall and Carter (1999), for example, examined whether a slew of individual diVerence factors increased or reduced the accuracy of people’s beliefs about diVerences between men and women (the most successful moderator was nonverbal sensitivity). With the modern methods that have cleared away the old obstacles, we may finally be able to obtain some answers to the question that dominated early accuracy research and seemed to have no answer: ‘‘Who is a more accurate judge of others?’’ Similarly, researchers have also begun addressing the other classic social psychological issue—situational moderators of accuracy. Wolsko et al. (1999), for example, found that White students, given a ‘‘multicultural’’ set (one emphasizing the importance of understanding and valuing bona fide racial and cultural diVerences), held more accurate racial stereotypes than those given a ‘‘color-blind’’ set (one emphasizing the sameness of all people and deemphasizing the importance of group diVerences). Those given

80

LEE JUSSIM

the multicultural set also used their stereotypes more appropriately when judging individual targets. Thus, situations evoking diVerent sociopolitical mindsets moderated stereotype accuracy. Others areas ripe for an investigation of situational moderators of accuracy include interpersonal expectations in general and teacher expectations in particular, managerial perceptions of employees, and first impressions. Accuracy may also vary by aspects of relationships. Length of relationship seems like an obvious moderator (the longer people get to know one another, the more accurately they should see one another). Although research on stereotypes is broadly consistent with this hypothesis (especially the overwhelming evidence that people rely less on stereotypes the more individuating information they have; Eagly, Makhijani, Ashmore, & Longo, 1991; Kunda & Thagard, 1996), research using the SRM has not always found that accuracy increases with length of relationship (Kenny, 1994). Whether accuracy also systematically varies by type of relationship (e.g., romantic, parent–child, occupational, etc.) is currently unknown. 2. Accuracy and Bias and Their Integration Whether lay social beliefs and perceptions are typically rational and accurate or inaccurate and biased has been controversial for decades. Clearly, however, those emphasizing error, bias, and the ways in which social beliefs create social reality have dominated the literature on social cognition (e.g., Fiske, 1998; Jones, 1986; Kahneman & Tversky, 1973; Nisbett & Ross, 1980; Snyder, 1984). These views have created an image of a social perceiver whose misbegotten beliefs and flawed processes construct not only illusions of social reality in the perceiver’s own mind but actual social reality through processes such as self-fulfilling prophecies. In this bleak view, the mind becomes primarily a product of cognitive shortcomings and distorted social interactions. Implicit in the prevalence of testaments to the power of error and bias is that something important has been discovered. Because error, bias, and accuracy can coexist simultaneously, however, merely demonstrating error or bias does not demonstrate inaccuracy. Although people undoubtedly commit errors and biases and are rarely perfectly accurate, almost none of the literature routinely cited as testaments to the power and prevalence of error and bias actually tests for accuracy. As a consequence, despite the manner in which it is cited, that literature provides little direct information about accuracy. One has to actually test for accuracy to reach conclusions about inaccuracy. Putting at least some of the huge existing literature on error and bias into some context by also testing for accuracy may be less diYcult than it seems,

ACCURACY IN SOCIAL PERCEPTION

81

because methods are being developed for extracting information about accuracy, agreement, and objectivity from studies that have focused exclusively on inaccuracy, disagreement, and subjectivity (Jussim, 2004; Jussim, Harber, Crawford, Cain, & Cohen, in press). Such methods suggest, for example, that classic demonstrations of bias and subjectivity, such as Rosenhan (1973) and Hastorf and Cantril (1954), actually provide far more evidence of accuracy and objectivity than of bias or subjectivity. At least under some conditions, furthermore, heuristics, biases, and accuracy may be integrated. Many of the classic treatises on flaws and biases (e.g., Kahneman & Tversky, 1973; Nisbett & Ross, 1980) include disclaimers and phrases along the following lines: The many biases, flawed processes, and faulty heuristics we have identified may often serve people quite well. Although the error and bias classics rarely expanded much on this type of claim to show how, why, or when it might actually be correct, the claim itself is correct, for a variety of reasons documented here: bias will sometimes increase accuracy, a flawed or imperfect process may still be good enough to generate mostly accurate predictions or judgments, and expert models are sometimes contradictory, implying at least one is wrong or not applicable, so that demonstrating that people deviate from such models does not always mean they are inaccurate. Both theory and empirical research have begun to address the role of heuristics and biases in increasing accuracy (Dawes, 1989; Jussim, 1991; Kenny & Acitelli, 2001). Indeed, it seems likely that the fact that many heuristics and biases do increase accuracy will help explain why they develop and are diYcult to eliminate. As such, a deeper understanding of the social cognitive processes underlying judgment, prediction, and decision-making will undoubtedly be obtained by research seeking to understand the benefits, and not just the costs, of well-documented biases and heuristics. 3. Accuracy of Implicit Beliefs The type of reaction time, categorization, and priming techniques common to research on implicit and automatic social cognition (Bargh & Chartrand, 1999; Fazio, Jackson, Dunton, & Williams, 1995; Greenwald, McGhee, & Schwartz, 1998) at first glance may not seem to readily lend themselves to the study of accuracy. It is not meaningful, for example, to discuss the ‘‘accuracy’’ of a reaction time. Furthermore, much of this type of research has focused on attitudes, and attitudes cannot be accurate or inaccurate. Nonetheless, research on the accuracy of implicit and automatic beliefs can and should be conducted. One key question involves the extent to which implicit and automatic beliefs correspond to social reality. For example, do

82

LEE JUSSIM

people’s implicit beliefs about diVerences between groups correspond to the size of the actual group diVerences? To address this type of question, people’s scores on the measures reflecting their implicit beliefs (priming eVects, implicit associations, etc.) can be correlated with criteria regarding those beliefs. For example, implicit beliefs about diVerences between African Americans and Whites can be correlated with Census reports, or implicit beliefs about a particular person can be correlated with that person’s attributes. Much of the work to date on implicit beliefs and attitudes emphasizes their dark side, in large part, because it has often focused on prejudice (Fazio et al., 1995; Greenwald et al., 1998). This emphasis is well-justified by the damage that prejudice wreaks on individuals and groups and is broadly consistent with social psychology’s historical emphasis on error and bias. Nonetheless, the extent to which implicit beliefs about the self, other individuals, and groups reflect social reality (in addition to, and, perhaps, sometimes rather than) perceivers’ errors and biases is currently unknown.

4. Evaluating the EVectiveness of Interventions Designed to Increase Accuracy Accuracy interventions, although they are rarely referred to in those terms, pervade everyday life:  Psychotherapy: The goals of personal insight, recognition of dysfunctional behavior patterns and ways to avoid them, and the reframing of unjustifiably damaging views of self or others are often goals of psychotherapy.  Education: A liberal arts background, especially with experience in the humanities and social sciences, can be viewed as, in part, one immense undertaking intended to dissuade students of their inaccurate and parochial views of themselves, their peers, their country, and the various religions, cultures, nationalities, and ethnicities of the world. Advanced instruction in social, organizational, and cognitive psychology; business programs; statistics; and philosophy often provides specific training relevant to eliminating many of the errors and biases that are so well documented in the psychological literature.  Multiculturalism/diversity: Interventions in both business and educational settings designed to reduce prejudice and intergroup conflict by increasing support for multiculturalism and diversity have become commonplace. Multiculturalism/diversity typically has a moral component (‘‘we should value diVerence’’) that is beyond the scope of this chapter, but multiculturalism also has an accuracy component (‘‘we should replace our false beliefs about other groups with more accurate ones’’). This latter point is rarely put so bluntly, but how can we be ‘‘sensitive to others’ cultural

ACCURACY IN SOCIAL PERCEPTION

83

diVerences’’ if we do not have some valid insights into and understandings of those diVerences? This type of point is frequently made in pleas for people to become more multiculturally sensitive. For example, Pinderhughes (1989) titled her book, which is essentially an extended treatise advocating multiculturalism in therapy, Understanding Race, Ethnicity, and Power. Presumably, she believes that such understanding can be attained, otherwise she would not have written the book or given it that title. From the perspective here, this constitutes a call for an increase in the accuracy of people’s beliefs about (understanding of, insight into) those from other cultural backgrounds. I would call that an increase in the accuracy of their stereotypes, because I simply define a stereotype as a belief about a group. Indeed, Pinderhughes (1989, p. 147) makes essentially the same points when she identifies several abilities that enhance people’s multicultural competencies, two of which are ‘‘The ability to control, and even change false beliefs, assumptions, and stereotypes’’ and ‘‘The ability to respect and appreciate the values, beliefs, and practices of persons who are culturally diVerent.’’ Respecting others’ values, beliefs, and practices makes sense primarily if one has a reasonably clear (accurate) sense of what those values, beliefs, and practices are. Unfortunately, however, the accuracy of the new beliefs created by interventions such as therapy, education, and multicultural programs is rarely tested. The validity and eVectiveness of psychotherapy remains controversial (Dawes, 1994; Seligman, 1995). The specific issue of whether clients’ beliefs typically become more accurate through therapy is largely unknown. As a consequence, this would appear to be an interesting and important area for future research. Similarly, little research has evaluated the extent to which education improves accuracy. On the one hand, direct explicit statistical training increases people’s ability to use appropriate statistical reasoning when solving problems (Fong & Nisbett, 1991). On the other, the overwhelming majority of studies demonstrating biased judgments, overreliance on faulty heuristics, and a myriad of failures to understand or apply principles of statistics or logic in everyday reasoning problems have relied on college student samples (e.g., Kahneman et al., 1982; Nisbett & Ross, 1980). Whether people who have received a college education generally reach more accurate conclusions about their friends, coworkers, and peers and about the larger groups in society than do their peers who have not gone to college but who are otherwise similar in age and intelligence is unknown. Presumably, the thousands of diversity and multiculturalism workshops, classes, and programs that have been instituted on campuses around the

84

LEE JUSSIM

country do achieve some degree of reduction in discrimination, if only because they convey the message that ‘‘people in power around here want you all to get along.’’ Furthermore, they are probably superior to doing nothing at all, but this is not guaranteed. Potential negative side eVects, such as a dampening of the free exchange of views crucial to both a college campus and a functioning democracy, and an increase in the willingness to accept immoral or illegal behavior (e.g., acceptance of sexual harassment because someone is from a culture emphasizing machismo) ‘‘because that’s how they do things in that culture,’’ must be evaluated alongside any reductions in prejudice or discrimination. Nonetheless, the eVectiveness of such programs has almost never been evaluated (e.g., McCauley, Wright, & Harris, 2000). Thus, their eVectiveness in leading people to become more acutely aware of (i.e., to more accurately perceive) how the beliefs, values, and practices of people from various backgrounds diVer from their own is both unknown and a potentially rich area for future research.

H. ‘‘THE COMMON AND MAINLY VERIDICAL CHARACTER OF THE BASIC HUMAN PERCEPTIONS’’ This review has suggested that it is no longer justified to dismiss this blossoming of accuracy research for reasons such as accuracy research provides little information about psychological processes (process is a central focus of most accuracy research); research on accuracy is unnecessary because research on social cognition already shows that social perception is generally inaccurate (social cognition research has uncovered a host of errors and biases, but these do not demonstrate that people are generally inaccurate); research on accuracy often has not addressed how perceivers explain others’ behaviors and attributes (it often has not, and this is a limit, but not a threat, to existing accuracy research); self-fulfilling prophecies may account for that which is ‘‘accurately’’ perceived (they may, but this does not preclude the study of accuracy); research on stereotype accuracy is unnecessary because even a stereotype that fits a group will inaccurately describe most members of that group (because this claim confounds levels of analysis, its challenge to stereotype accuracy research fails); it reifies individual and group diVerences (existing scholarship on accuracy has never claimed that accurately perceived diVerences constitute immutable characteristics); it is hopelessly politicized (it sometimes evokes hostile, politically motivated criticisms, and the complex interrelationships of accuracy, bias, and discrimination have received little discussion in the scholarly literature or even in everyday cultural discourse, but the research itself is probably no more

ACCURACY IN SOCIAL PERCEPTION

85

politicized than many other areas of social psychology); it exacerbates social problems (it is a powerful tool for the redress of many social problems); criteria are so fundamentally flawed or ambiguous as to undermine the viability of accuracy research (the criteria are identical to those used to assess self-fulfilling prophecies and to assess the validity of most psychological constructs); or much accuracy research has failed to perform a componential analysis (much has, and even when it has not, accuracy has been productively studied anyway). Fifty years ago, Floyd Allport (1955), in responding to the New Look in Perception research which, arguably, launched the now-five-decadelong social psychological emphasis on error and bias, made the following appeal: ‘‘What we are urging here is that social psychologists, in building their theories of perception, assume their share of the responsibility for reconciling and integrating their ‘social-perceptual’ concepts, fraught with all their deviations and special cognitive loadings, with the common and mainly veridical character of the basic human perceptions’’ (p. 372).

It now seems clear that the meta-theoretical, theoretical, and empirical paths have been suYciently cleared of obstructions for accuracy research to provide the balance to psychological perspectives on social perception that Floyd Allport called for 50 years ago. Acknowledgments I sincerely thank Richard Ashmore, Gretchen Chapman, David Funder, Joachim Krueger, and Clark McCauley, for comments on earlier drafts of this chapter. This chapter is based, in part, on sections from a book by Lee Jussim, currently in progress and tentatively titled Social Beliefs and Social Realities.

References Allport, F. H. (1955). Theories of perception and the concept of structure. New York: Wiley. Allport, G. W. (1958). The nature of prejudice (2nd ed.) Garden City, NY: Doubleday. Ambady, N., & Rosenthal, R. (1992). Thin slices of expressive behavior as predictors of interpersonal consequences: A meta-analysis. Psychological Bulletin, 111, 256–274. American Psychological Association. (1991). In the Supreme Court of the United States: Price Waterhouse v. Ann B. Hopkins (Amicus curiae brief ). American Psychologist, 46, 1061–1070. Archer, D., & Akert, R. M. (1977). Words and everything else: Verbal and nonverbal cues in social interpretation. Journal of Personality and Social Psychology, 35, 443–449.

86

LEE JUSSIM

Aronson, E. (1999). The Social Animal (8th ed.). New York: Worth Publishers. Asch, S. (1955). Opinions and social pressure. Scientific American, 11, 31–55. Ashmore, R. D., & Del Boca, F. K. (1981). Conceptual approaches to stereotypes and stereotyping. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 1–35). Hillsdale, NJ: Erlbaum. Bandura, A. (1977). Self-eYcacy: Toward a unifying theory of behavioral change. Psychological Bulletin, 84, 191–215. Barclay, C. R., & DeCooke, P. A. (1988). Ordinary everyday memories: Some of the things of which selves are made. In U. Neisser & E. Winograd (Eds.), Remembering reconsidered (pp. 91–125). New York: Cambridge University Press. Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of being. American Psychologist, 54, 462–479. Bernieri, F. J., & Gillis, J. S. (2001). Judging rapport: Employing Brunswik’s lens model to study interpersonal sensitivity. In J. A. Hall & F. J. Bernieri (Eds.), Interpersonal sensitivity: Theory and measurement (pp. 67–88). Mahwah, NJ: Lawrence Erlbaum. Biernat, M. (1995). The shifting standards model: Implications of stereotype accuracy for social judgment. In Y. T. Lee, L. Jussim, & C. R. McCauley (Eds.), Stereotype accuracy (pp. 87–114). Washington, DC: American Psychological Association. Block, J. (1993). Studying personality the long way. In D. C. Funder, R. D. Parke, C. Tomlinson-Keasey, & K. Widaman (Eds.), Studying lives through time: Personality and development (pp. 9–41). Washington, DC: American Psychological Association. Block, J. (1965). The challenge of response sets. New York: Meredith Publishing. Bodenhausen, G. V. (1988). Stereotypic biases in social decision making and memory: Testing process models of stereotype use. Journal of Personality and Social Psychology, 55, 726–737. Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley Interscience. Borgida, E., Rudman, L. A., & Manteufel, L. L. (1995). On the courtroom use and misuse of gender stereotyping research. Journal of Social Issues, 51, 181–192. Brewer, M. B. (1988a). A dual process model of impression formation. In T. K. Srull & R. S. Wyer, Jr. (Eds.), Advances in social cognition (Vol. 1, pp. 1–36). Hillsdale, NJ: Erlbaum. Brewer, W. F. (1988b). Memory for randomly sampled autobiographical events. In U. Neisser & E. Winograd (Eds.), Remembering reconsidered (pp. 21–90). New York: Cambridge University Press. Brigham, J. C. (1971). Ethnic stereotypes. Psychological Bulletin, 76, 15–38. Brodt, S. E., & Ross, L. D. (1998). The role of stereotyping in overconfident social prediction. Social Cognition, 16, 225–252. Brophy, J. (Ed.) (1998). Advances in Research on Teaching (Vol. 7). Greenwich, Connecticut: JAI Press. Brophy, J. (1983). Research on the self-fulfilling prophecy and teacher expectations. Journal of Educational Psychology, 75, 631–661. Brophy, J., & Good, T. (1974). Teacher-student relationships: Causes and consequences. New York: Holt, Rinehart, and Winston. Brunswik, E. (1952). The conceptual framework of psychology. Chicago: University of Chicago Press. Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs for research. Chicago: Rand-McNally. Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment. Beverly Hills, CA: Sage. Cialdini, R. B. (1993). Influence: Science and Practice (3rd ed.). New York: HarperCollins.

ACCURACY IN SOCIAL PERCEPTION

87

Chen, M., & Bargh, J. A. (1997). Nonconscious behavioral confirmation processes: The selffulfilling consequences of automatic stereotype activation. Journal of Experimental Social Psychology, 33, 541–560. Claire, T., & Fiske, S. (1998). A systemic view of behavioral confirmation: Counterpoint to the individualist view. In C. Sedikedes, J. Schopler, & C. A. Insko (Eds.), Intergroup cognition and intergroup behavior (pp. 205–231). Mahwah, NJ: Erlbaum. Clarke, R. B., & Campbell, D. T. (1955). A demonstration of bias in estimates of Negro of ability. Journal of Abnormal and Social Psychology, 51, 585–588. Cohen, C. E. (1981). Personal categories and social perception: Testing some boundaries of the processing eVects of prior knowledge. Journal of Personality and Social Psychology, 40, 441–452. Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design and analysis issues for field settings. Chicago: Rand McNally. Craik, K. H., & Appleyard, D. (1980). Streets of San Francisco: Brunswik’s lens model applied to urban inference and assessment. Journal of Social Issues, 36, 72–85. Cronbach, L. J. (1955). Processes aVecting scores on understanding of others and assumed similarity. Psychological Bulletin, 52, 177–193. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281–302. Darley, J. M., & Fazio, R. H. (1980). Expectancy-confirmation processes arising in the social interaction sequence. American Psychologist, 35, 867–881. Darley, J. M., & Gross, P. H. (1983). A hypothesis-confirming bias in labeling eVects. Journal of Personality and Social Psychology, 44, 20–33. Dawes, & Robyn, M. (1994). House of cards: Psychology and psychotherapy built on myth. New York: Free Press. Dawes, R. M. (1989). Statistical criteria for establishing a truly false consensus eVect. Journal of Experimental Social Psychology, 25, 1–17. Dawes, R. M. (1988). Rational Choice in an Uncertain World. San Diego, CA: Harcourt Brace Jovanovich. Dawes, R. M. (1979). The robust beauty of improper linear models in decision making. American Psychologist, 34, 571–582. Devine, P. (1995). Prejudice and outgroup perception. In A. Tesser (Ed.), Advanced Social Psychology (pp. 467–524). New York: McGraw-Hill. Diekman, A. B., Eagly, A. H., & Kulesa, P. (2002). Accuracy and bias in stereotypes about the social and political attitudes of women and men. Journal of Experimental Social Psychology, 38, 268–282. Dunning, D., & Sherman, D. A. (1997). Stereotypes and tacit inference. Journal of Personality and Social Psychology, 73, 459–471. Dusek, J. (Ed.) (1985). Teacher expectancies. Hillsdale, NJ: Erlbaum. Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. San Diego, CA: Harcourt Brace Jovanovich. Eagly, A. H., & Diekman, A. B. (1997). The accuracy of gender stereotypes: A dilemma for feminism. International Review of Social Psychology, 10, 11–30. Eagly, A. H., Makhijani, M. G., Ashmore, R. D., & Longo, L. C. (1991). What is beautiful is good, but . . .: A meta-analytic review of research on the physical attractiveness stereotype. Psychological Bulletin, 110, 109–128. Eccles, J., & Wigfield, A. (1985). Teacher expectations and student motivation. In J. Dusek (Ed.), Teacher expectancies (pp. 185–226). Hillsdale, NJ: Erlbaum.

88

LEE JUSSIM

Ekman, P., & Friesen, W. V. (1969). Nonverbal leakage and cues to deception. Psychiatry, 32, 88–105. Ekman, P., & Friesen, W. V. (1974). Detecting deception from the body or face. Journal of Personality and Social Psychology, 29, 288–298. Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation as an unobstrusive measure of racial attitudes: A bona fide pipeline? Journal of Personality & Social Psychology, 69, 1013–1027. Felson, R. B. (1984). The eVect of self-appraisals of ability on academic performance. Journal of Personality & Social Psychology, 47, 944–952. Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Festinger, L., Riecken, H. W., & Schachter, S. (1956). When prophecy fails. Minneapolis: University of Minnesota Press. Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In D. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 357–411). New York: McGraw-Hill. Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from categorybased to individuating processes: Influences of information and motivation on attention and interpretation. Advances in Experimental Social Psychology, 23, 1–74. Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York: McGraw-Hill. Fong, G. T., & Nisbett, R. E. (1991). Immediate and delayed transfer of training eVects in statistical reasoning. Journal of Experimental Psychology: General, 120, 34–45. Funder, D. C. (2001). Personality. Annual Review of Psychology, 52, 197–221. Funder, D. C. (1995). On the accuracy of personality judgment: A realistic approach. Psychological Review, 102, 652–670. Funder, D. C. (1999). Personality judgment: A realistic approach to person perception. San Diego, CA: Academic Press. Funder, D. C. (1987). Errors and mistakes: Evaluating the accuracy of social judgment. Psychological Bulletin, 101, 75–90. Funder, D. C., & West, S. G. (1993). Consensus, self-other agreement, and accuracy in personality judgment: An introduction. Journal of Personality, 61, 457–476. Gage, N. L., & Cronbach, L. J. (1955). Conceptual and methodological problems in interpersonal perception. Psychological Review, 62, 411–422. Gergen, K. J. (1985). The social constructionist movement in modern psychology. American Psychologist, 40, 266–275. Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton-MiZin. Gilbert, D. T. (1995). Attribution and interpersonal perception. In A. Tesser (Ed.), Advanced social psychology (pp. 99–147). New York: McGraw-Hill. Gilovich, T. (1991). How we know what isn’t so: The fallibility of human reason in everyday life. New York: The Free Press. Goldman, W., & Lewis, P. (1977). Beautiful is good: Evidence that the physically attractive are more socially skillful. Journal of Experimental Social Psychology, 13, 125–130. Gottfredson, L. S. (1994). The science and politics of race-norming. American Psychologist, 49, 955–963. Gould, S. J. (1978). The mismeasure of man. New York: Norton. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual diVerences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464–1480. Hall, J. A., & Carter, J. D. (1999). Gender-stereotype accuracy as an individual diVerence. Journal of Personality and Social Psychology, 77, 350–359.

ACCURACY IN SOCIAL PERCEPTION

89

Hamilton, D. L., Sherman, S. J., & Ruvolo, C. M. (1990). Stereotype-based expectancies: EVects on information processing and social behavior. Journal of Social Issues, 46, 35–60. Hastie, R., & Dawes, R. (2001). Rational choice in an uncertain world. Thousand Oaks, CA: Sage. Hastorf, A. H., & Cantril, H. (1954). They saw a game: A case study. Journal of Abnormal and Social Psychology, 47, 129–143. Hastie, R., & Rasinski, K. A. (1988). The concept of accuracy in social judgment. In D. Bar-Tal & A. W. Kruglanski (Eds.), The social psychology of knowledge (pp. 193–208). Cambridge: Cambridge University Press. Hedges, S. M., Jandorf, L., & Stone, A. A. (1985). Meaning of daily mood assessments. Journal of Personality and Social Psychology, 48, 428–434. Herrnstein, R. J., & Murray, C. (1994). The bell curve. New York: The Free Press. Hoge, R. D., & Butcher, R. (1984). Analysis of teacher judgments of pupil achievement levels. Journal of Educational Psychology, 76, 777–781. Ickes, W. J. (1997). Empathic accuracy. New York: The Guilford Press. Ickes, W. (1993). Empathic accuracy. Journal of Personality, 61, 587–610. Jacoby, R., & Glauberman, N. (1995). The bell curve debate. New York: Times Books. Jones, E. E. (1985). Major developments in social psychology during the past five decades. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (3rd ed., Vol. 1, pp. 47–107). New York: Random House. Jones, E. E. (1986). Interpreting interpersonal behavior: The eVects of expectancies. Science, 234, 41–46. Jones, E. E. (1990). Interpersonal perception. New York: W. H. Freeman and Company. Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system-justification and the production of false consciousness. British Journal of Social Psychology, 33, 1–27. Jost, J. T., & Kruglanski, A. W. (2002). The estrangement of social constructionism and experimental social psychology: History of the rift and prospects for reconciliation. Personality and Social Psychology Review, 6, 168–187. Judd, C. M., & Park, B. (1993). Definition and assessment of accuracy in social stereotypes. Psychological Review, 100, 109–128. Judd, C. M., Park, B., Ryan, C. S., Brauer, M., & Kraus, S. (1995). Stereotypes and ethnocentrism: Diverging interethnic perceptions of African American and White American youth. Journal of Personality & Social Psychology, 69, 460–481. Jussim, L. (1986). Self-fulfilling prophecies: A theoretical and integrative review. Psychological Review, 93, 429–445. Jussim, L. (1989). Teacher expectations: Self-fulfilling prophecies, perceptual biases, and accuracy. Journal of Personality and Social Psychology, 57, 469–480. Jussim, L. (1991). Social perception and social reality: A reflection-construction model. Psychological Review, 98, 54–73. Jussim, L. (2004). The goodness of judgment index. Behavioral and Brain Sciences, 27, 344–345. Jussim, L., & Harber, K. (In press). Teacher expectations and self-fulfilling prophecies: Knowns and unknowns, resolved and unresolved controversies. Personality and Social Psychology Review. Jussim, L., Harber, K., Crawford, J., Cain, T., & Cohen, F. (in press). Social reality makes the social mind: Self-fulfilling prophecy, stereotypes, bias, and accuracy in social perception. Interaction Studies. Jussim, L., Eccles, J., & Madon, S. J. (1996). Social perception, social stereotypes, and teacher expectations: Accuracy and the quest for the powerful self-fulfilling prophecy. Advances in Experimental Social Psychology, 29, 281–388.

90

LEE JUSSIM

Jussim, L., McCauley, C. R., & Lee, Y. T. (1995). Why study stereotype accuracy and inaccuracy? In Y. T. Lee, L. Jussim, & C. R. McCauley (Eds.), Stereotype accuracy: Toward appreciating group diVerences (pp. 3–27). Washington, DC: American Psychological Association. Kahneman, D., Slovic, P., & Tversky, A. (1982). Judgment under uncertainty: Heuristics and biases. New York: Cambridge University Press. Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Review, 80, 237–251. Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska symposium on motivation (Vol. 15). Lincoln: University of Nebraska Press. Kelly, G. A. (1955). The psychology of personal constructs (Vol. 1). New York: W. W. Norton. Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York: Guilford. Kenny, D. A., & Acitelli, L. K. (2001). Accuracy and bias in the perception of the partner in a close relationship. Journal of Personality and Social Psychology, 80, 439–448. Kenny, D. A., & Albright, L. (1987). Accuracy in interpersonal perception: A social relations analysis. Psychological Bulletin, 102, 390–402. Kenny, D. A., Kashy, D., & Bolger, N. (1998). Data analysis in social psychology. In D. Gilbert, S. I. Fiske, & J. Lindzey (Eds.), Handbook of Social Psychology (4th ed.). New York: McGraw Hill. Koriat, A., Goldsmith, M., & Panksy, A. (2000). Toward a psychology of memory accuracy. Annual Review of Psychology, 51, 481–537. Krueger, J. (1996). Probabilistic national stereotypes. European Journal of Social Psychology, 26, 961–980. Krueger, J., & Funder, D. C. (2004). Towards a balanced social psychology: Causes, consequences, and cures for the problem-seeking approach to social behavior and cognition. Behavioral and Brain Sciences, 27, 313–367. Krueger, J. I., Hasman, J. F., & Acevedo, M. (2003). Perceptions of trait typicality in gender stereotypes: Examining the role of attribution and categorization processes. Personality and Social Psychology Bulletin, 29, 108–116. Kruglanski, A. (1989). The psychology of being ‘‘right’’: The problem of accuracy in social perception and cognition. Psychological Bulletin, 106, 395–409. Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480–498. Kunda, Z., & Thagard, P. (1996). Forming impressions from stereotypes, traits, and behaviors: A parallel-constraint-satisfaction theory. Psychological Review, 103, 284–308. LaPiere, R. T. (1934). Attitudes versus actions. Social Forces, 13, 230–237. LaPiere, R. T. (1936). Type-rationalizations of group antipathy. Social Forces, 15, 232–237. Latane, B., & Darley, J. M. (1970). The unresponsive bystander: Why doesn’t he help. New York: Appleton-Century-Crofts. Lee, Y. T., Jussim, L., & McCauley, C. R. (Eds.) (1995). Stereotype Accuracy: Toward Appreciating Group DiVerences. Washington, DC: American Psychological Association. Locksley, A., Borgida, E., Brekke, N., & Hepburn, C. (1980). Sex stereotypes and social judgment. Journal of Personality and Social Psychology, 39, 821–831. Lord, C. G., Lepper, M. R., & Preston, E. (1984). Considering the opposite: A corrective strategy for social judgment. Journal of Personality and Social Psychology, 47, 1231–1243. Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimilation and attitude polarization: The eVects of prior theories on subsequently considered evidence. Journal of Personality and Social Psychology, 37, 2098–2109. Madon, S., Guyll, M., Spoth, R. L., Cross, S. E., & Hilbert, S. J. (2003). The self-fulfilling influence of mother expectations on children’s underage drinking. Journal of Personality and Social Psychology, 84, 1188–1205.

ACCURACY IN SOCIAL PERCEPTION

91

Madon, S. J., Jussim, L., Keiper, S., Eccles, J., Smith, A., & Palumbo, P. (1998). The accuracy and power of sex, social class and ethnic stereotypes: Naturalistic studies in person perception. Personality and Social Psychology Bulletin, 24, 1304–1318. Malkiel, B. G. (1999). A random walk down Wall Street. New York: Norton. McArthur, L. Z., & Baron, R. M. (1983). Toward an ecological theory of social perception. Psychological Review, 90, 215–238. McCauley, C. (1995). Are stereotypes exaggerated? A sampling of racial, gender, occupational, academic, and political stereotypes. In Y. T. Lee, L. Jussim, & C. R. McCauley (Eds.), Stereotype accuracy: Toward appreciating group diVerences (pp. 215–243). Washington, D.C: American Psychological Association. McCauley, C., Jussim, L., & Lee, Y. T. (1995). Stereotype accuracy: Toward appreciating group diVerences. In Y. T. Lee, L. Jussim, & C. R. McCauley (Eds.), Stereotype accuracy: Toward appreciating group diVerences (pp. 293–312). Washington, DC: American Psychological Association. McCauley, C., & Stitt, C. L. (1978). An individual and quantitative measure of stereotypes. Journal of Personality and Social Psychology, 36, 929–940. McCauley, C., Stitt, C. L., & Segal, M. (1980). Stereotyping: From prejudice to prediction. Psychological Bulletin, 87, 195–208. McCauley, C., & Thangavelu, K. (1991). Individual diVerences in strength of sex-occupation stereotyping: Reliability, validity, and some correlates. Social Psychology Quarterly, 54, 195–208. McCauley, C., Wright, M., & Harris, M. E. (2000). Diversity workshops on campus: A survey of current practice at U.S. colleges and universities. College Student Journal, 34, 100–114. McNatt, D. B. (2000). Ancient Pygmalion joins contemporary management: A meta-analysis of the result. Journal of Applied Psychology, 85, 314–322. Meehl, P. E. (1990). Appraising and amending theories: The strategy of Lakatosian defense and two principles that warrant it. Psychological Inquiry, 1, 108–141. Merton, R. K. (1948). The self-fulfilling prophecy. Antioch Review, 8, 193–210. Milgram, S. (1974). Obedience to authority. New York: Harper & Row. Miller, D. T., & Turnbull, W. (1986). Expectancies and interpersonal processes. Annual Review of Psychology, 37, 233–256. Murray, S. L., Holmes, J. G., & GriYn, D. W. (1996). The self-fulfilling nature of positive illusions in romantic relationships: Love is not blind, but prescient. Journal of Personality and Social Psychology, 71, 1155–1180. Myers, D. G. (1999). Social psychology (6th ed.). New York: McGraw-Hill. Neisser, U., Boodoo, G., Bouchard, T. J., Jr., Boykin, A. W., Brody, N., Ceci, S. J., Halpern, D. F., Loehlin, J. C., PerloV, Rt., Sternberg, R. J., & Urbina, S. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51, 77–101. Nelson, T. (2002). The psychology of prejudice. Boston: Allyn & Bacon. Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgment. Englewood CliVs, NJ: Prentice-Hall. Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231–259. Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social reality. Cambridge, MA: Blackwell. Olson, J. M., Roese, N. J., & Zanna, M. P. (1996). Expectancies. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 211–239). New York: Guilford Press.

92

LEE JUSSIM

Paulhus, D. L. (1991). Measurement and control of response bias. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes. San Diego, CA: Academic Press. Paulhus, D. L. (2002). Socially desirable responding: Evolution of a construct. In H. Braun, D. N. Jackson, & D. E. Wiley (Eds.), The role of constructs in psychological and educational measurement. Mahwah, NJ: Erlbaum. Paunonen, S. V. (1991). On the accuracy of ratings of personality by strangers. Journal of Personality and Social Psychology, 61, 471–477. Pickering, M. (2001). Stereotyping. New York: Palgrave. Pinderhughes, E. (1989). Understanding race, ethnicity, and power: The key to eYcacy in clinical practice. New York: The Free Press. Ra¨ikko¨nen, K., Matthews, K. A., Flory, J. S., Owens, J. F., & Gump, B. B. (1999). EVects of optimism, pessimism, and trait anxiety on ambulatory blood pressure and mood during everyday life. Journal of Personality and Social Psychology, 76, 104–113. Raudenbush, S. W. (1984). Magnitude of teacher expectancy eVects on pupil IQ as a function of the credibility of expectancy inductions: A synthesis of findings from 18 experiments. Journal of Educational Psychology, 76, 85–97. Redding, R. E. (2001). Sociopolitical diversity in psychology: The case for pluralism. American Psychologist, 56, 205–215. Reynolds, D. J., Jr., & GiVord, R. (2001). The sounds and sights of intelligence: A Lens Model channel analysis. Personality and Social Psychology Bulletin, 65, 187–2000. Robinson, J. P. Shaver, P. R., & Wrightsman, L. S. (Eds.) (1991). Measures of personality and social psychological attitudes. San Diego, CA: Academic Press. Robinson, J. P. Shaver, P. R., & Wrightsman, L. S. (Eds.) (1999). Measures of political psychological attitudes. San Diego, CA: Academic Press. Rosenhan, D. L. (1973). On being sane in insane places. Science, 179, 250–258. Rosenthal, R., & Jacobson, L. (1968). Pygmalion in the classroom: Teacher expectations and student intellectual development. New York: Holt, Rinehart, and Winston. Rosenthal, R., & Rubin, D. B. (1978). Interpersonal expectancy eVects: The first 345 studies. The Behavioral and Brain Sciences, 3, 377–386. Ross, L., & Nisbett, R. E. (1991). The person and the situation. New York: McGraw-Hill. Ryan, C. S. (2002). Stereotype accuracy. European Review of Social Psychology, 13, 75–109. Ryan, C. S. (1996). Accuracy of black and white college students’ in-group and out-group stereotypes. Personality and Social Psychology Bulletin, 22, 1114–1127. Ryan, C. S. (1995). Motivations and the perceiver’s group membership: Consequences for stereotype accuracy. In Y. T. Lee, L. Jussim, & C. R. McCauley (Eds.), Stereotype accuracy (pp. 189–214). Washington, DC: American Psychological Association. Ryan, C. S., & Bogart, L. M. (2001). Longitudinal changes in the accuracy of new group members’ in-group and out-group stereotypes. Journal of Experimental Social Psychology, 37, 118–133. Ryan, C., Park, B., & Judd, C. M. (1996). Assessing stereotype accuracy: Implications for understanding the stereotyping process. In C. N. Macrae, C. Stangor, & M. Hewstone (Eds.), Stereotypes and stereotyping. New York: Guilford. Schmitt, N., Noe, R. A., & Gottschalk, R. (1986). Using the lens model to magnify raters’ consistency, matching, and shared bias. Academy of Management Journal, 29, 130–139. Schneider, D. J., Hastorf, A. H., & Ellsworth, P. C. (1979). Person perception (2nd ed.). Reading, MA: Addison-Wesley. Seligman, & Martin, E. P. (1995). The eVectiveness of psychotherapy: The Consumer Reports study. American Psychologist, 50, 965–974.

ACCURACY IN SOCIAL PERCEPTION

93

Sidanius, J., & Pratto, F. (1999). Social dominance: an intergroup theory of social hierarchy and oppression. New York: Cambridge University Press. Snyder, M. (1984). When belief creates reality. Advances in Experimental Social Psychology, 18, 247–305. Snyder, M. (1992). Motivational foundations of behavioral confirmation. Advances in Experimental Social Psychology, 18, 247–305. Snyder, M., & Stukas, A. A., Jr. (1998). Interpersonal processes: The interplay of cognitive, motivational, and behavioral activities in social interaction. Annual Review of Psychology, 50, 273–303. Snyder, M., Tanke, E. D., & Berscheid, E. (1977). Social perception and interpersonal behavior: On the self-fulfilling nature of social stereotypes. Journal of Personality and Social Psychology, 35, 656–666. Spitz, H. H. (1999). Beleaguered Pygmalion: A history of the controversy over claims that teacher expectancy raises intelligence. Intelligence, 27, 199–234. Stangor, C. (1995). Content and application inaccuracy in social stereotyping. In Y. T. Lee, L. Jussim, & C. R. McCauley (Eds.), Stereotype accuracy (pp. 275–292). Washington, DC: American Psychological Association. Stangor, C., & McMillan, D. (1992). Memory for expectancy-congruent and expectancyincongruent information: A review of the social and social developmental literatures. Psychological Bulletin, 111, 42–61. Stinson, L., & Ickes, W. (1992). Empathic accuracy in the interactions of male friends versus male strangers. Journal of Personality and Social Psychology, 62, 787–797. Swann, W. B. (1984). Quest for accuracy in person perception: A matter of pragmatics. Psychological Review, 91, 457–477. Swim, J. K. (1994). Perceived versus meta-analytic eVect sizes: An assessment of the accuracy of gender stereotypes. Journal of Personality and Social Psychology, 66, 21–36. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193–210. Trouilloud, D., Sarrazin, P., Martinek, T., & Guillet, E. (2002). The influence of teacher expectations on students’ achievement in physical education classes: Pygmalion revisited. European Journal of Social Psychology, 32, 1–17. Tversky, A., & Kahneman, D. (1983). Extensions versus intuitive reasoning: The conjunction fallacy in probability judgment. Psychological Review, 90, 293–315. U. S. Census (2002). Money income in the United States: 2001 (Web address: http:// www.census.gov/prod/2002pubs/p60-218.pdf). West, C., & Anderson, T. (1976). The question of preponderant causation in teacher expectancy research. Review of Educational Research, 46, 613–630. Williams, T. (1976). Teacher prophecies and the inheritance of inequality. Sociology of Education, 49, 223–236. Wilson, T. D., Lindsey, S., & Schooler, T. Y. (2000). A model of dual attitudes. Psychological Review, 107, 101–126. Wolsko, C., Park, B., Judd, C. M., & Wittenbrink, B. (2000). Framing interethnic ideology: EVects of multicultural and color-blind perspectives on judgments of groups and individuals. Journal of Personality and Social Psychology, 78, 635–654. Word, C. O., Zanna, M. P., & Cooper, J. (1974). The nonverbal mediation of self-fulfilling prophecies in interracial interaction. Journal of Experimental Social Psychology, 10, 109–120. Wright, J. C., & Drinkwater, M. (1997). Rationality vs. accuracy of social judgment. Social Cognition, 15, 245–273.

This Page Intentionally Left Blank

OVER THIRTY YEARS LATER: A CONTEMPORARY LOOK AT SYMBOLIC RACISM

David O. Sears P. J. Henry The theory of symbolic racism was proposed over 30 years ago to explain new forms of racial conservatism appearing after the civil rights era (Sears and Kinder, 1970). It has been controversial and has stimulated much research. We here summarize the available evidence on twelve controversies surrounding symbolic racism. It remains a coherent, distinctive, and popular belief system: (1) the conceptualization of symbolic racism was originally somewhat fuzzy and has evolved over time; but (2) the measurement of it has been surprisingly constant over time; and (3) it seems to form a substantively meaningful and statistically consistent belief system, with two highly correlated variants that diVer slightly in the language they use but not in their eVects. Its eVects on racial politics are quite stable and consistent: (4) all four themes of symbolic racism have about the same eVect, and indeed, it seems not to matter at all which conventional measure of symbolic racism is used; (5) its eVects on racial policy preferences are not weakened if items that might be ‘‘close’’ in content to such preferences are removed from the measures of symbolic racism; (6) it has remarkably similar eVects on preferences about quite diverse racial policies; and (7) its eVects on racial policy *Previous versions of this chapter were presented at the annual meeting of the International Society of Political Psychology, Seattle, July, 2000, and the Attitudes Preconference of the Society for Personality and Social Psychology, Los Angeles, February 6, 2003. We would like to express our thanks to our colleagues Donald R. Kinder and the late John B. McConahay, with whom the original theory of symbolic racism was developed; to our research collaborators, Harris M. Allen, Mary Carrillo, Sharmaine Cheleden, Jack Citrin, Tom Jessor, Rick Kosterman, Christopher Tarman, Nicholas Valentino, and Colette van Laar; to our colleagues who have given us much valuable feedback over the years, usually without as much gratitude, Lawrence Bobo, Franklin Gilliam, Jr., and Jim Sidanius; and to our critics who have done much to help sharpen our thinking and our research. 95 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 37

Copyright 2005, Elsevier Inc. All rights reserved. 0065-2601/05 $35.00

96

DAVID O. SEARS AND P. J. HENRY

preferences are just as strong among the college educated as they are among the less educated. Its origins (8) do indeed appear to lie in part in a ‘‘blend’’ of racial animosity and conservative traditional values such as individualism. Several initially plausible alternative explanations for symbolic racism and its eVects do not bear fruit. It lies in the nomological net originally argued for it, and not others: (9) it is quite diVerent from (though modestly correlated with) old-fashioned racism; (10) it is correlated with political ideology, but its political eVects are quite distinct from and stronger than those of ideology; and (11) it is a distinctive construct necessary for the understanding of Whites’ responses to racial politics, not merely redundant with other constructs and hence disposable in the service of parsimony. Finally, (12) it focuses on antagonism toward Blacks, not White group consciousness, which has little to do with either symbolic racism or opposition to policies targeted for Blacks. All these findings replicate strongly over surveys. In our view these particular controversies ought properly to be set aside and regarded as settled unless and until contrary evidence emerges. We do, however, review seven other questions, less central to the validity of the theory of symbolic racism but relevant to it, about which we appraise the evidence as less clear. Racial conflicts have plagued the United States from its very beginnings, in particular driven by racial prejudice against Blacks. When Myrdal (1944) published his classic treatise on race relations in America, African Americans were second-class citizens, denied the pursuit of the American dream in all spheres of life—socially, economically, and politically. Since then, Jim Crow segregationism has been eliminated and Whites’ opinions about racial issues have liberalized in many areas (Schuman, Steeh, Bobo, & Krysan, 1997). However, neither development has ended the substantial disadvantages experienced by African Americans. Proponents of Blacks’ interests have therefore continued to push for further advances. These eVorts have often met with substantial White opposition ranging, over the years since the Civil Rights era, from the Republican ‘‘Southern strategy’’ of the 1960s to the furors over court-ordered busing in the 1970s to contemporary initiatives to outlaw aYrmative action and debates about the removal of Confederate symbols from state flags. Such opposition has generated considerable research interest in the determinants of Whites’ attitudes about racially targeted policies, Black political candidates, and White candidates charged with appealing to racial prejudice. This is the problem that has animated our own research agenda: how to understand Whites’ continuing resistance to eVorts to increase racial equality despite much evidence that in some measurable ways their racial attitudes have become substantially liberalized.

SYMBOLIC RACISM: A CONTEMPORARY LOOK

97

Theoretical explanations for the roots of this opposition, in terms of the White public’s attitudes, fall into four main categories. The first is that some new form of racism has taken over the political role played in pre–Civil Rights days by ‘‘old-fashioned,’’ ‘‘redneck,’’ or ‘‘Jim Crow racism.’’ Another treats contemporary racial politics as a largely unexceptional case of normal political processes, in which the key ingredient is not racial animosity in the mass public but elites’ agenda control and coalition formation and their appeals to the mass public’s nonracial political ideologies and values (Sniderman & Carmines, 1997; Sniderman & Piazza, 1993; Sniderman, Crosby, & Howell, 2000). A third emphasizes the inevitable group conflicts that stem from structural inequalities, such as realistic conflicts of interest (Bobo, 1988) or dominant groups’ desire for social dominance (Sidanius, Levin, Rabinowitz, & Federico, 1999), threatened ‘‘sense of group position’’ (Bobo, 1999), or use of a ‘‘dominant ideology’’ to legitimatize privilege (Kluegel & Smith, 1986). Finally, a fourth looks to an ‘‘implicit’’ racism that is largely automatic and unconscious (Devine, 1989; Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997; Fazio, Jackson, Dunton, & Williams, 1995; Greenwald & Banaji, 1995). Our focus here is on the first explanation—the ‘‘new racism’’ theories and challenges to them. Each of these theories contains the assumptions that new forms of prejudice became politically potent after the civil rights revolution of the 1960s and that they embody both negative feelings toward Blacks as a group and some conservative nonracial values. One version has variously been described as ‘‘symbolic racism’’ (Kinder & Sears, 1981; McConahay & Hough, 1976; Sears & Kinder, 1971), ‘‘modern racism’’ (McConahay, 1986), or ‘‘racial resentment’’ (Kinder & Sanders, 1996). Although these terms have some slight conceptual diVerences, they have been operationalized with similar survey items, and we will not distinguish among them here. There are other distinctive, but related, concepts of a ‘‘new racism,’’ but we will not focus on them in this chapter. These include ‘‘subtle prejudice’’ (Pettigrew & Meertens, 1995), reflecting ‘‘defense of traditional values,’’ ‘‘exaggeration of cultural diVerences,’’ and ‘‘denial of positive emotions’’ toward the target group; ‘‘racial ambivalence’’ (Katz, 1981), indicating that many Whites are ambivalent about blacks, with ‘‘pro-Black’’ attitudes rooted in humanitarianism and egalitarianism, whereas ‘‘anti-Black’’ attitudes are based in values such as the Protestant ethic; ‘‘aversive racism,’’ combining a belief that equal treatment of Blacks is very important with negative attitudes toward them (Gaertner & Dovidio, 1986); and ‘‘laissez-faire racism’’ (Bobo & Smith, 1994), combining denial of continuing racial discrimination with attributions of Blacks’ disadvantages to their own lack of eVort in a market-driven society.

98

DAVID O. SEARS AND P. J. HENRY

I. The Theory of Symbolic Racism The concept of symbolic racism was the first of these ‘‘new racisms.’’ It and its brethren have perhaps stimulated the most research and the most controversy. A recent comprehensive social-psychological review of measures of racial attitudes (Biernat & Crandall, 1999) describes measures of symbolic racism as the most widely used measures of explicit racism today. A leading sociological review of racial attitudes treats the theory of symbolic racism as one of the major social-scientific challenges to post–Civil Rights optimism about the winding down of racism in America (Schuman et al., 1997). Social structural theorists concur that measures of symbolic racism are the most powerful predictors of White opposition to racial policies (Bobo, 2000; Sidanius et al., 1999). Political scientists treat the symbolic racism thesis as central to the study of race in American politics (Hutchings & Valentino, 2004). In like fashion, those who believe that racial prejudice has lost much of its force in American politics have focused most of their critical attention on the theory of symbolic racism (e.g., Hurwitz & PeZey, 1998; Sniderman, Crosby, & Howell, 2000; Sniderman & Tetlock, 1986a,b; Tetlock, 1994). The original theory (Kinder & Sears, 1981; McConahay & Hough, 1976; Sears & Kinder, 1971; Sears & McConahay, 1973) embodied three essential propositions. First, a new form of racism, ‘‘symbolic racism,’’ had replaced the Jim Crow or ‘‘old-fashioned’’ racism, which had embodied social distance between the races, beliefs in the biological inferiority of Blacks, and support for formal discrimination and segregation in pre–Civil Rights days. Symbolic racism had replaced Jim Crow racism in two senses: the latter was no longer very popular (indeed, it had almost disappeared in the Los Angeles suburbs we looked at), and because only a tiny minority still accepted it, oldfashioned racism no longer could be very influential in ordinary politics. Second, opposition to Black politicians and racially targeted policies were more influenced by symbolic racism than by realistic self-interest; that is, more than by real or perceived racial threats to Whites’ own personal lives. Finally, the origins of symbolic racism lay in a blend of early-socialized negative feelings about Blacks with traditional conservative values. Symbolic racism was labeled ‘‘symbolic’’ to highlight its roots in abstract moral values, rather than in concrete self-interest or personal experience, and because it targets Blacks as an abstract collectivity rather than specific Black individuals. The label ‘‘racism’’ was chosen because the construct was thought in part to reflect racial antipathy. A simple example of the original finding is reproduced in Table I. It shows that support for Sam Yorty, the race-baiting White opponent of the Black mayoral candidate in Los Angeles in 1969, Tom Bradley, was greatest

99

SYMBOLIC RACISM: A CONTEMPORARY LOOK TABLE I Percentage Preferring the White (Yorty) over Black (Bradley) Mayoral Candidate as a Function of Symbolic Racism and Ideological Conservatism (San Fernando Valley, 1969) High symbolic racism

Low symbolic racism

DiVerence

Conservatives

76

42

þ34

Liberals

58

17

þ41

Conservatives

71

50

þ21

Liberals

45

15

þ30

Conservatives

80

43

þ37

Liberals

67

4

þ63

Could most Negroes get along without welfare if they tried?

Negroes shouldn’t push themselves where they’re not wanted:

Do LA city oYcials pay more, less, or the same attention to a request from a black person as from a white person?

Source: Adapted from Sears and Kinder, 1971. Note: Each entry is the proportion of Whites preferring the White candidate, Yorty, to the Black candidate, Bradley, among those with the specified combination of symbolic racism and political ideology. The high symbolic racism responses to the three items are ‘‘could get along without it,’’ ‘‘agree,’’ and ‘‘more attention,’’ respectively.

among Whites who endorsed three tenets of measuring symbolic racism: that Blacks have a poor work ethic and really do not need welfare, that they are making demands that are unwanted and illegitimate, and that they get excessive attention from government. The table also shows that all three eVects of symbolic racism hold up with general political ideology controlled; that is, among both ideological conservatives and liberals. This latter finding reflects a point that has often been misunderstood ever since: The eVects of symbolic racism do partially overlap those of political conservatism, but they are not reducible to general conservatism. They occur above and beyond the eVects of general ideology, adding the unique eVects of racism itself. The measurement of symbolic racism at that time was somewhat happenstance rather than operationalizing a refined conceptual definition. It mainly reflected a few items that intuitively seemed to reflect a common construct and happened to be available in a particular survey. Since then, the concept has come under considerable scrutiny and, as a result, has evolved. It now is

100

DAVID O. SEARS AND P. J. HENRY

usually described as a coherent belief system reflecting a unidimensional underlying prejudice toward Blacks, with four specific themes: the beliefs that Blacks no longer face much prejudice or discrimination, that their failure to progress results from their unwillingness to work hard enough, that they are demanding too much too fast, and that they have gotten more than they deserve (Henry & Sears, 2002; Sears, Henry, & Kosterman, 2000).2 Table II shows the items most commonly used to measure these themes. The theory of symbolic racism has also become more specific with time. It is graphically displayed in Fig. 1. It depicts the sources of symbolic racism as rooted originally in early socialization experiences. The individual’s early social environment is said to pass on its conventional attitudes about blacks. Those attitudes evolve into a ‘‘blend’’ of primitive anti-Black aVects with traditional values that on their face have little to do with Blacks. That leads to the acceptance of beliefs supporting the four specific themes of the symbolic racism belief system. And that, in turn, is the most powerful proximal predictor of opposition to Black candidates and to contemporary racial policies. Simultaneously, other socialization experiences generate a more general political conservatism that has (as shown in Table I) independent eVects on these same political dependent variables. Since those early findings, many studies have shown that symbolic racism, conceptualized and measured in this form, is strongly associated with Whites’ racial policy preferences. Its explanatory power typically outweighs that of other important political attitudes, such as ideology, party identification, and attitudes toward the size of the federal government, or of older and more traditional racial attitudes, such as beliefs in Blacks’ genetic inferiority, support for racial segregation, negative stereotypes, or anti-Black aVect (e.g., Bobo, 2000; Hughes, 1997; Kinder & Mendelberg, 2000; Kinder & Sanders, 1996; Sears & Henry, 2003; Sears, van Laar, Carrillo, & Kosterman, 1997; Sidanius, Devereux, & Pratto, 1992; Sidanius et al., 1999; Tarman & Sears, 2005). A standard design has been to include all such variables in a regression equation predicting Whites’ racial policy attitudes. Typical findings are shown in Fig. 2. The standardized regression coeYcients for symbolic racism are far higher than those for either political conservatism or traditional racial attitudes. Symbolic racism has also been demonstrated to have systematic eVects on Whites’ candidate preferences in racialized election campaigns (e.g., Howell, 1994; Kinder & Sanders, 1996; 2 Items directly measuring emotions toward Blacks have also been used, but much less often. McConahay did employ an item on understanding Blacks’ anger in his studies (McConahay & Hough, 1976; McConahay, 1986), and another study used the absence of positive emotions toward Blacks (Sears et al., 1997), but these are the exceptions.

SYMBOLIC RACISM: A CONTEMPORARY LOOK

101

Kinder & Sears, 1981; McConahay & Hough, 1976; Mendelberg, 2001; Sears et al., 1997), and it has been shown to drive the shift of White Southerners from predominantly Democratic to predominantly Republican party aYliation over the last several decades (Valentino & Sears, 2005).

II. Controversies About Symbolic Racism The symbolic racism claim is an important one—that the politics of race are not merely ‘‘politics as usual’’ but instead are significantly distorted by the underlying racial prejudice held by many racial conservatives, and that race-neutral conservative rhetoric often disguises underlying racial animosity. It goes to the substantive core of America’s longest-running and most diYcult social problem. Not surprisingly, then, the theory has been controversial and has stimulated some heated criticism. Some of it has come from those who believe that racial prejudice has become a minor political force and that the theory of symbolic racism mistakenly treats ordinary political conservatism and its eVects as reflecting racial prejudice (Carmines & Merriman, 1993; Hurwitz & PeZey, 1998; Sniderman & Piazza, 1993; Sniderman & Tetlock, 1986a,b; Tetlock, 1994). Other criticism has come from those who believe the symbolic racism theory ignores the vested interest that Whites, as the dominant group in a racially hierarchical society, have in maintaining their privileged position. To these scholars, defense of that vested interest is the key force motivating Whites’ opposition to greater racial equality. In their models, symbolic racism serves largely as a myth to legitimize Whites’ superior racial status (Bobo, 1983, 1988, 2000; Sidanius et al., 1992, 1999). These controversies are of more than mere academic relevance. If the symbolic racism claim is right, much remedial work of a variety of kinds needs to be done on the White side of the racial divide. If it is wrong, and racial conservatives’ views about the optimal relative balance of governments and markets in modern societies are largely free of underlying racial prejudice, much obligation would be placed on Blacks to adapt to a society in which they no longer are being treated much less fairly than other Americans. Alternatively, if Whites are primarily motivated by a defense of their privileged position, a diVerent strategy is called for, perhaps one that calls for more direct exercise of whatever powers minorities have. Criticism of any line of research generally challenges either its empirical findings or its interpretations of those findings. In this particular case, most of it has centered on the interpretation of the primary findings rather than disputing those findings themselves. Even the theory’s critics conclude

TABLE II Widely Used Symbolic Racism Items Theme 1: Denial of Continuing Discrimination 1. Generations of slavery and discrimination have created conditions that make it diYcult for blacks to work their way out of the lower classes (disagree) [3 NES, 5 LACSS]

Mean .51

2. How much discrimination against blacks do you feel there is in the United States today, limiting their chances to get ahead? (none at all) [1 NES, 4 LACSS]

.34

3. Discrimination against blacks is no longer a problem in the United States (agree) [1 LACSS]

.24

4. Has there been a lot of real change in the position of black people in the past few years? (a lot) [2 NES, 1 LACSS]

.58

Theme 2: Work Ethic and Responsibility for Outcomes 5. It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well oV as whites (agree) [3 NES, 5 LACSS]

.54

6. Irish, Italian, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors (agree)* [4 NES, 2 LACSS, 1 GSS]

.67

7. Blacks work just as hard to get ahead as most other Americans (disagree) [1 LACSS]

.30

8. Most blacks who receive money from welfare programs could get along without it if they tried (agree)* [2 NES, 1 LACSS]

.61

Theme 3: Excessive Demands 9. Some say that the Civil Rights people have been trying to push too fast. Others feel that they haven’t pushed fast enough. How about you? (push too fast) [5 NES, 1 LACSS]

.56

10. Blacks are getting too demanding in their push for equal rights (agree) [6 LACSS]

.49

11. Blacks are demanding too much from the rest of society (agree) [2 LACSS]

.47

12. How much of the racial tension that exists in the United States today do you think blacks are responsible for creating? (all of it) [3 LACSS]

.33

Theme 4: Undeserved Advantage 13. Over the last few years, blacks have gotten less than they deserve (disagree) [4 NES, 5 LACSS]

.57

14. Over the past few years, blacks have gotten more economically than they deserve (agree) [2 LACSS]

.27

15. Government oYcials usually pay less attention to a request or complaint from a black person than from a white person (disagree)* [2 NES, 1 LACSS]

.52

16. Do blacks get much more attention from the government than they deserve, more attention, about the right amount, less attention, or much less attention from the government than they deserve? (much more)* [3 LACSS, 1 GSS]

.55

Sources: 1985, 1986, 1990, 1992, 2000 National Election Studies (NES); 1994 to 2001 Los Angeles County Social Surveys (LACSS); 1994 General Social Survey (GSS). Note: The response option oVering the highest symbolic racism is shown in parentheses. The number and type of surveys on which the means are based are shown in brackets. Means are based on a 0–1 scale where 1 is the highest possible symbolic racism. The mean reflects a simple average across surveys, not weighted for numbers of cases. Where items used a 5-point scale in some surveys and a 4-point scale in other surveys, the means were averaged across surveys. Items purged in later analyses because of content overlap with racial policy items are starred. For item 16, the Los Angeles government was specifically cited in the LACSS. Starred items are vulnerable to a ‘‘common content’’ critique and deleted from ‘‘purged’’ symbolic racism scales (see Table V).

104

DAVID O. SEARS AND P. J. HENRY

Fig. 1.

The symbolic racism model.

Fig. 2. The basic eVects of symbolic racism.

that symbolic racism is the dominant predictor of racially relevant political attitudes (e.g., Bobo, 2000; Hughes, 1997; Sidanius et al., 1999). Nor do they dispute the weakness of Whites’ personal self-interests as explanations for such attitudes (Bobo, 1988). As a result, most have oVered alternative interpretations without new data. These alternative interpretations are often quite plausible a priori, but empirical research is ultimately required to assess their validity. We have identified 12 points of controversy surrounding theory and research on symbolic racism. Some points concern the theory that symbolic racism is a coherent belief system that cleaves the White public down the middle, including the contentions that (1) symbolic racism has been

SYMBOLIC RACISM: A CONTEMPORARY LOOK

105

conceptualized, and (2) measured in inconsistent ways over time, and that (3) it is not a single psychologically consistent construct but reflects multiple and diverse elements. Others challenge the consistency of the eVects of symbolic racism: contentions that (4) diVerent subsets of the symbolic racism belief system have diVerent eVects, (5) that measures of symbolic racism are so similar to the policy preferences it purports to predict that the associations between them are mere tautologies, (6) that Whites now respond to diverse racial policies in terms of their individual merits rather than responding in terms of their common racial content, and that (7) racial prejudice is politically important nowadays only among less educated Whites. Another question is whether or not (8) the origins of symbolic racism truly lie in the blend of anti-Black aVect and traditional conservative values originally specified. Several alternative substantive explanations have been oVered for the apparent eVects of symbolic racism on the grounds that it is not a distinctive psychological construct from other, more traditional constructs, such as that (9) symbolic racism is not really diVerent from old-fashioned racism; (10) that it confounds prejudice with political conservatism, so its association with policy preferences may largely reflect the influence of nonracial political ideologies; or that (11) it merely collects pieces of other psychological constructs under a single rubric, expropriating their explanatory variance, and so failing the principle of parsimony because it is not a necessary addition to ‘‘the usual suspects.’’ Finally, some suggest that (12) symbolic racism merely reflects a threatened sense of White privilege and group consciousness, rather than antagonism toward a racial outgroup. The theory was first proposed over 30 years ago (Sears & Kinder, 1971). These controversies have bubbled to the surface at diVerent points in that time, and there have been responses to some of these points in the past (e.g., Kinder, 1986; Kinder & Sanders, 1996; McConahay, 1986; Sears, 1988, 1994). However, there has been no concerted and comprehensive eVort to address the empirical questions they raise. That has been the primary agenda of a recent program of research (see Henry & Sears, 2002; Sears & Henry, 1999, 2003; Sears et al., 1997, 2000; Tarman & Sears, 2005). The purpose of this article is to enumerate these critiques and controversies, to assemble this new empirical evidence, to formulate a comprehensive assessment of the validity of these critiques, and so to provide an overall appraisal of where things stand now, after over 30 years of research and scholarly conflict. In doing so, we rely only on evidence from general population surveys; specifically, the National Election Studies (NES) and the General Social Survey (GSS), both periodic face-to-face surveys with nationally representative adult samples, and the Los Angeles County Social Surveys (LACSS), an annual random-digit-dial device telephone

106

DAVID O. SEARS AND P. J. HENRY

survey representing adults in Los Angeles. Although most social psychological research is conducted in the laboratory, with college student subjects, that methodological paradigm is more useful for some purposes than others (Sears, 1986). Because our primary interest is in the political thinking of White American adults, we believe it necessary to rely primarily on representative samples of that population. Our essay has two parts. First we assess the empirical evidence on these specific controversies and what it implies for the validity of the theory. Obviously this will be an abbreviated treatment of this research, but we present what we believe to be the best evidence on each point, using data from multiple surveys in each case. As will be seen, the findings tend to replicate quite closely across surveys. Second, we indicate where we believe the evidence is incomplete.

III. The Symbolic Racism Belief System A. INCONSISTENT CONCEPTUALIZATION One critique is that symbolic racism has been conceptualized inconsistently over time (Sniderman & Tetlock, 1986b; also see Bobo, 1988; Schuman et al., 1997; Stoker, 1998). It began life as an inductively derived, largely intuitive account of the most politically potent form of racism outside the South in the post–Civil Rights era. The concept was heavily influenced by whatever items were available in early surveys. With such beginnings, it would not be surprising if it were not sharply and consistently conceptualized. Indeed, on looking back over the printed record, that charge seems true enough, in several ways. For one thing, symbolic racism was originally described as a single construct (Sears & Kinder, 1971). Later it was described variously as having two themes (Kinder & Sears, 1981), or three (Sears, 1988), or four (Henry & Sears, 2002), or even five (Sears et al., 1997). Today it is consistently described as having the four content themes described above (Henry & Sears, 2000; Sears & Henry, 2003; Sears, Henry, & Kosterman, 2000; Tarman & Sears, 2005), but that explicit focus on four essential themes is a relatively recent development. A second set of conceptual inconsistencies stems from the contention that it is a ‘‘blend’’ of anti-Black aVect with conservative traditional values that themselves have no intrinsic link to race (Kinder & Sears, 1981; Sears, 1988). But that ‘‘blend’’ has been defined in diVerent ways over time. Sometimes it is presented as a simple additive combination of those two separate elements,

SYMBOLIC RACISM: A CONTEMPORARY LOOK

107

measured separately (Sears, 1988). At other times it is described as a ‘‘fusion’’ of those two elements in a single attitude that needs to be measured as a unique attitude distinct from anti-Black aVect and individualism, each measured separately (Kinder, 1986; Sears & Henry, 2003; Wood, 1994). This latter conceptualization is both more in harmony with the original theorizing and with current empirical evidence about the origins of symbolic racism, as will be seen below, so it is the standard current usage (see Sears & Henry, 2003). Another question is how this assertion—that symbolic racism is rooted in a blend of anti-Black aVect and traditional values—squares with the assertion that the symbolic racism belief system has four diVerent themes—none of which involve simple negative aVect toward Blacks, and only one of which implicates Blacks’ violations of individualistic values. Are these contradictory assertions about what symbolic racism ‘‘is’’? Or are they assertions about two diVerent aspects of the symbolic racism phenomenon? The current conceptualization of symbolic racism is shown in Fig. 1. There the ‘‘blend’’ of anti-Black aVect and individualism is described as ‘‘Black individualism.’’ It is an underlying psychological construct like ‘‘prejudice’’ or ‘‘stereotype,’’ and by itself it has no particular political content. ‘‘Black individualism’’ would seem to be a construct that has been alive in White America, in force, for a long time. In contrast, the ‘‘symbolic racism belief system’’ is represented as the set of conventional contemporary beliefs that links this psychological construct to the political and social world. It is the political belief system that came into common currency after the demise of the Jim Crow or ‘‘old-fashioned racism’’ belief system, following the Civil Rights revolution of the 1960s. Some data relevant to the more developed model shown in Fig. 1 will be reviewed later in connection with the origins of symbolic racism.

B. INCONSISTENT MEASUREMENT On the second point, the inconsistencies in its measurement over time seem to have been overstated. All four themes described above were present both in the earliest measures of symbolic racism (e.g., see Kinder & Sears, 1981, and McConahay & Hough, 1976) and in the most recent measures of it (e.g., see Henry & Sears, 2002; Sears & Henry, 2003; Tarman & Sears, 2005). Studies using the related concepts of ‘‘modern racism’’ and ‘‘racial resentment’’ have departed from this practice only slightly, using items from three of the four themes, the former omitting only the work ethic (McConahay, 1986; McConahay & Hough, 1976) and the latter omitting only excessive demands (Kinder & Sanders, 1996; Kinder & Mendelberg,

108

DAVID O. SEARS AND P. J. HENRY

2000). However, the specific items used, as well as the balance among the four themes, have varied across studies, depending on the availability of items in any particular study. Our conclusion is that the underlying themes appearing in measures of symbolic racism have remained surprisingly constant over the long period of time involved. That said, consistency in measurement is clearly desirable. To move toward that goal, an updated scale of symbolic racism has been developed, called the Symbolic Racism 2000 Scale (SR2K; see Table III). It covers all four of its major themes, has superior scaling properties, and has been shown to have both discriminant and predictive validity (Henry & Sears, 2002).

C. THE INTERNAL CONSISTENCY OF THE SYMBOLIC RACISM BELIEF SYSTEM A third controversy surrounds the assertion that symbolic racism, as defined and measured above, reflects a single unitary construct rather than the four multiple and diverse themes by which it is operationalized. The four themes might reflect four diVerent attitudes that are only artificially conjoined in a single measure, or they might reflect other subsets of attitudes within the overall package of items. For example, the first two themes have to do with beliefs about the society’s system of social stratification, and the second two with resentments of Blacks’ specifically political activities. Or, the first and third themes reflect attributions external to Blacks, whereas the second and fourth reflect attributions internal to Blacks. However, tests of its internal consistency, replicated across three separate surveys, reveal that it is indeed a coherent construct, with two highly correlated subcomponents focusing on structural and individual attributions for Blacks’ disadvantages (Tarman & Sears, 2005). An example is displayed in Fig. 3. A two-factor attributional model (individual vs. structural attributions for Blacks’ disadvantage) proved to be the best fitting in both confirmatory and exploratory factor analyses, but the diVerence between these two factors is so slight that they seem best interpreted as two quite similar variants of the same underlying psychological construct: they are highly correlated with each other, and with the second-order factor of global symbolic racism; their coeYcients for predicting racial policy preferences are almost identical in all racial policy domains; after estimating the variance in racial policy preferences that one variant contributes, the other adds no meaningful additional unique increment to R2; and their correlations with other attitudes are similar (although some diVer in strength), indicating that they have similar social and psychological origins. Cumulatively, these

SYMBOLIC RACISM: A CONTEMPORARY LOOK

109

findings indicate to us that the symbolic racism belief system is best interpreted as a single logically and psychologically consistent belief system, expressed in terms of two attributional variants that reflect much the same underlying psychological dimension. It does make sense to talk about a consistent symbolic racism belief system, then. The reason for its consistency, presumably, is that it is a substantively meaningful package of beliefs to most White Americans, whichever side they come down on.

IV. Consistency of the Political Effects of Symbolic Racism As indicated earlier, numerous studies have shown that symbolic racism is strongly associated with Whites’ racial policy preferences. Analyses by both supporters and critics of the theory find that its explanatory power typically outweighs that of other important political attitudes, such as ideology, party identification, and attitudes toward the size of the federal government, or of older and more traditional racial attitudes, such as beliefs in Blacks’ genetic inferiority, support for racial segregation, negative stereotypes, or anti-Black aVect (e.g., Bobo, 2000; Hughes, 1997; Kinder & Sanders, 1996; Sears et al., 1997; Sidanius et al., 1992, 1999). Typical examples are shown in Fig. 2.

A. CONSISTENCY ACROSS THEMES OF SYMBOLIC RACISM The fourth controversy concerns the generality of such associations with policy preferences across the four themes of symbolic racism. It is possible that they do not all have the same eVects on policy attitudes. No surveys have included enough items on each theme to permit exact assessment of their separate eVects, but two surveys have provided enough detail to allow comparisons of the most plausible substantive subdivisions. One is the diVerence between the themes referring to structural attributions for Blacks’ disadvantage (denial of discrimination, undeserved advantage) and those referring to individual attributions about Blacks’ own characteristics (work ethic, excessive demands). Another is the distinction between dominant ideology, which refers only to social stratification (including the denial of discrimination and work ethic themes), and political resentment, which focuses instead on Blacks’ demands and treatment (including the excessive demands and undeserved advantage themes). These diVerent subscales are

TABLE III The Symbolic Racism 2000 Scale 1. It’s really a matter of some people not trying hard enough; if Blacks would only try harder they could be just as well oV as Whites. <1> Strongly agree <2> Somewhat agree <3> Somewhat disagree <4> Strongly disagree 2. Irish, Italian, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same. <1> Strongly agree <2> Somewhat agree <3> Somewhat disagree <4> Strongly disagree 3. Some say that Black leaders have been trying to push too fast. Others feel that they haven’t pushed fast enough. What do you think? <1> Trying to push very much too fast <2> Going too slowly <3> Moving at about the right speed 4. How much of the racial tension that exists in the United States today do you think Blacks are responsible for creating? <1> All of it <2> Most <3> Some <4> Not much at all 5. How much discrimination against Blacks do you feel there is in the United States today, limiting their chances to get ahead? <1> A lot <2> Some

<3> Just a little <4> None at all 6. Generations of slavery and discrimination have created conditions that make it diYcult for Blacks to work their way out of the lower class. <1> Strongly agree <2> Somewhat agree <3> Somewhat disagree <4> Strongly disagree 7. Over the last few years, Blacks have gotten less than they deserve. <1> Strongly agree <2> Somewhat agree <3> Somewhat disagree <4> Strongly disagree 8. Over the last few years, Blacks have gotten more economically than they deserve. <1> Strongly agree <2> Somewhat agree <3> Somewhat disagree <4> Strongly disagree Note: The following is the standard procedure for combining the items into a scale: After collecting the data, items 1, 2, 4, and 8 need to be recoded so that a 1 ¼ 4, 2 ¼ 3, 3 ¼ 2, and 4 ¼ 1. Item 3 needs to be recoded so that 1 ¼ 3, 2 ¼ 1, and 3 ¼ 2. For combining the items into a scale, there are several options, ranging from the simplest to the most precise: (1) One could simply add the raw scores together for each item, so that each individual has a score that could range from 8 to 31. (2) To compensate for any missing data, one could average the raw scores. (3) To compensate for the diVerences in the number of response alternatives, one could recode each of the items on a 0 to 1 scale, so for item #3, a 1 ¼ 1, 2 ¼ 0, and 3 ¼ .50, and for the other items the high response is a 1, the next a .66, the next a .33, and the low response is a 0. (This third technique is the one used in Henry & Sears, 2002.) (4) To equate the variability across items, one could create standardized (z) scores for each of the items in the scale, then average the responses.

112

DAVID O. SEARS AND P. J. HENRY

Fig. 3. Two-Factor Model—Dominant Ideology and Political Resentment, 1997 LACSS. Note: The statistical significance of the standardized parameter estimates was evaluated at  ¼ .05 using Robust standard errors. For identification purposes, factor loadings for the Generations of Slavery and Black Welfare items were fixed at 1.0 and the two second-order factors loadings were constrained to be equal.

113

SYMBOLIC RACISM: A CONTEMPORARY LOOK

TABLE IV Correlations of Different Symbolic Racism Scales with Composite Racial Policy Scales 1986 NES

1997 LACSS

.60

.65

Structural (themes 1 and 4)

.54

.58

Individual (themes 2 and 3)

.52

.58

Symbolic racism scales Full scale Structural vs. individual attributional subscales

Dominant ideology vs. political resentment subscales Dominant ideology (themes 1 and 2)

.54

.63

Political resentment (themes 3 and 4)

.52

.60

Source: 1986 National Election Study and 1997 Los Angeles County Social Survey. Note: Entries are Pearson correlations. Positive entries mean that higher symbolic racism is associated with more opposition to policies designed to help Blacks. All variables are coded 0–1. Pairwise deletion of missing values is employed. All coeYcients significant at p < .001. The references to specific themes are to Table III.

not mutually exclusive; for example, the ‘‘denial of discrimination’’ theme falls under both the ‘‘structural’’ subscale and the ‘‘dominant ideology’’ subscale. Table IV presents the correlations of each symbolic racism subscale with composite racial-policy scales (from Tarman & Sears, 2005). It shows that it does not matter which subset of the four symbolic racism themes is employed. All contribute to the same degree to opposition to racial policies.

B. CONTENT OVERLAP BETWEEN MEASURES OF SYMBOLIC RACISM AND POLICY ATTITUDES A fifth criticism is that the symbolic racism items are excessively close in content to the racial policy preferences they are intended to predict (Sniderman et al., 2000). This criticism has taken two forms. First, in some early studies opposition to racial policies such as aYrmative action and busing were themselves included in measures of symbolic racism (though the dependent variables were voting preferences in a biracial mayoralty election rather than policy attitudes; e.g., Kinder & Sears, 1981; McConahay & Hough, 1976). This was criticized for assuming, rather than demonstrating empirically, that opposition to such policies was based in racism (Sniderman and Tetlock, 1986a). It was also confusing because in some studies, anti–aYrmative action and antibusing attitudes were used as measures of

114

DAVID O. SEARS AND P. J. HENRY

symbolic racism and so as independent variables predicting other political attitudes, and in other studies as dependent variables (policy preferences) predicted by symbolic racism. Both points seemed well taken, so such items have not been used to measure symbolic racism for many years (most recently, we believe, by Sears & Citrin, 1982). Nevertheless, these complaints, though now quite dated, continue to be prominently featured in critical treatments of the theory (see Hurwitz & PeZey, 1998; Stoker, 1998; Tetlock, 1994; Thernstrom & Thernstrom, 1997). Another version of this critique involves symbolic racism items that seem ‘‘too close’’ in content to the policy attitudes that serve as the dependent variables they are intended to predict. If that is the case, correlations between them might reflect mere tautologies rather than the causal eVect of symbolic racism. For example, one standard symbolic racism question has asked whether or not Blacks get more attention from government than they deserve. Any correlation between responses to that item and evaluations of race-targeted government programs could simply reflect two expressions of the same attitude (e.g., Hurwitz & PeZey, 1998; Sniderman et al., 2000). However, on closer inspection this point is only occasionally relevant to the political eVects of symbolic racism. First of all, relatively few symbolic racism items are vulnerable to this critique—only four of the 16 items in Table II are (they are starred). Second, purging items from measures of symbolic racism that refer to possible or actual government action (starred in Table II above) scarcely changes their association with racial policy preferences at all, as shown in Table V. Furthermore, limiting the measurement of symbolic racism to the two wholly nongovernmental themes—denial of discrimination and beliefs in Blacks’ poor work ethic—does not reduce its explanatory power at all. Table IV shows that a ‘‘dominant ideology’’ scale has just as potent a predictive power as the wholly political ‘‘political resentment’’ subscale (Tarman & Sears, 2005). Finally, if the ‘‘tautological’’ explanation of the eVects of symbolic racism were correct, purging items relevant to government should lower the R2 estimate of variance explained (by removing essentially tautological items from both sides of the equation). As can be seen in Table V, the R2 is reduced only very slightly. For all these reasons, this ‘‘tautological’’ critique does not seem to explain the eVects of symbolic racism well (also see Hutchings & Valentino, 2004). The eVects of symbolic racism on racial policy preferences cannot be reduced to a more general hostility to government action in general, but obviously resentment of government attention to Blacks is quite relevant to the symbolic racism phenomenon. Such resentments of Black activism in politics are just as central to symbolic racism as are the other themes.

115

SYMBOLIC RACISM: A CONTEMPORARY LOOK

TABLE V The Relative Impacts of the Full and Purged Symbolic Racism Scales on Whites’ Opposition to Policies Designed to Help Blacks 1986 NES Racial policies Predictors

b

(se)

1997 LACSS Racial policies

b

(se)

b

(se)

b

(se)

Symbolic racism Full scale

.57***

Purged scale

—

(.04) —

—

—

.46***

(.04)

.73***

(.08)

—

—

—

—

.55***

(.08)

Racial attitudes Anti-Black aVect

.13***

(.03)

.15***

(.03)

.08

(.06)

.10

(.07)

Old-fashioned racism

.03

(.03)

.03

(.03)

.07

(.05)

.07

(.05)

Party identification

.04

(.03)

.04

(.03)

.06

(.06)

.09

(.06)

Ideology

.13**

(.05)

.15**

(.05)

.14*

(.06)

.16**

(.06)

Role of government

.08**

(.03)

.07**

(.03)

.05

(.04)

.05

(.04)

Individualism

.05

(.04)

.02

(.04)

—

—

—

—

Antiegalitarianism

.12***

(.03)

.13***

(.03)

.19**

(.07)

.21**

(.07)

Moral traditionalism

.02

(.04)

.04

(.04)

—

—

—

—

.14**

(.05)

.19***

(.05)

.03

(.06)

.14*

(.06)

Political attitudes

Nonracial values

Constant Adjusted R2 (%) Standard error Number of cases

44.6

41.6

.17 769

.17 769

52.0

47.5

.21 257

.22 257

Source: 1986 National Election Study and 1997 Los Angeles County Social Survey. Note: Entries are unstandardized regression coeYcients, with standard errors in parentheses. Positive entries mean that negative Black attitudes, conservative political attitudes, and conservative nonracial values are associated with more opposition to racial policies. Individualism and moral traditionalism not available in the 1999 LACSS. Demographic controls for years of age, levels of education, gender, and Southern region are all included in the equations, but the terms are not shown. All variables are coded (0–1). Pairwise deletion of missing values is employed. *Significant at p < .05, **significant at p < .01, ***significant at p < .001.

Nevertheless, at a methodological level, the criticism seems to us well taken with respect to a few of the symbolic racism items used in the past. As a result, items referring to government action have been eliminated from the new symbolic racism scale referred to earlier (Henry & Sears, 2002).

116

DAVID O. SEARS AND P. J. HENRY

C. CONSISTENCY ACROSS DIVERSE RACIAL POLICIES A sixth controversy concerns the consistency of the symbolic racism eVects across diverse racial policies. If racial animosity is one of the components of symbolic racism, it should form a common racialized basis for opposition to all racial policies. Sniderman and his colleagues have argued instead that the unique politics of each issue (or ‘‘policy agenda’’) dominate its popular reception (Sniderman & Piazza, 1993; Sniderman et al., 2000). Consistent with their view, aggregate White opinion varies a great deal across various racial policies (Sniderman & Carmines, 1997; Steeh & Krysan, 1996; Stoker, 1998; Virtanen & Huddy, 1998). However, the eVects of symbolic racism on policy preferences are remarkably similar across diVerent racial policy domains, as shown in Table VI (also see Sears et al., 1997; Tarman & Sears, 2005). These findings suggest two things. To White Americans, the racial content of all such policies seems to be their most salient feature. The result is that they strongly evoke Whites’ basic predispositions about race.

D. SYMBOLIC RACISM AMONG COLLEGE-EDUCATED WHITES A seventh claim is that racial prejudice only aVects the policy preferences of poorly educated Whites. There is much evidence that education contributes both to reduced racial prejudice and to greater ideological sophistication (e.g., Converse, 1964; Schuman et al., 1997). From this Sniderman and Piazza (1993) took the further step of suggesting that poorly educated Whites’ policy preferences might be more influenced by simple racial prejudice, and those of the well educated by the more complex general political ideology. From our symbolic politics perspective (Sears, 1993), however, we expected that even a moderately complex belief system like symbolic racism was most likely, like ideology, to be as influential or even more so among the college educated. Sniderman and Piazza (1993) only selectively analyzed the eVects of prejudice or ideology on a few racial policies. Our systematic comparison of the eVects of symbolic racism on six racial policy scales in two surveys, measuring all major dimensions of racial policy preferences, found that symbolic racism had, if anything, slightly more influence over Whites’ preferences among the college educated than among those with no college (Sears et al., 1997; also see Federico & Sidanius, 2002a). Greater political information also strengthens the consistency of policy preferences with traditional prejudice (Federico & Sidanius, 2002b). Such findings reflect the more general principle that

TABLE VI Multiple Regression Equations Showing the Origins of Whites’ Opposition to Race-Based Policies Equal opportunity

Federal assistance

1986

1992

2000

Symbolic racism

.40***

.35***

Anti-Black aVect

.11*

.17**

—

.04

.10*

.06

—

—

Ideology Party identification Social welfare

AYrmative action

1986

1992

2000

1986

1992

2000

.29***

.41***

.39***

.43***

.42***

.44***

.34***

.04

.19***

.16***

.12***

.08*

.08*

.09*

—

.08*

.07**

—

.01

.01

—

.04

—

—

.14**

—

—

.02

.01

—

.01

.04

—

.01

.04

.12*

.03

.08

.09*

.03

.05

.02

.08

.07

.03

.06

.05

.07*

.02

.02

.09*

.03

.04

.25***

.15**

.16***

.21***

.22***

.16

.12*

.02

.16***

Racial attitudes

Stereotypes Old-fashioned racism Pro-White aVect Partisanship

Nonracial values Individualism

.09*

—

—

.03

—

—

.03

—

—

Morality/sexuality

.02

.09

.03

.01

.03

.03

.07

.08*

.06

Authoritarianism

—

.06

—

—

.02

—

—

.13**

—

22.3%

40.4%

33.0%

26.4%

Adjusted R2

42.8%

31.1%

37.7%

25.5%

21.3%

Note: Data are from the National Election Studies (NES). Entries are standardized beta coeYcients ( ). *p < .05, **p < .01, ***p < .001. A positive entry means opposition to aYrmative action for Blacks is associated with more negative racial attitudes and more conservative political attitudes and values. The full equations include age, education, gender, and region; these terms are not shown. In this and later tables, a dash (—) indicates a variable not measured in that particular survey (e.g., stereotypes were not measured in 1986).

118

DAVID O. SEARS AND P. J. HENRY

education and political sophistication usually increase the consistency of attitudes toward politically related issues (Converse, 1964; Sears, 1969; Zaller, 1992).

V. Origins of Symbolic Racism An eighth question concerns the hypothesized origins of symbolic racism. The original theory argued that symbolic racism was rooted in a ‘‘blend’’ of two elements—anti-Black aVect and such traditional conservative values as individualism. Until recently, those origins had not been directly empirically examined. Critics thought that approach depicted a ‘‘confounded’’ mixture of prejudice with traditional nonracial values (Sniderman & Tetlock, 1986b). They argued instead for disentangling the two elements and testing them as independent predictors of symbolic racism. When that is done, however, these two attitudes do not provide a very powerful explanation of symbolic racism, regardless of whether they are combined additively or interactively (Hughes, 1997; Sears & Henry, 2003; Sears et al., 1997; Sniderman et al., 2000). A. THE ‘‘BLEND’’ OF ANTI-BLACK AFFECT AND TRADITIONAL VALUES A closer look at the theory indicates that it did not specify clearly whether this ‘‘blend’’ referred to additive or interactive models that treat anti-Black aVect and conservative values as two separate attitudes (e.g., feeling distaste for Blacks and the quite separate belief that a life of hard work has more moral value than a life of leisure) or referred to a ‘‘fusion’’ of the two elements in a single attitude, such as in the belief that Blacks too often violate individualistic values like the work ethic (Kinder, 1986; Kinder & Sears, 1981; Sears & McConahay, 1973; Wood, 1994). These issues have been examined directly only recently (Sears & Henry, 2003). First of all, symbolic racism does prove to have strong components of both racial and nonracial attitudes. Factor analyses using eight surveys that contained traditional measures of racial prejudice, nonracial political attitudes (ideology and party identification), and symbolic racism are shown in Table VII. Prejudice and the nonracial attitudes consistently fell onto separate factors. The two factors were entirely nonoverlapping except that symbolic racism loaded about equally on both factors. This indicates that symbolic racism is the glue linking prejudice to conventional conservatism, consistent with the theory.

SYMBOLIC RACISM: A CONTEMPORARY LOOK

119

Then a direct measure of the fusion of anti-Black aVect and general individualism was created by rewording standard individualism items to focus on Blacks (‘‘Black individualism’’) rather than on people in general (‘‘general individualism’’). For example, the general individualism item, ‘‘If people work hard they almost always get what they want,’’ was changed into a Black individualism item by rewording it, ‘‘If Blacks work hard they almost always get what they want.’’ This rewording measured a fusion of anti-Black aVect and individualism in a single attitude, rather than measuring them in two separate attitudes. This fusion was validated by data showing that target persons high in Black individualism were rated as more unfavorable to Blacks than were target persons low in it. Consistent with the theory, Black individualism proved to have a fairly strong eVect on symbolic racism above and beyond anti-Black aVect and general individualism as measured separately and combined in either additive or interactive fashion (see Table VIII). That is, the ‘‘blend’’ had some emergent properties those separate elements did not. Moreover, Black individualism had the eVects expected of it. It generally explained racial policy preferences more eVectively than did the pooled eVects of racial aVect and general individualism variables measured separately. Table IX shows that in eight equations predicting racial policy preferences, Black individualism had significant and stronger eVects than either anti-Black aVect or general individualism in every case. However, the additive model had something to contribute, as well. Five of 16 of the separate aVect and general individualism terms also reached statistical significance. Finally, Black individualism had the expected strong association with symbolic racism, and its eVects on racial policy preferences were largely (though not completely) mediated by symbolic racism. These latter two findings are depicted in Fig. 4 (the entries are drawn from the appropriate regression equations). Both the fusion and additive models have explanatory value.

B. IS THIS ‘‘BLEND’’ TRULY RACIAL? However previous researchers had suggested that Black individualism might merely reflect general economic individualism without much racial content. Carmines and Merriman (1993) constructed a scale of Black individualism, describing it as ‘‘economic individualism,’’ and found that it did not have much stronger associations with racial policy preferences than did an analogous measure derived from items on women’s individualism. So they concluded that both scales were mainly measuring a more general individualism, or ‘‘classical liberal values,’’ and that respondents were not much influenced by which group was mentioned in the items, whether Blacks or women.

TABLE VII Factor Analyses: Symbolic Racism as a Blend of Traditional Racial Attitudes and Conservation

Symbolic racism

1985

1986

1992

2000

1995

1997

1998

2001

NES

NES

NES

NES

LACSS

LACSS

LACSS

LACSS

1

2

1

2

1

2

1

2

.31

.29

.03

1

2

1

2

1

2

1

2

.48

.26

.34

.47

.37

.54

.49

.45

.47

.48

.47

.66

.50

.43

.40

.55

.07

.07

.63

.10

.63

.13

.66

.09

.59

—

—

.18

.53

Traditional racial attitudes Unintelligent/lack of ability Violent/group stereotype

—

—

—

—

.11

.75

—

—

.17

.56

—

—

.08

.40

—

—

God’s will

.00

.82

.61

.00

—

—

—

—

—

—

—

—

—

—

—

—

Anti-Black aVect

.10

.35

.29

.00

.14

.55

.13

.67

—

—

.10

.29

.25

.53

.27

.60

Ideology

.74

.07

.19

.77

.94

.21

.84

.26

.86

.17

.83

.29

.68

.30

.84

.31

Party identification

.78

.05

.01

.61

.87

.16

.77

.14

.77

.19

.83

.15

.83

.29

.71

Political predispositions

Interfactor correlation

.16

.12

.23

.26

.28

.30

.38

.26 .41

Source: Sears and Henry, 2003. Note: Items are keyed such that higher scores reflect more racial animosity or more conservatism. All loadings greater than .30 are bolded. A dash (—) indicates the variable was not measured in that survey.

121

SYMBOLIC RACISM: A CONTEMPORARY LOOK TABLE VIII Antecedents of Symbolic Racism 1995 LACSS

2001 LACSS

The blend Black individualism

.33***

.32***

.01

.16*

Anti-Black aVect

—

.15*

Stereotypes

.27***

—

.03

.07

Its parts General individualism Traditional racial attitudes

Interaction term Adjusted R2

19.2%

22.6%

Source: Los Angeles County Social Surveys. Note: Columns show standardized regression coeYcients. The ‘‘Interaction Term’’ is a multiplication of General Individualism by Traditional Racial Attitudes. A dash (—) indicates the variable was not measured in that survey. Variables are keyed such that high numbers indicate greater racial antipathy and more individualistic values. *p < .05, ***p < .001.

On close inspection, however, it turns out that these authors’ racial and gender items diVered in a number of respects other than merely race and gender, suggesting the need for a more precise manipulation of that group-specificity in item content. We repeated the comparison more rigorously by substituting the word ‘‘Blacks’’ for ‘‘people’’ in general individualism items to produce a measure of Black individualism, as just described, and then substituting the word ‘‘women’’ for ‘‘people’’ to produce a measure of gender individualism (e.g., ‘‘Any person who is willing to work hard has a good chance of succeeding’’ was changed to ‘‘Any woman who is willing to work hard has a good chance of succeeding’’; Sears & Henry, 2003). Black individualism proved to have considerably stronger eVects on racial policy preferences than did an analogously constructed gender individualism. Moreover, Black individualism had stronger eVects on racial policy preferences than it did on preferences about policies targeted for women. Gender individualism had stronger eVects on policies targeted for women than did Black individualism. So the group-specific component of either form of individualism proved quite central to its eVects. Black individualism is crucial to racial policy preferences, and gender individualism to gender preferences. There are other, more indirect tests of this ‘‘blend’’ notion that yield much the same picture (see Gilens, 1998; Kinder & Mendelberg, 2000; PeZey & Hurwitz, 1998). As a set, all these findings point to a key role for

TABLE IX The Blend is Greater Than the Sum of Its Parts: Consequences for Racial Policy Preferences 1983 NES Federal assistance

Equal opportunity

1995 LACSS AYrmative action

2001 LACSS

Federal assistance

AYrmative action

Federal assistance

Equal opportunity

AYrmative action

The blend Black individualism

.37***

.24*

.38***

.36***

.29***

.27***

.22**

.18*

.09

.01

.04

.08

.09

.03

.20**

.05

Its parts General individualism Traditional racial attitudes Anti-Black aVect Stereotypes Adjusted R2 (%)

.26*

.23*

.19*

—

—

—

24.7

11.0

20.1

— .05 10.7

— .15* 9.0

.11

.02

.06

—

—

—

7.5

11.4

3.6

Source: National Election Studies and Los Angeles County Social Surveys. Note: Columns show standardized regression coeYcients. A dash (—) indicates the variable was not measured in that survey. Variables are keyed such that high numbers indicate greater policy opposition, greater racial antipathy, and more individualistic values. *p < .05, **p < .01, ***p < .001.

SYMBOLIC RACISM: A CONTEMPORARY LOOK

123

Fig. 4. Path analysis of the origins of symbolic racism. Source: Sears & Henry, 2003.

the blend of anti-Black aVect and individualism in forming symbolic racism and, through it, opposition to liberal racial policies. We would agree with Kinder and Mendelberg (2000, p. 60), who, describing Whites’ racial prejudice in the post–Civil Rights era, point to ‘‘individualism within prejudice . . . prejudice [is] preoccupied with black Americans’ specifically individualistic shortcomings’’ (also see Hutchings & Valentino, 2004).

VI. Distinctiveness of Symbolic Racism from Alternative Constructs A variety of alternative constructs have been proposed that might explain the apparent political eVects of symbolic racism. A. OLD-FASHIONED RACISM A ninth controversy concerns the assertion that symbolic racism has replaced old-fashioned racism as the dominant vehicle for Whites’ opposition to racial equality. The skeptics believe that symbolic racism is just old wine in new bottles and really is no diVerent from old-fashioned racism. Empirically, however, that is clearly not true. The two are indeed correlated, but they are distinctive in several ways. The aggregate level of White support for symbolic racism is far greater than it is for old-fashioned racism. As can be seen in Table II, symbolic racism items generally divide the White public in the middle. That is not true at all for the key elements of old-fashioned

124

DAVID O. SEARS AND P. J. HENRY

racism, such as support for segregation, legalized discrimination, or beliefs in the genetic inferiority of blacks. On average these attitudes corral only about 10% of the White public, or less, in the contemporary era (Kinder & Sanders, 1996; Kluegel, 1990; Schuman et al., 1997). Moreover, when entered together in factor analyses with no other variables, old-fashioned and symbolic racism fall out on separate factors. A typical example is shown in Table X, using four diVerent national surveys. Using the standard .30 cutoV threshold for factor loadings, all 16 of the symbolic racism items load on symbolic racism factors, and 18 of the 19 oldfashioned racism items load on old-fashioned racism factors. Only two of the symbolic racism items load on the old-fashioned racism factors, and none of the old-fashioned racism items load on the symbolic racism factors. Parenthetically, which items blend into the other construct is instructive. The theme of Blacks’ insuYcient work motivation is a very old one and is the clearest bridge between the two constructs. The issue of miscegenation is the last vestige of Jim Crow racism to be contested (Schuman et al., 1997), but the larger story here is that the constructs of symbolic and old-fashioned racism are very diVerent. Of course when they are entered into factor analyses with nonracial variables, their common racial content leads them to load on the same factor (see Table X; also Sears & Henry, 2003). Finally, symbolic racism has much greater eVect in today’s racial politics than does traditional prejudice. At a bivariate level, the latter sometimes has a significant eVect (e.g., Federico & Sidanius, 2002b; Sears et al., 1997; Virtanen & Huddy, 1998). But in multiple regression equations, it is overshadowed by symbolic racism. For example, the most comprehensive comparative study estimated regression equations using four surveys and three dependent variables, using both symbolic and old-fashioned racism among the predictors (Sears et al., 1997). The average standardized regression coeYcient for symbolic racism was .41, and all were statistically significant. The average coeYcient for old-fashioned racism was .01, and only one yielded a significant contribution to opposition to racial policy.

B. POLITICAL IDEOLOGY A tenth critique is that symbolic racism confounds prejudice with political conservatism (Fazio et al., 1995; Hurwitz & PeZey, 1998; Sniderman & Tetlock, 1986a,b; Tetlock, 1994). The eVects of symbolic racism on racial policy preferences could then be caused either by racial prejudice or by unprejudiced conservatives’ aversion to big government, which would have very diVerent implications.

TABLE X Factor Analyses of Symbolic and Old-Fashioned Racism Items

126

DAVID O. SEARS AND P. J. HENRY

To be sure, ideological conservatism is correlated with both symbolic racism and racial policy preferences (Hughes, 1997; Sears, 1994; Sears et al., 1997; Sidanius et al., 1992), and symbolic racism does prove also to have both strong racial and nonracial components. As shown earlier, in Table VII, symbolic racism loads strongly on both racial and nonracial factors in factor analyses of eight surveys. However, symbolic racism and its eVects on racial policy preferences cannot be reduced to a nonracial ideological conservatism. This can be seen in four ways. First, it invariably has far greater explanatory value than ostensibly race-neutral conservatism when predicting racial policy preferences, as shown in Fig. 2 and Tables V and VI (also see Henry & Sears, 2002; Sears et al., 1997; Tarman & Sears, 2005). Second, these eVects of symbolic racism are scarcely diminished at all when ideology is controlled, as shown in a number of studies (see especially Sears et al., 1997; Tarman & Sears, 2005). Third, if the explanatory power of ideology has been underestimated because its variance has been ‘‘expropriated’’ by similarly worded symbolic racism items, purging the symbolic racism measures of references to policy or to government should reduce its political eVects, but it does not, as we have seen (Table V). Fourth, such purges do not enhance the explanatory power of ideology, as it should if such purging allowed its variance to flow back to its ‘‘proper’’ source. For these reasons, then, we believe that the eVects of symbolic racism cannot be explained as mainly being eVects of nonracial ideological conservatism (though one recent experiment suggests symbolic racism and ideology may sometimes interact; see Feldman & Huddy, 2005, as well as Table I).

C. OTHER CONSTRUCTS However, ideology and symbolic racism (as well as traditional racial prejudice) contribute overlapping variance to the prediction of racial policy preferences. How is that to be interpreted? The eleventh claim is that symbolic racism may just be a construct overlapping other such traditional constructs and does not contribute anything new. To the contrary: Symbolic racism contributes unique variance to explaining racial policy preferences that other variables cannot. Moreover, it is a distinctive construct in its own right. On the first point, the most demanding test of the role of symbolic racism asks whether it has significant unique eVects on racial policy preferences beyond the eVects of such other variables. Operationally, this can be tested by entering symbolic racism as the final stage in a hierarchical regression after both putatively nonracial attitudes (such as ideology and party identification) and traditional racial attitudes have been entered.

127

SYMBOLIC RACISM: A CONTEMPORARY LOOK

TABLE XI Variance in Opposition to Racial Policies Accounted For by Racial and Nonracial Attitudes: Hierarchical Regressions First stage

Last stage

All variables

Racial attitudes

Nonracial attitudes

Symbolic racism alone

Nonracial attitudes

1986 NES

44.8%

33.8%

24.8%

9.9%

9.6%

1992 NES

32.7

26.4

16.3

7.5

5.2

Mean

38.8

30.1

20.6

8.7

7.4

1986 NES

41.5

32.6

18.3

10.4

7.3

1992 NES

38.6

32.6

16.6

9.6

5.6

1994 GSS

29.4

23.8

9.1

15.1

2.2

Mean

36.5

29.7

14.7

11.7

5.0

1986 NES

27.6

21.8

11.2

11.3

3.5

1992 NES

26.6

21.5

9.9

12.1

2.4

1994 GSS

20.8

18.6

3.9

11.6

0.1

1995 LACSS

25.5

17.9

12.8

7.3

4.7

25.1

20.0

9.5

10.6

2.7

Equal opportunity

Federal assistance

AYrmative action

Mean

Source: Adapted from Sears et al. (1997). Note: Entries in column 1 are the R2 when all variables are considered; in column 2, the variance explained by both symbolic racism and traditional racial attitudes when entered as the first stage; in column 3, the R2 explained by all nonracial attitudes when entered as the first stage; in column 4, the increment in R2 when symbolic racism is added as a final stage after demographics, traditional racial attitudes, and nonracial attitudes are considered; in column 5, the increment in R2 when nonracial attitudes are entered in the last stage. The R2 for all variables shown in column 1 is not adjusted for the number of variables in each equation.

The most comprehensive such study used preferences about nine racial policies in four diVerent surveys (Sears et al., 1997). Table XI shows a series of hierarchical regressions to test this hypothesis. The first column shows the R2 from regressions that include all variables, including symbolic racism, traditional racial attitudes, partisan attitudes, nonracial values, and demographic variables (the variables are the same as those in Table VI above; for the specific variables in each survey, see Sears et al., 1997, Tables I–III). Columns 2 and 3 compare the R2 for the first stages of regressions that include either all those racial attitudes or all those nonracial attitudes. As can

128

DAVID O. SEARS AND P. J. HENRY

be seen, racial attitudes are, cumulatively, considerably more powerful than all the nonracial attitudes together. Most important, however, is the fourth column, which shows the incremental variance explained by symbolic racism when added as a final stage after all other variables have been considered by the prior stage. As can be seen, it adds more than 10% in variance explained, on average, and indeed quite regularly across surveys. The final column shows that nonracial attitudes do not explain nearly as much variance when they are added as a final stage, beyond symbolic racism and all other variables, averaging under 3% in explained variance. Whatever the overlaps in explained variance between symbolic racism and other variables, then, symbolic racism proves to be the most important contributor of uniquely nonshared, explained variance. Symbolic racism consistently added about 10% in unique explained variance when entered even after all the usual other nonracial and racial attitudes had already been considered (Sears et al., 1997). But is symbolic racism really a distinctive construct in its own right? Another set of analyses has used structural equation models to determine whether symbolic racism is an independent psychological construct or merely draws elements from other more traditional constructs, and so could easily be abandoned on the grounds of parsimony. If this latter view were correct, presumably models that did not separate out symbolic racism as an independent construct but that distributed symbolic racism items into other, more traditional constructs, would fit the data just as well as a model that identified symbolic racism as a separate, distinctive construct. In two NES surveys we compared the symbolic racism model with four alternatives that distributed the symbolic racism items into other factors (Tarman & Sears, 2005). We found that symbolic racism, old-fashioned racism, conservatism, and nonracial individualism all constituted separate and independent factors. Models that distributed symbolic racism items among other constructs did not fit the data nearly as well. Symbolic racism is an independent construct (though correlated with the others, of course) and a necessary and key component to explaining Whites’ contemporary racial attitudes as they apply to racial politics.

VII. Outgroup Antagonism or White Group Consciousness? A twelfth and final set of questions concern alternative explanations that assume the importance in racial politics of Whites’ group consciousness, centered on their racial identity as Whites. Social structural theories, such as those of realistic group conflict (Bobo, 1983), sense of group position (Bobo,

SYMBOLIC RACISM: A CONTEMPORARY LOOK

129

1999), and social dominance (Sidanius & Pratto, 1999), argue that Whites oppose liberal racial policies because they would jeopardize the racial hierarchy that privileges them and their fellow Whites. Symbolic racism, in this view, is just an ideology that legitimizes White privilege. Such theories focus more on aYnity to the ingroup, in the form of White group consciousness, than outgroup antagonism, such as in symbolic racism, as the engine of White resistance to change. They especially focus on Whites’ senses of self interest, group interest, or senses of collective racial threat as driving motives for opposition to greater racial equality. We have pursued these alternatives. Our evidence, at least, seems to be inconsistent with the main thrust of these alternative conceptions.

A. WHITES’ SELF-INTEREST Among these various possibilities, Whites’ personal self-interest is the factor that has been subject to the most empirical tests from the beginning of research on symbolic racism. This research has shown that self-interest, as operationalized by racial threats to Whites’ personal lives, is rarely a significant factor in White opposition to racial policies. That is true for selfinterest operationalized in either objective terms (e.g., having children in public schools in a district with busing for racial integration) or subjective terms (e.g., perceiving it likely that Black families would be moving into one’s neighborhood). The evidence seems to be convincing to supporters and critics alike (for the former, see Kinder & Sears, 1981; Sears & Kinder, 1971; for the latter, see Bobo, 2000; Sniderman & Tetlock, 1986a; for comprehensive summaries, see Citrin & Green, 1990; Sears & Funk, 1991).

B. WHITES’ SENSE OF GROUP IDENTITY AND GROUP INTEREST Whites’ sense of group interest is another matter. To test its eVects would require direct measures of the psychological constructs involved. For Whites to act politically on the basis of their sense of shared group interest as White people, one would think that several subjective states would be relevant: strong White group identity, perceived common fate with other Whites about valued resources, perceived threat or competition for scarce resources between Blacks and Whites, and perceived collective threat from Black people to Whites’ well-being (Sears & Kinder, 1985). Such direct measures of a sense of White consciousness have not been common in the empirical

130

DAVID O. SEARS AND P. J. HENRY

work of proponents of these theories (for exceptions, see Bobo & Hutchings, 1996; Bobo & Johnson, 2001). Our research on this topic indicates that Whites’ political attitudes only unevenly fit the portrait of a dominant ingroup that is self-consciously devoted to maintaining its hegemony over minority outgroups. Neither symbolic racism nor racial policy preferences is as much influenced by White group consciousness as by antagonism toward Blacks. For example, in the 1985 and 1986 NES surveys, as shown in Table XII, an ingroup identity as ‘‘White,’’ self-categorizing as White, or liking Whites in general seems not to contribute to more conservative preferences about racial policy. Nor did individual Whites’ perceptions of common fate with other Whites or their sense of personal threat from minorities seem to have much political eVect. The one variable central to group interest theories that did have consistent eVects among Whites was the perception of zero-sum racial conflict: Whites who perceived more Black–White conflict were also more opposed to policies that might benefit Blacks. In contrast with this generally null eVect of group-interest, symbolic racism had consistently strong political eVects. This suggests that it is resentment of Blacks, not feeling protective of fellow Whites, that drives opposition to race-targeted policies. We also had relevant measures in five Los Angeles County Social Surveys. We had three measures of strength of White identity (‘‘How strongly do you identify with other White people?’’ ‘‘How important is being White to your identity?’’ and ‘‘How often do you think of yourself as a White person?’’). We also used feeling thermometers toward ‘‘Whites’’ and, for comparison, toward ‘‘Blacks.’’ In one survey we measured a sense of common fate with other Whites (‘‘How much do you think that what happens generally to White people in this country will have something to do with what happens in your life?’’). None of these indicators of White group consciousness had any eVects: The strength of identification as ‘‘White,’’ evaluations of Whites, and a sense of common fate with Whites were scarcely related to symbolic racism at all, as shown in Table XIII. They had an average correlation of r ¼ .07 with symbolic racism, and only 2 out of 11 correlations were even statistically significant. The Black feeling thermometer was more strongly correlated with symbolic racism than were any of these indicators of White group consciousness in all five surveys. The average correlation was a noticeable, if not enormous, r ¼ .22 for the Black feeling thermometer, and most of its correlations with symbolic racism were significant. White group consciousness had this same lack of eVect over opposition to racial policies, as shown in Table XIV, again using the five Los Angeles County surveys. As can be seen, symbolic racism has consistently strong eVects, with an average regression coeYcient of .43 over five surveys, statistically significant in each case. Evaluations of Blacks have more modest

131

SYMBOLIC RACISM: A CONTEMPORARY LOOK TABLE XII Predicting Whites’ Opposition to Racial Policy 1985

Variable

1986

r

Regression

r

Regression

1

2

3

4

Blacks as attitude object Symbolic racism

.52***

.426*** (.049)

.60***

.503*** (.048)

Dislike Blacks

.27***

.185*** (.065)

.31***

.096* (.039)

Dislike Black militants

.19***

.108 (.043)

.31***

.101 (.035)

Like Whites

.04

.073 (.035)

—

—

White identity

.05

N/A

—

—

White self-categorization

.02

N/A

—

—

.41***

.189*** (.041)

.48***

.157*** (.035)

Black gains hurt self

.24**

.051 (.037)

.32***

White gains help self

.02

.068 (0.040)

—

Whites as attitude object

Perceived group conflict Black gains hurt Whites Self interest

R2

.42

.017 (.030) — .43

Source: 1985 and 1986 American National Election Studies (adapted from Jessor, 1988; Sears & Jessor, 1996). Note: Regressions include controls for age (in years), education (in years), region (South vs. rest of country), subjective social class identification (middle class vs. working class), and employment problems. Entries in columns 1 and 3 are Pearson correlation coeYcients. Entries in columns 2 and 4 are unstandardized regression coeYcients; entries in parentheses below are standard errors. White identity and self-categorization are not included in the regression equation in column 2. All scales coded 0 to 1, with racially conservative position high (weighted n ¼ 333 in 1985; n ¼ 534 in 1986). *p < .05, **p < .01, ***p < .001.

132

DAVID O. SEARS AND P. J. HENRY TABLE XIII Ingroup vs. Outgroup: Correlations of Social Structural Variables with Symbolic Racism Los Angeles County Social Surveys 1997

1998

1999

2000

2001

.12

.35**

.17**

.15

.31**

Outgroup attitudes Anti-Black aVect Ingroup attitudes White ethnic identity

.07

.14*

.15**

.10

.06

Pro-White aVect

.00

.06

.10

.03

.07

Common fate with Whites

—

—

—

—

.10

Source: Los Angeles County Social Surveys. Note: Entries are Pearson r correlations. *p < .05, **p < .001.

eVects, averaging .06 and with only one significant. The three indicators of a sense of White group consciousness, however, have almost no eVect at all. White ethnic identity yields an average coeYcient of .04; evaluations of Whites, .06; and sense of common fate with other Whites, .08. Only 1 of the 11 coeYcients is significant in the expected direction, and 2 in the opposite direction (those with stronger White identity are more supportive of racial policies).3 Symbolic racism, and opposition to racial policies, mainly reflect resentment of Blacks and what are seen as their misdeeds; they have little to do with any sense of White consciousness or defense of White privilege on the part of individual Whites.

VIII. Summary We have taken up 12 controversies surrounding symbolic racism. We agree that the conceptualization of symbolic racism has not been terribly clear or consistent over time, though hopefully that has now evolved to a 3 It should be noted that these weak coeYcients for white group consciousness are not artifacts of collinearity with symbolic racism; as shown in Table XIII their correlations with symbolic racism are modest, and as shown in Table XIV, their average bivariate correlation with opposition to racial policies is actually r ¼ .00.

TABLE XIV Outgroup vs. Ingroup: Symbolic Racism and White Group Consciousness as Predictors of Whites’ Racial Policy Preferences 1997

1998

1999 Beta

2000

r

Beta

Symbolic racism

.66

.49**

.56

.45**

.52

.42**

.49

.35**

.51

.45**

Anti-Black aVect

.19

.16*

.09

.04

.21

.00

.11

.25

.17

.00

.07

.15*

.09

.01

.09

.00

.12

.06

.14

.13*

Pro-White aVect

.08

.03

.14

.10

.18

.13

.03

.27*

.00

.04

Common fate with Whites

—

—

—

—

—

—

—

—

.00

.08

Ideology

.53

.20*

.44

.16*

.38

.17*

.37

.15

.42

.03

Party identification

.52

.20*

.49

.22**

.30

.07

.30

.16*

.34

.12

r

Beta

2001

r

Beta

r

Beta

r

Outgroup attitudes

Ingroup attitudes White ethnic identity

Partisan attitudes

Adj. R n

2

55.6

43.1

31.4

32.2

30.9

228

282

290

169

231

Source: Los Angeles County Social Surveys. Note: Only White respondents are included. For each year, left-hand column reflects a Pearson r of predictor with the racial policy scale; righthand column reflects the standardized regression coeYcient. Regression equations also included age, education, and gender; those coeYcients are not shown. Policy opposition and anti-Black, pro-White, and conservative attitudes are keyed as high. *p < .05, **p < .001.

134

DAVID O. SEARS AND P. J. HENRY

clear and consistent conceptualization. The measurement of symbolic racism has centered on a reasonably constant set of content themes, although the specific items used have varied considerably over time. It seems to form a substantively meaningful and statistically consistent belief system; it has two variants that diVer slightly in the language they use but that are extremely highly correlated and have almost identical eVects. The political eVects of symbolic racism are quite consistent regardless of how it is measured. It has strong eVects on racial policy preferences that are not weakened if items thought ‘‘close’’ in content to those dependent variables are removed from the measure of symbolic racism. Indeed, it seems not to matter at all which conventional measure of symbolic racism is used; they all have approximately the same eVects. It has remarkably similar eVects on preferences about quite diverse racial policies—even those that diVer substantially in the overall level of White support they attract. Its eVects on racial policy preferences are just as strong among the college educated as they are among the less educated. Its origins do appear to lie in a ‘‘blend’’ of racial animosity and conservative traditional values such as individualism, combined in either additive fashion or in the fusion we called ‘‘Black individualism.’’ Symbolic racism seems to mediate the eVects of that ‘‘blend’’ on racial policy preferences, as would be the case if symbolic racism were the contemporary political belief system that gives it political power. And Black individualism is distinctly racial: It has stronger eVects on racial policy preferences than on gender policy preferences, whereas the reverse holds for an analogous gender individualism. Symbolic racism is quite diVerent from old-fashioned racism. It is much more widely accepted, has far stronger political eVects, and falls on diVerent factors in factor analyses, though their common racial content generates significant correlations between them. Its political eVects are quite distinct from and stronger than those of political conservatism, though correlated with it. Its eVects do not seem to stem from any very self-conscious defense of White privilege, because a variety of measures of White group consciousness have little to do with either symbolic racism or opposition to policies targeted for Blacks. Symbolic racism is all about antagonism toward Blacks, not ingroup favoritism for Whites. In all these respects the theory seems to have been sustained quite well by available empirical evidence, contrary to the speculative alternative interpretations alluded to initially. The data we have presented have been collected in a wide variety of surveys. None of our basic findings seem to be specific to any one survey; rather, they replicate strongly over surveys. In our view these particular critiques ought to be set aside and these controversies regarded as settled unless and until contrary evidence emerges (see also Hutchings & Valentino, 2004).

SYMBOLIC RACISM: A CONTEMPORARY LOOK

135

IX. Further Questions There are a number of other questions that might reasonably be raised about symbolic racism. In these cases we appraise the evidence as less clear. We have seven in mind.

A. A LIFE-SPAN APPROACH TO THE ORIGINS OF SYMBOLIC RACISM To start with, we want to qualify our assertion that the ‘‘origins’’ of symbolic racism lie in a blend of anti-Black aVect and individualism. There are two distinctively diVerent ways to think about the ‘‘origins’’ of any attitude. One is its grounding at any given point in time in more basic or fundamental psychological constructs, such as values or basic attitudes or personality predispositions or social identities. That is what we tested above, grounding symbolic racism in ‘‘Black individualism.’’ The other approach locates its origins in a developmental or historical process across the individual’s life course. In our case this would involve a causal model about the origins of symbolic racism within the individual’s own life history. These two kinds of explanations of the origins of symbolic racism are not the same, and they are probably not best tested with the same data. That is, the evidence presented above is based on cross-sectional data on adult samples and is quite appropriate for examining the links between symbolic racism and presumably more fundamental values and attitudes. However, cross-sectional data on adult samples are not very useful for testing a lifehistory model. Either longitudinal evidence on adults or cross-sectional data on preadults would be more appropriate. Thus, we have addressed only the first approach to the ‘‘origins’’ of symbolic racism. We could, however, speculate about what such a life-history model might look like. First, most early-acquired attitudes toward social groups seem to be primitive and cognitively rudimentary (e.g., Aboud, 1988; Lambert & Klineberg, 1967; Sears & Levy, 2003). They involve little more than a sense of similarity/dissimilarity and some favorable or unfavorable feeling (indeed, such primitive aVects might result ultimately in adults’ ‘‘implicit prejudice’’; e.g., Devine, 1989; Dovidio et al., 1997). Values such as the work ethic or morality or equality presumably are acquired a little later, perhaps usually initially with cognitive referents to the child’s world rather than to the broader society. As children become aware of the political world, in late childhood and in adolescence, they are exposed to political and social rhetoric from others.

136

DAVID O. SEARS AND P. J. HENRY

That rhetoric normally includes prepackaged, oV-the-shelf beliefs about society such as that Blacks violate individualistic values. The earlier socialization might well provide some fertile ground when these relatively simple abstract assertions about social groups are encountered, perhaps in adolescence. The result may be that this ‘‘Black individualism’’ might itself become an underlying ‘‘symbolic predisposition’’ in the language of symbolic politics theory (e.g., Sears, 1993). Symbolic predispositions generally facilitate the acceptance, through standard cognitive processes, of more complex political ideas that pass into and out of the political firmament later on in life. Among these would be the several concrete political beliefs held by racial conservatives that today form the symbolic racism belief system. We might speculate, given the high visibility of racial issues in our society, that young people are exposed to the symbolic racism belief system, and acquire attitudes about it, in late adolescence and early adulthood, when the individual’s attitudes about highprofile political attitude objects tend to become crystallized (Alwin, Cohen, & Newcomb, 1991; Jennings & Stoker, 1999; Miller & Sears, 1986). Examining the cross-sectional grounding of political attitudes in fundamental predispositions and tracing the life-span development of those attitudes are thus two quite diVerent ways of explaining ‘‘origins.’’ They are diVerent, rather than contradictory, models of analysis. We have evidence that symbolic racism is grounded, among adults, in a ‘‘blend’’ of anti-Black aVect and individualism. We have no direct evidence on this life-history model. The contrast is important, because it represents a key respect in which symbolic racism theory departs from social structural theories. These latter postulate that Whites identify with their racial group, perceive their privileged position as a member of it, desire to protect that privilege, and as a result adopt ideologies that legitimize and justify that inequality. For example, social dominance theory views symbolic racism as a ‘‘legitimizing myth’’ that helps to rationalize and legitimate that racial order hierarchy and the resulting structural inequalities (Sidanius et al., 1999). That would be a complex cognitive process that would probably not be very common among preadults. In contrast, the symbolic racism account looks for the historical roots of symbolic racism in preadult socialization. In our view, such conflicting views cannot be resolved by tinkering with the causal ordering in structural models based on cross-sectional data. They will require longitudinal data, or at least cross-sectional data at diVerent ages, that will allow for the life-span causal ordering to be sorted out. Surveys of adults, such as those we have employed here, collected decades after their preadult years, on average, would be poor bases for assessing either the acquisition of anti-Black aVect and traditional values in childhood

SYMBOLIC RACISM: A CONTEMPORARY LOOK

137

or their role in promoting acceptance of Black individualism, perhaps in adolescence. Indeed our adult data may in fact only pick up the residues of the process by which Black individualism promotes the acceptance of the symbolic racism belief system, if that occurs and becomes crystallized as the preadult emerges into adulthood. Presumably our cross-sectional data are more relevant for examining the process by which symbolic racism in turn influences responses to the various political issues that come onto the political agenda through the individual’s adulthood. B. BEYOND INDIVIDUALISM We find that the blend of anti-Black aVect and general individualism makes a substantial contribution to explaining symbolic racism, and in turn, symbolic racism mediates its eVects on racial policy preferences (Sears & Henry, 2003). So we take this as an instance of the more general proposition that symbolic racism is grounded in both racial prejudice and ostensibly nonracial traditional conservative values. Individualism has always been the most prominently mentioned of these values, but is it the only one, or even the most important? Black individualism by no means explains all the variance in symbolic racism. We think a case can be made for including traditional morality as well (Sears et al., 1997), and perhaps one also can be made for religious conservatism (Sears & Valentino, 2002). All of this is to say that there may be a richer story yet to tell about the value side of the determinants of symbolic racism, as Kinder and Sanders (1996) believe. Racism has permeated American society for over 350 years, and it has wormed its way into the warp and woof of many of the ways we all think about society. We doubt that Black individualism exhausts that penetration. C. VALUING EQUALITY VS. SOCIAL ‘‘DOMINANCE’’ Historians and other social observers have long noted the importance of Americans’ responses to the idea of ‘‘equality,’’ from its invocation by JeVerson in the Declaration of Independence to Lincoln at Gettysburg, and from de Tocqueville to modern social scientists such as Lipset (1996) and Pole (1993). Today all Americans support to some degree the general idea of equality, although there are substantial cleavages in degree of support for general egalitarian values (Federico & Sidanius, 2002b; Feldman, 1988; Sears et al., 2000). Individual diVerences in support for general egalitarian values are quite central to contemporary partisan divisions (Miller & Shanks, 1996).

138

DAVID O. SEARS AND P. J. HENRY

Some social structural theories of racial stratification in the United States, particularly realistic group conflict theory and sense of group position theory, posit a standing racial hierarchy, with the dominant racial group, Whites, exercising a strong vested interest in their privileged position. As a result, Whites they should be generally more supportive of that hierarchy and opposed to policies that threaten it than subordinate groups should be (e.g., Bobo, 1983, 1988, 1999). Another empirical implication is that the strongest White supporters of such a racial hierarchy—that is, either general inegalitarians or racial inegalitarians—ought also to be the strongest opponents of policies that threaten that hierarchy. Other social structural theories focus on variations in support for equality in all racial and ethnic groups. Social dominance theory (Sidanius & Pratto, 1999) argues that people vary in their support for group-based hierarchies as a general feature of social stratification. Each society may diVer somewhat in the exact group hierarchy that is most politically pivotal. In America, the racial hierarchy is viewed as very important. However, the measure of that individual diVerence dimension, ‘‘social dominance orientation,’’ is neutral with respect to which groups are involved; it refers to ‘‘groups’’ in general without further specification. How does the theory of symbolic racism deal with the question of general values about equality? The original theory was largely silent on the role of egalitarian values in racial politics. It did assert that the opposition to racial equality, as expressed in old-fashioned racism, was rapidly waning in the 1960s and thereafter and therefore was unlikely to continue to have a pivotal political role, especially outside the South (Sears & Kinder, 1971). As anticipated, opposition to general principles of racial equality has continued to drop. It is now almost nonexistent outside the South and is comparatively rare even there. In the language of Schuman et al. (1997), the White public today shows generally consensual support for ‘‘general principles of racial equality.’’ Beyond the question of race, a long strain in American social science has emphasized the political importance of variations in Americans’ commitment to ‘‘equality’’ as a general value. Lipset (1996) makes it a centerpiece of ‘‘American exceptionalism’’ and argues that as White Americans’ support for equality in general waxes and wanes, so does their support for racial liberalism. Much research has reported significant eVects of general egalitarian values on support for liberal racial policies (Federico & Sidanius, 2002b; Kinder & Sanders, 1996; Sears, 1988; Sears et al., 2000). What are the roles of attitudes toward equality in general, as a general principle, in Whites’ responses to racial issues today? There are two measures of egalitarian values in common use today. One is the scale developed by Feldman (1988) for the National Election Studies, and the other is the measure

SYMBOLIC RACISM: A CONTEMPORARY LOOK

139

of social dominance orientation developed by Sidanius and Pratto (1999) and others. We have more data on the latter, so let us turn to it.4 Social dominance orientation is intended to be a coherent, singledimensional construct. In practice we find that it has two rather weakly correlated subcomponents, at least among ordinary adults (Sears et al., 2000; but perhaps not among college students; see Jost & Thompson, 2000). The first we have labeled as ‘‘less equal treatment.’’ It seems to reflect the general value of group equality (e.g., ‘‘group equality should be our ideal,’’ or ‘‘it would be good if all groups could be equal’’; the questions do not specify which groups are involved). The second, ‘‘dominance,’’ reflects support for a more malignant forced group inequality (e.g., ‘‘sometimes other groups must be kept in their place,’’ or ‘‘superior groups should dominate inferior groups’’). To assess the correlation between the two components, we have analyzed five general population surveys that contained measures of both, the 1997–2001 LACSS surveys. The correlations among Whites ranged from r ¼ .08 to r ¼ .23, with a mean (and median) of .14. Two of the five are not statistically significant. This leads us to have some confidence that the general White population responds to the two subcomponents as two diVerent constructs, not as closely intertwined aspects of the same construct. The two subcomponents of social dominance orientation also prove to have strikingly diVerent bivariate associations with racial policy preferences, as shown in Table XV. Inegalitarianism is quite closely linked to opposition to liberal racial policies, whereas dominance is not. The diVerence in strength of the correlations is striking in each survey. The average correlations with racial policy preferences, across five surveys, are r ¼ .39 and .11, respectively. Two of the latter are not even statistically significant. When included in regression equations with symbolic racism, the slight eVects of ‘‘dominance’’ disappear altogether; in no case does it significantly aVect racial policy preferences. Inegalitarianism does have residual significant eVects in each case, although they are substantially weaker than the eVects of symbolic racism. In short, symbolic racism again is the construct that best helps us explain contemporary White opposition to race-targeted political alternatives. General inegalitarianism, as a presumably general and fundamental value, con4 Sears et al. (2000) also argued that the NES egalitarianism scale (Feldman, 1988; Miller & Shanks, 1996) was also composed of two components, one that is close to general egalitarian values (‘‘less equal treatment’’) and the other, closer to continuing animosity to blacks (‘‘we’ve gone too far’’). They found that the latter was considerably more closely linked to symbolic racism in the 1985 and 1986 NES studies. We find the same is true in the 1992 and 2000 NES studies as well; in regressing symbolic racism on them, the standardized coeYcients for ‘‘we’ve gone too far’’ are .42 and .36, and for ‘‘less equal treatment,’’ .18 and .20. All coeYcients are statistically significant.

140

DAVID O. SEARS AND P. J. HENRY TABLE XV Symbolic Racism and Social Dominance as Predictors of Whites’ Racial Policy Preferences 1997

1998

Beta

2000

r

Beta

r

.65

.55**

.56

.41**

.52

.40**

.49

Inegalitarianism

.45

.16**

.41

.20**

.35

.18**

Dominance

.16

.05

.12

.13*

.15

.05

Symbolic racism

Beta

1999 r

r

Beta

.34**

.51

.42**

.34

.24**

.40

.26**

.01

.11

.12

.03

r

Beta

2001

Social dominance

Adj. R

2

59.9%

45.9%

32.8%

36.3%

35.0%

Source: Los Angeles County Social Surveys. Note: Only White respondents are included. For each year, left-hand column reflects a Pearson r of predictor with the racial policy scale; right-hand column reflects the standardized regression coeYcient. ‘‘Inegalitarianism’’ and ‘‘dominance’’ are the two separate subcomponents of social dominance orientation. Models also included ideology, party identification, age, education, and gender; those coeYcients are not shown.

tinues to have some resonance in its own right. This indicates that the American case may be best understood in terms of Lipset’s (1996) focus on historic debates over the trade-oV of equality against values of liberty and freedom. Although the kind of aggressive ‘‘dominance’’ of subordinate groups by privileged groups is the most distinctive and direct index of the process that social dominance theory focuses on, it recedes somewhat to the background, seemingly peripheral to contemporary racial politics. Perhaps this is because the language of ‘‘dominance’’ is interpreted by contemporary White Americans as much the same as the widely repudiated Jim Crow racism (Sears et al., 2000). This is also suggested by the earlier data showing that White group consciousness seems to have very little to do with how Whites approach divisive racial issues. That is, the more authoritarian, aggressive, actively subordinating element of social dominance orientation seems to have little to do with Whites’ opposition to policies that would increase racial equality. These data do not directly address the principal assertion of social dominance theory about symbolic racism—that it is a mere ‘‘legitimizing myth’’ justifying maintenance of the existing racial hierarchy. The evidence for that assertion is that symbolic racism can statistically mediate the eVects of social dominance orientation on racial policy preferences, if assumed to be more proximal to those dependent variables (Sidanius et al., 1992, 1999). However, that analysis cannot, by its statistical logic, convincingly determine

SYMBOLIC RACISM: A CONTEMPORARY LOOK

141

causal ordering in cross-sectional data. Our point here is, rather, that ‘‘dominance’’ is the most unique feature of social dominance theory, and it seems peripheral to racial politics in our analyses.

D. SYMBOLIC RACISM AS AN EXPRESSION OF WHITES’ GROUP INTERESTS The social structural theories place much emphasis on privileged groups’ sense of group interest as a driving motive for their political opposition to greater racial equality. We argue above that to test that view would require direct measures of the psychological constructs involved, perhaps along with longitudinal data or at least cross-sectional data on preadults. But that does not exhaust the number of open questions about the diVerences between symbolic racism and social structural theories. One puzzle is how to disentangle simple aVect toward symbols of the ingroup from the sort of group identification that group interest theories describe (Conover, 1988). Second, the strongest evidence for a group-interested dynamic is perhaps the several findings that perceived interracial conflict is associated with greater opposition to racial policies (Bobo & Johnson, 2001; Sears & Jessor, 1996). Unfortunately, that finding does not distinguish clearly between the group-interest account and a symbolic politics story that features the role of outgroup antagonism instead. Third, we earlier proposed contrasting self-oriented and more symbolic versions of group interest (Sears & Kinder, 1985). We find little evidence for the self-oriented version, in that perceived interdependence of the self with other Whites has little eVect on the policy preferences we have examined (see Tables XII and XIV above; also Sears & Jessor, 1996). However, we are reluctant to conclude that a more symbolic version of group interest is at work without more direct and positive evidence (see also Hutchings & Valentino, 2004).

E. POPULAR AND UNPOPULAR RACIALLY RELEVANT POLICIES We have found that the eVects of symbolic racism are almost identical across a wide variety of racial policies (Table VI above; Sears et al., 1997). This evidence seems to us quite convincing that the most salient feature of such policies is that they are specifically racial, but there still are very large diVerences in aggregate White support across various racial policies; e.g., between voluntary and mandatory, or between ‘‘equal opportunity’’ and ‘‘equal outcome’’ programs. What else is going on that explains

142

DAVID O. SEARS AND P. J. HENRY

those aggregate diVerences? Survey-based experiments have documented numerous such diVerences, but typically they have provided little direct evidence about why they occur (see especially Hurwitz & PeZey, 1998; Sniderman & Carmines, 1997). The explanations oVered by these authors are usually little more than intuitive interpretations of complex item-wording diVerences across diVerent policies. For example, Sniderman and Carmines (1997) interpret White opposition to race-targeted aYrmative action as violating Whites’ senses of ‘‘fairness’’ or ‘‘race-blind equality’’ but oVer no direct evidence that such values are implicated. Such direct evidence is needed if we are to understand why some policies are more palatable to Whites than others, beyond the observation that ‘‘race matters.’’ Already some recent research has begun to consider some of this direct evidence. For example, one series of studies (Bobocel, Son Hing, Davey, Stanley, & Zanna, 1998) suggest that non-prejudice-based justice beliefs may help to explain diVerent levels of opposition to diVerent kinds of aYrmative action programs (see also Stoker, 1998; Virtanen & Huddy, 1998).

F. DO CONTROLS ON IDEOLOGY ‘‘OVERCONTROL,’’ UNDERESTIMATING THE EFFECTS OF RACISM? In assessing the racial contribution to symbolic racism and racial policy preferences, we have controlled on party identification and ideology. To provide a conservative test of the symbolic racism theory, we have interpreted such controls as entirely nonracial in nature. This protects us from falsely interpreting any possibly nonracial components of symbolic racism as reflecting racial animosity, but it also might risk overcontrolling. These ostensibly nonracial predispositions in fact are likely to have some racial component, in many observers’ judgments. The Edsalls (1991) argued that ‘‘When the oYcial subject is presidential politics, taxes, welfare, crime, rights, or values . . . the real subject is race.’’ Sidanius and his colleagues (1999) argue that conservative ideology itself is in part a product of eVorts to legitimize the American racial hierarchy, and in that sense itself is racialized despite its ostensibly nonracial content. How far does racial prejudice really range into issues that have no direct or exclusive racial focus? There is substantial evidence that some putatively nonracial issues have become racialized over the past few decades. Racial attitudes had strong eVects on preferences for sharp tax and social-spending cuts and for cutting the general size of government in California 20 years ago, even with controls on ideology and party identification (Sears & Citrin, 1982). Poverty and welfare appear to have become strongly racialized over that period (Gilens, 1999). So has violent crime (Mendelberg, 2001).

SYMBOLIC RACISM: A CONTEMPORARY LOOK

143

Have party identification, conservative political ideology, and presidential voting preferences all become racialized as well? Some say not very much (Abramowitz, 1994), some say quite a lot (Kinder & Sanders, 1996), and still others say it depends on how you interpret attitudes that might be used as controls (Miller & Shanks, 1996). We have generally fallen somewhere in between, suggesting that symbolic racism does aVect White support for Black candidates and White candidates hovering on the edges of overtly racist appeals, but not support for mainstream White presidential candidates (Sears et al., 1997). We also interpret ideology and party identification conservatively, treating them as if they were wholly nonracial. Our suspicion is that assigning all eVects of these partisan dimensions to the nonracial side has indeed resulted in overcontrolling, thereby underestimating some of the true pervasive political eVects of race. That is, we would argue that the racially based realignment of the 1960s that Carmines and Stimson (1989) identified still polarizes the political parties around the now-less-obviously racial substrate of symbolic racism. Indeed we have found evidence of a geographically based realignment driven by conservative racial attitudes in three ways (Sears & Valentino, 2002; Valentino & Sears, 2005). White Southerners are on average higher in symbolic racism than are white Northerners. Symbolic racism is more closely associated with both Republican party identification and Republican presidential voting among White Southerners than among their Northern counterparts. Third, the association between symbolic racism and party identification has increased over the last two decades in the White South, but not in the White North. The sectionally based realignment of Southern Whites to the Republicans seems to have had much to do with that region’s historic legacy of racial inequality.

G. SOCIAL DESIRABILITY AND IMPLICIT PREJUDICE An early assertion about symbolic racism was that it was a nonreactive, indirect measure of racism. In contrast, measures of old-fashioned racism were readily interpreted by respondents as measuring racism, leading to artifical underestimates of levels of prejudice (McConahay, Hardee, & Batts, 1981). Survey researchers have reported lower levels of anti-Black attitudes on traditional measures of racial attitudes when Black interviewers are employed (relative to the results from White interviewers; Kinder & Sanders, 1996; Schuman et al., 1997). The same has been reported in some experiments with students given symbolic racism measures (Fazio et al., 1995). Other unobtrusive measurement of racial attitudes seems to suggest that

144

DAVID O. SEARS AND P. J. HENRY

conventional surveys underestimate Whites’ racial hostility (Kuklinski et al., 1997). Such findings have led to suspicions that self-report measures of racial attitudes underestimate Whites’ actual level of racism. Among other things it has led to the development of ‘‘implicit’’ measures of prejudice, which presumably tap unconscious, automatic, and uncontrollable forms of prejudice (Greenwald & Banaji, 1995). A ‘‘dual process’’ model of prejudice has developed that separates uncontrollable, unconscious ‘‘implicit’’ prejudice from controllable, conscious, ‘‘explicit’’ prejudice (Devine, 1989). These implicit measures also presumably have the advantage of not being influenced by social desirability pressures, thereby perhaps constituting more valid indicators of prejudice. One way of describing the threat this poses to symbolic racism research is that the associations of symbolic racism with opposition to racial policies might simply reflect shared error variance (i.e., variation in the tendency to control the overt expression of whatever racial animosity the individual possesses). We have several reasons for believing this is not an adequate explanation of symbolic racism eVects. For one thing, other measures of racial animosity, such as old-fashioned racism, stereotypes, and anti-Black aVect, have much smaller correlations with racial policy preferences than does symbolic racism, even though by this argument they too should share such error variance with the dependent variables. More generally, White respondents are usually interviewed by White interviewers, not Black interviewers, and it is unknown how much such artificial suppression of prejudice occurs with White interviewers. Implicit prejudice itself actually turns out to be influenced by a variety of contextual pressures (Blair, 2002). The association between implicit and explicit measures of prejudice remains cloudy, and there are disquieting methodological issues about implicit prejudice, such as divergences among alternative measures of it as well as lower levels of stability over time (e.g., Cunningham et al., 2001). Nevertheless, this strikes us as an extraordinarily fruitful line of research, and we stand to gain much more insight about racial attitudes from it.

X. Conclusions There is mostly good news for the theory of symbolic racism, if not for American society, in this story. Looking back now, the original theory seems itself rather intuitive and unformed, and it has been sharpened considerably by the many scholarly exchanges on this topic. The original research testing it can fairly be said to have been vulnerable to a number of

SYMBOLIC RACISM: A CONTEMPORARY LOOK

145

alternative explanations, but in the light of considerable later research, its core propositions have largely proved to be on target. Readers should justifiably be somewhat skeptical of this account, suspecting it to be subject to self-serving biases. Perhaps it is to some degree; we would not be the most trustworthy judges of that. However, we believe we have fairly and comprehensively summarized the available research and are not aware of compelling data contrary to the views summarized above. The new racism is not the same as the old. It combines racial and nonracial elements. A kind of racialized individualism lies most prominently behind symbolic racism. It is a very powerful determinant of Whites’ opposition to policies designed to produce more racial equality. The rhetoric of that opposition rarely contains explicitly racist or even racial argument. Rather, its manifest content focuses on basic political ideology and practical matters. Much of that nonracial focus may be sincere, at least in conscious intent, but behind much of it, we believe, lies racial prejudice. It is an element not heard very much these days in open political debate (and thank goodness for that), but it is very powerful nonetheless. Where does this leave us, living in a society that seems unable to put race aside or deal with racism openly? Racial inequality continues to pervade most areas of social and economic life, and it does not seem to be diminishing rapidly (Smelser et al., 2000). How is that to be rectified? One option would seem to be to turn to non-race-targeted policies that benefit Blacks, which seem to be more popular among Whites (Bobo & Kluegel, 1993; Sniderman & Carmines, 1997). The diYculty with that is that putatively nonracial issues quickly become racialized, with welfare (Gilens, 1999), tax cuts (Sears & Citrin, 1982), crime (Mendelberg, 2001), and a variety of other issues (Kinder & Sanders, 1996) as case studies. Issues perceived as at least correlated with race do not long stay above the fray of racial prejudice. Others advocate cooling down racial liberals’ moralizing rhetoric. Rather, liberals should treat racial issues as just as another case of conflicting group interests, much like wage disputes between labor and management. They then might get adjudicated through simple negotiations that might reach stable and mutually agreed-upon compromises (Bobo, 2000). The problem here too, in our eyes, is that such issues tend not to remain simple matters of bargaining over emotionally neutral resources. Rather, they get bogged down in racialized anxieties and accusations. A longer and more diYcult way may be necessary. White America has developed a case of collective amnesia that allows it to treat racism as an ugly chapter in ancient American history, a chapter closed for good with the passage of civil rights legislation in the 1960s. But in our view, neither people nor societies change that abruptly. Our data testify to the continuing political power of racism many years later. It may be that anti-Black racism in

146

DAVID O. SEARS AND P. J. HENRY

America, like German anti-Semitism, cannot be put aside without demonstrating and acknowledging the continuing role of racial prejudice and discrimination in our society. Indeed, we may all need first to find our own racism in our hearts and minds before we can take the steps necessary to rectify its eVects.

References Aboud, F. E. (1988). Children and prejudice. New York: Basil Blackwell. Abramowitz, A. I. (1994). Issue evolution reconsidered: Racial attitudes and partisanship in the U.S. electorate. American Journal of Political Science, 38, 1–24. Alwin, D. F., Cohen, R. L., & Newcomb, T. M. (1991). Political attitudes over the life span: The Bennington women after fifty years. Madison: University of Wisconsin Press. Biernat, M., & Crandall, C. S. (1999). Racial attitudes. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of political attitudes (pp. 297–412). San Diego, CA: Academic Press. Blair, I. (2002). The malleability of automatic stereotypes and prejudice. Personality and Social Psychology Review, 6, 242–261. Bobo, L. (1983). Whites’ opposition to busing: Symbolic racism or realistic group conflict? Journal of Personality and Social Psychology, 45, 1196–1210. Bobo, L. (1988). Group conflict, prejudice, and the paradox of contemporary racial attitudes. In P. Katz & D. Taylor (Eds.), Eliminating racism: Profiles in controversy (pp. 85–116). New York: Plenum. Bobo, L. (1999). Prejudice as group position: Micro-foundations of a sociological approach to racism and race relations. Journal of Social Issues, 55, 445–472. Bobo, L. (2000). Race and beliefs about affirmative action: Assessing the effects of interests, group threat, ideology and racism. In D. O. Sears, J. Sidanius, & L. Bobo (Eds.), Racialized politics: The debate about racism in America (pp. 137–164). Chicago: University of Chicago Press. Bobo, L., & Kluegel, J. R. (1993). Opposition to race-targeting: Self-interest, stratification ideology, or racial attitudes? American Sociological Review, 58, 443–464. Bobo, L., & Smith, R. A. (1994). From Jim Crow racism to laissez-faire racism: An essay on the transformation of racial attitudes in America. In W. Katkin & A. Tyree (Eds.), Beyond pluralism: Essays on the conceptions of groups and identities in America (pp. 182–220). Urbana-Champaign: University of Illinois Press. Bobo, L., & Hutchings, V. L. (1996). Perceptions of racial group competition: Extending Blumer’s theory of group position to a multiracial social context. The American Sociological Review, 61, 951–973. Bobo, L., & Johnson, D. (2001). Racial attitudes in a prismatic metropolis: Mapping identity, stereotypes, competition, and views on affirmative action. In L. D. Bobo, M. L. Oliver, J. H. Johnson, Jr., & A. Valenzuela, Jr. (Eds.), Prismatic metropolis: Inequality in Los Angeles (pp. 81–163). New York: Russell Sage Foundation. Bobocel, D. R., Son Hing, L. S., Davey, L. M., Stanley, D. J., & Zanna, M. P. (1998). Justicebased opposition to social policies: Is it genuine? Journal of Personality and Social Psychology, 75, 653–669.

SYMBOLIC RACISM: A CONTEMPORARY LOOK

147

Carmines, E. G., & Merriman, W. R., Jr. (1993). The changing American dilemma: Liberal values and racial policies. In P. M. Sniderman, P. E. Tetlock, & E. G. Carmines (Eds.), Prejudice, politics, and the American dilemma (pp. 237–255). Stanford, CA: Stanford University Press. Carmines, E. G., & Stimson, J. A. (1989). Issue evolution: Race and the transformation of American politics. Princeton, NJ: Princeton University Press. Citrin, J., & Green, D. P. (1990). The self-interest motive in American public opinion. In S. Long (Ed.), Research in micropolitics (Vol. 3, pp. 1–28). Greenwich, CT: JAI Press. Conover, P. J. (1988). The role of social groups in political thinking. British Journal of Political Science, 18, 51–76. Converse, P. E. (1964). The nature of belief systems in mass publics. In D. E. Apter (Ed.), Ideology and discontent (pp. 206–261). New York: Free Press of Glencoe. Cunningham, W. A., Preacher, K. J., & Banaji, M. R. (2001). Implicit attitude measures: Consistency, stability, and convergent validity. Psychological Science, 12, 163–170. Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5–18. Dovidio, J. F., Kawakami, K., Johnson, C., Johnson, B., & Howard, A. (1997). On the nature of prejudice: Automatic and controlled processes. Journal of Experimental Social Psychology, 33, 510–540. Edsall, T. B., & Edsall, M. D. (1991). When the official subject is presidential politics, taxes, welfare, crime, rights, or values . . . the real subject is race. The Atlantic Monthly, 267, 53–86. Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation as an unobtrusive measure of racial attitudes: A bona fide pipeline? Journal of Personality and Social Psychology, 69, 1013–1027. Federico, C. M., & Sidanius, J. (2002a). Racism, ideology, and affirmative action revisited: The antecedents and consequences of ‘‘principled objections’’ to affirmative action. Journal of Personality and Social Psychology, 82, 488–502. Federico, C. M., & Sidanius, J. (2002b). Sophistication and the antecedents of whites’ racial policy attitudes: Racism, ideology, and affirmative action in America. Public Opinion Quarterly, 66, 145–176. Feldman, S. (1988). Structure and consistency in public opinion: The role of core beliefs and values. American Journal of Political Science, 32, 416–440. Feldman, S., & Huddy, L. (2005). Racial resentment and white opposition to race-conscious programs: Principles or prejudice? American Journal of Political Science, 49, 168–183. Gaertner, S. L., & Dovidio, J. F. (1986). The aversive form of racism. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 61–89). Orlando, FL: Academic Press. Gilens, M. (1998). Racial attitudes and race-neutral social policies: White opposition to welfare and the politics of racial inequality. In J. Hurwitz & M. Peffley (Eds.), Perception and prejudice: Race and politics in the United States (pp. 171–201). New Haven, CT: Yale University Press. Gilens, M. (1999). Why Americans hate welfare: Race, media, and the politics of antipoverty policy. Chicago: University of Chicago Press. Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 4–27. Henry, P. J., & Sears, D. O. (2002). The symbolic racism 2000 scale. Political Psychology, 23, 253–283. Howell, S. E. (1994). Racism, cynicism, economics, and David Duke. American Politics Quarterly, 22, 190–207.

148

DAVID O. SEARS AND P. J. HENRY

Hughes, M. (1997). Symbolic racism, old-fashioned racism, and whites’ opposition to affirmative action. In S. A. Tuch & J. K. Martin (Eds.), Racial attitudes in the 1990s: Continuity and change. (pp. 45–75). Westport, CT: Praeger. Hurwitz, J., & Peffley, M. (1998). Introduction. In J. Hurwitz & M. Peffley (Eds.), Perception and prejudice: Race and politics in the United States. (pp. 1–16). New Haven, CT: Yale University Press. Hutchings, V. L., & Valentino, N. A. (2004). The centrality of race in American politics. Annual Review of Political Science, 7, 383–408. Kuklinski, J. H., Cobb, M. D., & Gilens, M. (1997). Racial attitudes and the ‘‘New South.’’ The Journal of Politics, 59, 323–349. Jennings, M. K., & Stoker, L. (1999). The persistence of the past: The class of 1965 turns fifty. Prepared for presentation at the Midwest Political Science Association Convention, Chicago, IL, April. Jessor, T. (1988). Personal interest, group conflict, and symbolic group affect: Explanations for whites’ opposition to racial equality. Unpublished doctoral dissertation, Department of Psychology, University of California, Los Angeles. Jost, J. T., & Thompson, E. P. (2000). Group-based dominance and opposition to equality as independent predictors of self-esteem, ethnocentrism, and social policy attitudes among African Americans and European American. Journal of Experimental Social Psychology, 36, 209–232. Katz, I. (1981). Stigma: A social psychological analysis. Hillsdale, NJ: Erlbaum. Kinder, D. R. (1986). The continuing American dilemma: White resistance to racial change 40 years after Myrdal. Journal of Social Issues, 42, 151–171. Kinder, D. R., & Mendelberg, T. (2000). Individualism reconsidered: Principles and prejudice in contemporary American opinion. In D. O. Sears, J. Sidanius, & L. Bobo (Eds.), Racialized politics: The debate about racism in America (pp. 44–74). Chicago: University of Chicago Press. Kinder, D. R., & Sanders, L. M. (1996). Divided by color: Racial politics and democratic ideals. Chicago: University of Chicago Press. Kinder, D. R., & Sears, D. O. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life. Journal of Personality and Social Psychology, 40, 414–431. Kluegel, J. R. (1990). Trends in whites’ explanations of the black-white gap in socioeconomic status, 1977–1989. American Sociological Review, 55, 512–525. Kluegel, J. R., & Smith, E. R. (1986). Beliefs about inequality: Americans’ views of what is and what ought to be. New York: Aldine de Gruyter. Lambert, W. E., & Klineberg, O. (1967). Children’s views of foreign peoples. New York: Appleton-Century-Crofts. Lipset, S. M. (1996). American exceptionalism: A double-edged sword. New York: Norton. McConahay, J. B. (1986). Modern racism, ambivalence, and the modern racism scale. In J. Dovidio & S. Gaertner (Eds.), Prejudice, discrimination, and racism. (pp. 91–126). Orlando, FL: Academic Press. McConahay, J. B., Hardee, B. B., & Batts, V. (1981). Has racism declined in America? It depends upon who is asking and what is asked. Journal of Conflict Resolution, 25, 563–579. McConahay, J. B., & Hough, J., Jr. (1976). Symbolic racism. Journal of Social Issues, 32, 23–45. Mendelberg, T. (2001). The race card: Campaign strategy, implicit messages, and the norm of equality. Princeton, NJ: Princeton University Press. Miller, S., & Sears, D. O. (1986). Stability and change in social tolerance: A test of the persistence hypothesis. American Journal of Political Science, 30, 214–236. Miller, W. E., & Shanks, J. M. (1996). The new American voter. Cambridge, MA: Harvard University Press.

SYMBOLIC RACISM: A CONTEMPORARY LOOK

149

Myrdal, G. (1944). An American dilemma. New York: Harper and Row. Peffley, M., & Hurwitz, J. (1998). Whites’ stereotypes of blacks: Sources and political consequences. In J. Hurwitz & M. Peffley (Eds.), Perception and prejudice: Race and politics in the United States (pp. 58–99). New Haven, CT: Yale University Press. Pettigrew, T. F., & Meertens, R. W. (1995). Subtle and blatant prejudice in Western Europe. European Journal of Social Psychology, 25, 57–75. Pole, J. R. (1993). The pursuit of equality in American history. Berkeley and Los Angeles: University of California Press. Schuman, H., Steeh, C., Bobo, L., & Krysan, M. (1997). Racial attitudes in America: Trends and interpretations. Cambridge, MA: Harvard University Press. Sears, D. O. (1969). Political behavior. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (rev. ed., Vol. 5, pp. 315–458) . Reading, MA: Addison-Wesley. Sears, D. O. (1986). College sophomores in the laboratory: Influences of a narrow database on social psychology’s view of human nature. Journal of Personality and Social Psychology, 51, 515–530. Sears, D. O. (1988). Symbolic racism. In P. Katz & D. Taylor (Eds.), Eliminating racism: Profiles in controversy (pp. 53–84). New York: Plenum Press. Sears, D. O. (1993). Symbolic politics: A socio-psychological theory. In S. Iyengar & W. J. McGuire (Eds.), Explorations in political psychology (pp. 113–149). Durham, NC: Duke University Press. Sears, D. O. (1994). Ideological bias in political psychology: The view from scientific hell. Political Psychology, 15, 547–556. Sears, D. O., & Citrin, J. (1982). Tax revolt: Something for nothing in California. Cambridge, MA: Harvard University Press. Sears, D. O., & Funk, C. L. (1991). The role of self-interest in social and political attitudes. In M. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 24, pp. 1–91). Orlando, FL: Academic Press. Sears, D. O., & Jessor, T. (1996). Whites’ racial policy attitudes: The role of white racism. Social Science Quarterly, 77, 751–759. Sears, D. O., & Henry, P. J. (1999). Ethnic identity and group threat in American politics. Political Psychologist, 4, 12–19. Sears, D. O., & Henry, P. J. (2003). The origins of symbolic racism. Journal of Personality and Social Psychology, 85, 259–275. Sears, D. O., Henry, P. J., & Kosterman, R. (2000). Egalitarian values and contemporary racial politics. In D. O. Sears, J. Sidanius, & L. Bobo (Eds.), Racialized politics: The debate about racism in America. (pp. 75–117). Chicago: University of Chicago Press. Sears, D. O., & Kinder, D. R. (1970). The good life, ‘‘white racism,’’ and the Los Angeles voter. Paper delivered at 50th annual meeting of Western Psychological Association. Los Angeles, April 15, 1970. Sears, D. O., & Kinder, D. R. (1971). Racial tensions and voting in Los Angeles. In W. Hirsch (Ed.), Los Angeles: Viability and prospects for metropolitan leadership. (pp. 51–88). New York: Praeger. Sears, D. O., & Kinder, D. R. (1985). Whites’ opposition to busing: On conceptualizing and operationalizing group conflict. Journal of Personality and Social Psychology, 48, 1141–1147. Sears, D. O., & Levy, S. (2003). Childhood and adult political development. In D. O. Sears, L. Huddy, & R. Jervis (Eds.), Handbook of political psychology (pp. 114–159). New York: Oxford University Press. Sears, D. O., & McConahay, J. B. (1973). The politics of violence: The new urban blacks and the Watts riot. Boston: Houghton-Mifflin.

150

DAVID O. SEARS AND P. J. HENRY

Sears, D. O., & Valentino, N. A. (2002). Is the proof in the pulpit? Race, religion, and party realignment in the South. Presented at the annual meeting of the American Political Science Association, Boston, MA, August 31. Sears, D. O., van Laar, C., Carrillo, M., & Kosterman, R. (1997). Is it really racism? The origins of white Americans’ opposition to race-targeted policies. Public Opinion Quarterly, 61, 16–53. Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press. Sidanius, J., Devereux, E., & Pratto, F. (1992). A comparison of symbolic racism theory and social dominance theory as explanations for racial policy attitudes. Journal of Social Psychology, 132, 377–395. Sidanius, J., Levin, S., Rabinowitz, J. L., & Federico, C. M. (1999). Peering into the jaws of the beast: The integrative dynamics of social identity, symbolic racism, and social dominance. In D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcoming group conflict (pp. 80–132). New York: Russell Sage Foundation. Smelser, N. J., Wilson, W. J., & Mitchell, F. (Eds.) (2001). America becoming: Racial trends and their consequences (Vol. II). Washington, DC: National Academy Press. Sniderman, P. M., & Carmines, E. G. (1997). Reaching beyond race. Cambridge, MA: Harvard University Press. Sniderman, P. M., & Piazza, T. (1993). The scar of race. Cambridge, MA: Harvard University Press. Sniderman, P. M., & Tetlock, P. (1986a). Symbolic racism: Problems of motive attribution in political debate. Journal of Social Issues, 42, 129–150. Sniderman, P. M., & Tetlock, P. (1986b). Reflections on American racism. Journal of Social Issues, 42, 173–187. Sniderman, P. M., Crosby, G. C., & Howell, W. G. (2000). The politics of race. In D. O. Sears, J. Sidanius, & L. Bobo (Eds.), Racialized politics: The debate about racism in America (pp. 236–279). Chicago: University of Chicago Press. Steeh, C., & Krysan, M. (1996). Trends: Affirmative action and the public, 1970–1995. Public Opinion Quarterly, 60, 128–158. Stoker, L. (1998). Understanding whites’ resistance to affirmative action: The role of principled commitments and racial prejudice. In J. Hurwitz & M. Peffley (Eds.), Perception and prejudice: Race and politics in the United States (pp. 135–170). New Haven, CT: Yale University Press. Tarman, C., & Sears, D. O. (2005). The conceptualization and measurement of symbolic racism. Journal of Politics. In press. Tetlock, P. E. (1994). Political psychology or politicized psychology: Is the road to scientific hell paved with good moral intentions? Political Psychology, 15, 509–529. Thernstrom, S., & Thernstrom, A. (1997). America in black and white: One nation, indivisible. New York: Simon & Schuster. Valentino, N. A., & Sears, D. O. (2005). Old times there are not forgotten: Race and partisan realignment in the contemporary South. American Journal of Political Science. Forthcoming. Virtanen, S. V., & Huddy, L. (1998). Old-fashioned racism and new forms of racial prejudice. The Journal of Politics, 60, 311–332. Wood, J. (1994). Is ‘‘symbolic racism’’ racism? A review informed by intergroup behavior. Political Psychology, 15, 673–86. Zaller, J. (1992). The nature and origins of mass opinion. New York: Cambridge University Press.

MANAGING GROUP BEHAVIOR: THE INTERPLAY BETWEEN PROCEDURAL JUSTICE, SENSE OF SELF, AND COOPERATION

David De Cremer Tom R. Tyler

This chapter focuses on how and why procedural fairness is an important group tool for managing cooperation. We introduce a motivational model of procedural fairness that links the fairness of procedures explicitly to the construction of the social self, a process that, in turn, aVects psychological processes and elicits cooperation. First, a review is provided of studies supporting the argument that procedural fairness is related to the social self. Second, an overview is given addressing how the social self relates to cooperation via the processes of trust and goal transformation. An integrative conceptual framework is then introduced and used for understanding the interplay of procedures, self, and cooperation. Finally, some implications of this motivational model are discussed, and conclusions are drawn.

I. Introduction and Overview Group membership and interaction within groups are central to modern life. We all belong to groups, and within this social context we are influenced by what goes on within the relationships that we have with others in those groups. This group context also has an important influence on our perceptions, motives, cognitions, and feelings. Central to the group context is the interdependence of group members (Cartwright & Zander, 1968; Lewin, 1948). As a consequence, our actions in groups aVect not only our own interests but also other people’s interests. 151 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 37

Copyright 2005, Elsevier Inc. All rights reserved. 0065-2601/05 $35.00

152

DAVID DE CREMER AND TOM R. TYLER

Although people in groups are interdependent, this does not mean that their goals are always the same. People’s relationship to the group is better described as a mixed-motive situation, in which some of their personal goals are compatible with the goals of other individuals in the group, and some are not. Because of this mixed-motive nature of people’s actions in groups, cooperation is often diYcult to achieve. The important question then becomes ‘‘How can we promote cooperation within groups?’’ A focus on cooperation is one that is important to social psychologists. To understand why people are motivated to cooperate, we need to understand how people define themselves when they are in group contexts. This definition involves self-interest, group welfare, and judgments about the person’s relationship to the group. Cooperation requires an individual to be willing to pursue the group’s welfare. In other words, it is linked to a motivation that is broader in scope than the pursuit of short-term individual self-interest. The central thesis of this chapter is that procedural justice is a key aspect of groups that defines the degree to which the people within those groups will be motivated to engage in cooperation with others in the group. As justice researchers have recently noted, ‘‘Individuals are likely to be motivated on behalf of the organization when the organization is motivated on behalf of them’’ (Cropanzano & Schminke, 2001), a motivation evaluated by looking at the procedural justice of groups. As a result, procedural justice helps groups to manage the problem of obtaining cooperation—if people feel that group procedures are fair, they are more highly motivated to cooperate. In this chapter, we identify and discuss one source of information through which the group shows that it does or does not care about its group members; that is, the fairness of its procedures. In the past the literatures on procedural fairness and cooperation have largely existed independent of each other (for exceptions, see De Cremer & Van Vugt, 2002; Tyler & Dawes, 1993; Tyler & Degoey, 1995). In this chapter, we argue that a successful linkage between these two literatures provides important insights into the dynamics of groups. To show that procedural fairness is a key antecedent to cooperation in groups, we adopt a motivational strategy in which we define motivation as a process that arouses, directs, and maintains human behavior (Blau, 1993; Pittman, 1998; Young, 1961; Zander, 1971). Adopting such a motivational approach we review the existing procedural fairness models—the group value model (Lind & Tyler, 1988), the relational model of authority (Tyler & Lind, 1992), and the group engagement model (Tyler & Blader, 2000)—and argue that each of these models stands for the existence of a similar pervasive inner drive, need, or concern that ultimately influences people’s reactions to fair procedures. In reviewing these models and their motivational components, we present the results of our own experimental studies to provide evidence that all of the

PROCEDURAL JUSTICE, SELF, AND COOPERATION

153

psychological concerns underlying these models (e.g., self-esteem, need to belong, self-uncertainty, and concern for reputation) are important for explaining fair procedure eVects (De Cremer, 2002, 2003a,b; De Cremer & Blader, 2004; Brockner, De Cremer, Van den Bos, & Chen, in press; De Cremer & Sedikides, in press; De Cremer & Tyler, 2004; Tyler, 1999; Tyler, Degoey, & Smith, 1996). In other words, a first purpose of this chapter is to integrate all these various psychological models into a unifying framework based on one common human motivation. We argue that this mechanism points to the argument that procedures motivate people by communicating information relevant to their needs and drives. This information is expected to influence people’s social self, which is considered to include and specify people’s motives and needs (Sedikides & Brewer, 2001; Tyler & Smith, 1999). The role of procedural justice in shaping motivation is only the first part of the definition of motivation. This definition also includes two other parts; namely, directing behavior and maintaining that behavior over time and across situations. These two additional aspects of the motivational equation also have to be supported to understand why procedural fairness is able to promote cooperation. That is, if people’s needs and concerns are influenced, this will both activate behavior and maintain that behavior over time. How will procedural fairness guide the direction that people’s behavior takes and whether that behavioral direction is maintained over time? Answering this question will be the second main purpose of this chapter. In determining the answer to this question, we use insights from the cooperation literature and, more specifically, research on social dilemmas. This literature has shown that promoting cooperation is facilitated when trust is reinforced and the interests of the group are experienced as one’s own interests (De Cremer & Stouten, 2003; De Cremer & Van Dijk, 2002; Pruitt & Kimmel, 1977). We demonstrate that the ability of procedures to aVect people’s needs and concerns leads to the creation of positive beliefs within the group (e.g., the judgment that others in the group are trustworthy) and to a transformation of motives from the personal level (acting out of self-interest) to the group level (viewing self-interest and group interest as the same or acting out of group interest). This linkage will be demonstrated using past and recent research (Brockner, Siegel, Daly, Tyler, & Martin, 1997; De Cremer, 2004a; De Cremer & Van Knippenberg, 2002; De Cremer & Van Vugt, 1999). It is assumed that reinforcing the group’s interest will direct people’s actions (i.e., cooperating for the group’s welfare) and the existence of trust will ensure that people will maintain such cooperative behaviour over time and across situations. After discussing procedural fairness as a motivational component comprising the elements of arousal, direction, and maintenance,

154

DAVID DE CREMER AND TOM R. TYLER

our line of reasoning will be summarized in a model of cooperation in which procedural fairness will figure as the regulating factor. Thereafter, we will conclude by outlining the strengths of this motivational approach by discussing its integrative potential, that is, its ability to integrate the existing procedural fairness models and to link it to other areas such as the social dilemmas literature. Finally, a brief overview will be presented how our findings and model can relate to other social issues addressing important instances of group behavior.

II. The Relation Between Procedural Fairness and Cooperation Groups are an essential part of our social lives and influence much of our psychological life. Indeed, previous research has shown that memberships in groups can help us to define ourselves; that is, they may tell us who we are (Sherman, Hamilton, & Lewis, 1999). Also, groups may create a setting in which important psychological goods such as approval, self-worth, belongingness, and self-definition are provided to its members (Baron & Kerr, 2003; Tajfel & Turner, 1986). Groups help people to meet identity-based and social needs. Of importance to this chapter is that groups also have functional value because they are better able to protect us and to promote the cooperation necessary for obtaining the survival benefits that can be provided by membership in groups (Baumeister & Leary, 1995; Brewer, 1991). Indeed, for functional reasons linked to survival it is necessary that groups be cooperative and successful in solving internal problems of organization and coordination (Caporeal & Brewer, 1991; Leakey, 1978). So, groups help people to meet material and instrumental needs, as well as helping them meet identity-based and social needs. One way to motivate cooperation is to appeal to people’s material and instrumental needs by promising people rewards when they cooperate or threatening them with sanctions when they fail to cooperate. Although these approaches do motivate cooperative behavior, their influence is weak and diYcult to maintain over time (Tyler & Blader, 2000). Further, their use fails to activate, and can even undermine, intrinsic motivations, leading to longrun problems in groups (De Cremer & Van Knippenberg, 2002; Tenbrunsel & Messick, 1999; Tyler & Blader, 2000). It is much more desirable for groups to encourage intrinsic motivation among group members, and it is our belief that the use of procedural fairness may also be an important antecedent that activates people’s intrinsic motivation to promote the interests and viability of the group. To achieve all of

PROCEDURAL JUSTICE, SELF, AND COOPERATION

155

this, we believe that procedures are required that promote a positive climate within the group by responding to the needs and concerns of group members. If group members’ needs and concerns are addressed, group members will be motivated to cooperate with the group. Indeed, in most groups ‘‘tragedy strikes when self-interest and social interest diverge’’ (Baden, 1998; p.51), and therefore it has been advocated that social arrangements need to emerge to restrain competitive actions and promote the cooperative urges of group members (Posner, 2000; Schroeder, Steel, Woodell, & Bembenek, 2003; Tyler, Boeckmann, Smith, & Huo, 1997). The establishment of such arrangements will foster a group climate in which ‘‘procedural fairness’’ is a given (Schroeder et al., 2003). The important role of procedural fairness in promoting cooperation has, however, received very little attention. The literature on procedural fairness often does not include behavioral reactions as part of the fairness construct, and the social dilemma literature has not generally focused on the role of procedures in monitoring group behavior to promote cooperation (Cropanzano, Byrne, Bobocel, & Rupp, 2001; De Cremer & Van Vugt, 2002; Lind, 2001a; Tyler & Blader, 2000; Tyler & Dawes, 1993; Tyler & Degoey, 1995). However, the two literatures share a focus on how people’s social self is influenced by the relationships they have with the group and its members (also including authorities and institutions). The social self is defined in this chapter as based on people’s relationships with other group members; within these relationships, people evaluate interpersonal relatedness, intimacy, and interdependence within the relationship (see Baumeister & Leary, 1995; Sedikides & Brewer, 2001). Thus, we argue that people’s self is fundamentally relational and, as such, can be conceived of as a social self (see also Andersen, Chen, & Miranda, 2002; Leary, 2002; Tyler & Lind, 1992; Tyler & Smith, 1999). Indeed, as Sedikides and Gregg (2003, p. 110) note, ‘‘the self operates predominantly within the social world. This means that the psychological phenomena that fall under its umbrella typically arise in interaction with others, real or imagined.’’ In social psychology, it is well accepted that people’s social self is developed and constructed by information that one receives through social interactions with others (De Cremer, 2003a; Leary, 2001; Leary & Baumeister, 2000; Sedikides & Gregg, 2003). In turn, this relational information defines one’s level of identity and goals and ultimately regulates one’s social actions (Carver & Scheier, 1998). On the basis of this argument, we reason that the fairness of enacted procedures communicates information relevant to the social self and, in turn, motivates group members to engage in cooperative behavior aimed at promoting the group’s interest (e.g., De Cremer, Tyler & den Ouden, in press; De Cremer & Van Knippenberg, 2002; Tyler & Blader, 2000, 2003b). Thus, procedural fairness (as a group means) serves as a

156

DAVID DE CREMER AND TOM R. TYLER

motivational antecedent of cooperation, because it aVects people’s social self and their self-regulation. As a consequence, via these self-related processes group members will adopt the goal of promoting the interests of themselves and the other group members (De Cremer & Van Dijk, 2002; De Cremer & Van Vugt, 1999). This line of reasoning is based on the assumption that the social self is dynamic (i.e., the content of the self-concept will vary across situations and experiences) and that by means of procedures, diVerent aspects of people’s social self can be altered (e.g., having an individual versus a collective goal/ perspective; see also Hogg, 2003; Turner, 1987). Thus, the social self is knowledge that people have about their socially constructed self-concept and that involves goals, values, and beliefs. By activating or reinforcing certain aspects of the social self, particular goals, beliefs, and values will be activated (cf. Lord, Brown, & Freiberg, 1999). Once these goals are accessible and salient, they will energize and direct behavior (Carver, 2001; Pervin, 1982). In other words, this accessible information will influence people’s actions and thus guide the process of self-regulation (cf. Kunda, 1999). Our argument is that the social self has major implications for motivation (see also Sedikides & Brewer, 2001). The way that those implications play out depends on which aspects of the social self are accessible and salient at any given time and in particular situations. From the group perspective, therefore, the key to motivating cooperation is to activate aspects of the social self that are linked to collective goals and perspectives. Once those aspects are activated, people will be motivated to engage in cooperative behavior based on those aspects of the self. This motivational approach has been suggested by procedural fairness researchers (Tyler & Blader, 2000, 2003a,b; Tyler & Lind, 1992), but many procedural fairness models have shown a tendency to focus more on the role that cognitive factors play in responding to variations in procedural fairness (see e.g. Folger, 1986; Folger & Cropanzano, 1998; Lind, 2001a; Van den Bos, Lind, & Wilke, 2001; Van den Bos & Lind, 2002). For example, Lind (2001a), in his fairness heuristic theory, advocates the use of more cognitive analyses of fairness judgment processes when he describes the use of fair procedures in terms of heuristics, and Folger and Cropanzano (1998), in their fairness theory, explain justice eVects by referring to the eVect of attribution-based cognitive analyses. Furthermore, even Thibaut and Walker (1975), in their pioneering research, were more inclined to a cognitive analysis, as their main concern was to examine how people deal with uncertainty surrounding the decisions they experience in the courtroom (i.e., being seen as guilty or innocent). One important way to deal with this type of uncertainty is to focus on how fairly

PROCEDURAL JUSTICE, SELF, AND COOPERATION

157

the trials were conducted; that is, to use procedural information to address cognitive uncertainty about the fairness of an outcome. Recently, Van den Bos and Lind (2002) elaborated on this original conception of uncertainty by introducing the uncertainty management model. This model suggests that the higher the degree of uncertainty that people experience, the more people will make use of procedural fairness information. Procedural fairness is then used as a cognitive heuristic to make more sense of the uncertainty experienced (i.e., fair procedures indicate that the situation is a safe one and will thus reduce uncertainty). It is important to note, however, that the uncertainty management model does not specify which particular types of uncertainty were being talked about (Tyler & Blader, 2003a,b).1 As their model seems to build, at least partly, on Thibaut and Walker’s prior work, it is our impression that in this model, uncertainty reflects how uncertain people feel about their outcomes (see, e.g., Van den Bos, Lind, Vermunt, & Wilke, 1997).2 As such, the uncertainty management model deals to a certain extent with how people use procedural fairness in a cognitive way (i.e., as a heuristic) to manage uncertainty about outcomes. For example, if an outcome is unexpected (regardless of the valence of the outcomes), then people will experience uncertainty and will try to make sense of the outcome by using procedural fairness information as substitutability information (see Lind, 2001a; see also Brockner & Wiesenfeld, 1996). In this chapter, however, our motivational account focuses on how procedural fairness plays a role in dealing with uncertainty about the social self and its related social motives. Thus, we argue that these cognitive analyses have paid too little attention to the role of subjective elements of procedures such as motives, needs, and aVect. However, we do wish to note that some models do include some motivations but do not consider defining motivation to be of primary importance, or have not elaborated on or created a motivational account (Van den Bos, 2003). In addition, these procedural fairness models were not developed to explain behavioral eVects like cooperation (see Tyler & Blader, 2000, 2003b). Hence, we distinguish our eVort to articulate a motivational account from eVorts to deal with outcome uncertainty.

1 Van den Bos (2001) also acknowledged that their model discusses uncertainty in a rather general way. For example, Van den Bos (2001) argues that ‘‘all uncertainties are not the same and cannot be expected to have the same eVects,’’ and he went on to invite justice researchers to focus on ‘‘relationships between uncertainty, self-esteem, confidence and control’’ (p. 940). 2 This model may also partly include a self-relevant analysis (see Van den Bos, 2001) modeled after Tyler (1990), as noted by Van den Bos (2004).

158

DAVID DE CREMER AND TOM R. TYLER

Why is it so important to emphasize such a motivational account? Literature on social cognition, for example, has shown that cognition and motivation are closely connected (Higgins & Kruglanksi, 2001), so it follows that it is important that a social psychological analysis of procedural fairness also focuses on the role of motivation. Furthermore, a motivational account of procedural justice events becomes even more important to look at when we wish to investigate procedural fairness eVects on behavioral instances such as cooperation. Indeed, as argued by Carver (2001, p. 307), ‘‘social behavior is built . . . on motivational processes.’’ Although behavioral reactions to procedural fairness are not considered to be part of the fairness construct, it is essential that a psychology of procedural fairness include a motivational component to account for the eVect of procedural fairness on behavior. In the justice literature (e.g., Cropanzano et al., 2001), it is readily acknowledged that to understand the fundamentals of procedural fairness we particularly need to know which motives and concerns drive people to pay attention to the fairness of procedures enacted by authorities, thus advocating a motivational approach. In the following section, we provide an overview of how previous procedural fairness models have accounted for the possible role of motives in the psychology of procedural fairness.

III. Procedural Fairness Models: Motives and Social Self The issue of fairness is a dominating theme in our daily lives, and concerns about the meaning of justice go back to ancient moral philosophers such as Plato and Socrates (Rawls, 1971; Ryan, 1993; for a review see Tyler et al., 1997). Indeed, the concept of justice is intuitively related to humanitarian and ethical standards that describe how we should act and treat others (e.g., Folger & Cropanzano, 1998; Miller, 2001). In eVect, fairness can be considered one of the most important guidelines in our lives, leading both scholars and lay people to frequently pose the question of why we care so much about fairness. One specific instance of fairness that people attend to very carefully is whether groups and organizations use correct and fair procedures in making decisions and allocating outcomes; that is, procedural fairness (Leventhal, 1980; Thibaut & Walker, 1975). Not surprisingly, the topic of procedural fairness is and should be an important topic of inquiry in the field of social psychology (Lind & Tyler, 1988). Indeed, historically, we can point out the influence of various social phenomena and issues that highlight the significance of procedural fairness. The most well-known example is the problem with eVective third-party conflict resolution eVorts that led John Thibaut, a psychologist, and Laurens

PROCEDURAL JUSTICE, SELF, AND COOPERATION

159

Walker, a lawyer, to collaborate, study, and make more prominent issues of procedural fairness (Thibaut & Walker, 1975; Tyler, 1990). In essence, the study of the role of procedural fairness in social phenomena has become important because social psychologists have recognized a need to understand why people were and are behaving in particular ways. In particular, a wide variety of studies indicates that procedural justice is central to the motivation of both immediate and long-term cooperation with the decisions made within groups, organizations, and societies (Tyler, 2000). This procedural justice eVect emerges in studies conducted in legal settings, in work organizations, in educational settings, and in families (Lind & Tyler, 1988). The one important question that we wish to address here is why people are motivated to cooperate as a function of procedural fairness. As noted above, there are many illustrations in the fields of organizational behavior, economics, and psychology that procedural fairness leads to many desirable organizational and group consequences (see Tyler, 2000, for a review), but these examples do not tell us what drives and motivates these findings. Why is people’s willingness to cooperate within groups shaped by the fairness they associate with their fellow group members and group authorities? In particular, why do people react to their evaluations of procedural justice? After all, people could potentially react to a wide variety of aspects of their social and group environments when they are trying to manage cognitive and social uncertainty (cf. Van den Bos et al., 1997). To understand such findings and to respond in a satisfactory manner to the above questions, a motivational analysis of procedural fairness eVects is necessary. In doing this analysis, we advocate the need to examine the central role of the social self in the psychology of procedural fairness. Several fairness models have been proposed to understand the importance of motives in relationships and the importance of social self in the psychology of procedural fairness. We will be focusing here on a set of models that share an emphasis on the relational and self-relevant implications of fairness evaluations. These models represent a programmatic research program designed to understand the noninstrumental value of the psychology of procedural fairness (i.e., the relevance to the social self ). In doing so, these models point out the importance of motivational processes in understanding why procedural fairness matters within groups. Below, we will first discuss how each of these models provides contributions to the development of a motivational account of procedural fairness that ultimately may predict cooperation in groups.

A. THE GROUP-VALUE MODEL The group-value model (Lind & Tyler, 1988) focuses on the antecedents of judgments of procedural fairness and predicts that noninstrumental factors

160

DAVID DE CREMER AND TOM R. TYLER

will influence procedural fairness judgments, and that these factors will matter more within the context of one’s own group. That is, not only are procedures used by the group valued to obtain control of the resources one may receive from this group but, even more importantly, the fairness of procedures is important because fair procedures also signal a symbolic message that one is valued by the group and, therefore, is viewed by others as worthy of receiving respectful treatment. An experiment by Lind, Kanfer, and Earley (1990) provided evidence that such relational or noninstrumental concerns do indeed play a role in the psychology of procedural fairness. These authors used a goal-setting context, in which participants were allowed to express their opinion—or not— about the type of goal that had to be reached. In this study, providing voice or not was the manipulation of procedural fairness (Folger, 1977). Participants received an opportunity to voice their opinion before the experimenter had decided on the goal (predecision voice), after the experimenter had decided on the goal (postdecision voice), or not at all (no voice). The results showed that perceptions of fairness were enhanced by the possibility of voice, even when there was no chance of influencing the decision (i.e., postdecision voice). In addition, predecision voice was found to lead to even greater perceived fairness than postdecision voice. Similarly, field studies of voice indicate that people are influenced by whether they receive opportunities for voice, even when they feel that what they say has little or no influence on the outcomes they obtain (Tyler, 1987). Taken together, studies supporting the group-value model make it clear that procedures have the potential to communicate to group members that they are valued members and, as such, can use the group as a reference point to define themselves (see also Tajfel & Turner, 1986). Thus, the motive of defining oneself in terms of a valued group membership is clearly acknowledged in this model.

B. THE RELATIONAL MODEL OF AUTHORITY The relational model of authority (Tyler & Lind, 1992) focuses on the possible social psychological factors shaping reactions to treatment by authorities. This model predicts that procedural fairness will influence reactions to authorities, and that the eVect of these procedures is influenced by relational concerns—in particular, how neutral, trustworthy, and respectful an authority is when dealing with group members. For example, research by Tyler (1989) showed that the fairness of police oYcers and judges determined people’s judgments and attitudes about legal authorities. More specifically, being treated fairly by the police or court indicates to people that

PROCEDURAL JUSTICE, SELF, AND COOPERATION

161

one is respected and, as such, receives the treatment good citizens ‘‘deserve’’ (see also Tyler, 1994). In addition, such positive relational information indicates that one’s social standing within society or the group is regarded as positive. In turn, this type of relational information has been shown in several studies to significantly influence people’s level of self-esteem, thus illustrating how strongly these relational concerns are related to people’s social self and self-worth (Tyler et al., 1996). Taken together, studies supporting this model provide evidence for a motivational account by showing that procedural fairness has strong selfrelevant implications that shape people’s sense of social worth and social status or social reputation (De Cremer & Tyler, 2004; Tyler & Smith, 1999). Also, in line with a motivational approach, Cropanzano et al. (2001, p. 177) recently concluded that ‘‘the need for belonging could serve as one mechanism for the relational model’’ of authority. Further, in a recent commentary, Lind (2001b) argued that one reason people value fair procedures is that these procedures aYrm their sense of belongingness to the group or organization at hand. In other words, the key assumption of these relational models of justice ‘‘focuses on messages of inclusion’’ (Lind, 2001b). Tyler and Blader (2002) provided empirical support for this argument by showing that judgments about justice linked to inclusion influenced reactions to group membership. Thus, in support of a motivational account, the relational model of authority points out that the fairness of procedures communicates information relevant to one’s need for belonging and concerns for social reputation.

C. THE GROUP-ENGAGEMENT MODEL The group-engagement model (Tyler & Blader, 2000, 2003b) integrates the insights of the group-value model (Lind & Tyler, 1988) and the relational model of authority (Tyler & Lind, 1992) but is broader in its scope. In essence, the group-engagement model broadens the focus of procedural fairness models by positing a general model of the relationships between people and groups. In doing so, this model tries to understand why people engage in their groups psychologically and behaviorally and to what extent procedures may play a role in this process. According to this model, the reason procedures have the potential to motivate people toward group engagement is that they influence how people define themselves (e.g., interdependent versus independent, collective versus personal identity; Tyler, 1999). In turn, these identity judgments (which take the form of a certain type of social self; see Tyler & Smith, 1999) are considered to be the primary factors shaping attitudes, values, and behaviors

162

DAVID DE CREMER AND TOM R. TYLER

in groups. Taken together, the contribution of the group engagement model to a motivational account of procedural fairness is that it assumes that procedures influence people’s social self by shaping their self-definition.

D. INTEGRATING MOTIVES AS PART OF THE SOCIAL SELF Reviewing these procedural fairness models, it can be concluded that procedures enacted within the group have implications for the social self. More precisely, these models assume that within group settings, people extract relational information from procedures that are being used to make outcome distribution decisions. As such, people infer their social reputation, their degree of belongingness, their level of identity, and their degree of selfuncertainty within the group from the perceived fairness of procedures. Thus, procedures viewed as accurate or as allowing voice opportunities are considered fair and, consequently, self-validating at the social level. In contrast, inaccurate or no-voice granting procedures are considered unfair and, consequently, self-devaluing. This line of reasoning fits well with findings derived from the selfevaluation literature (Sedikides, 2002; Sedikides & Gregg, 2003; Sedikides & Strube, 1997). That is, people, at least in part, base their feelings of social self-worth on relational information, such as direct interpersonal appraisals (Tice & Wallace, 2003), clues about social acceptance or rejection (Leary & Baumeister, 2000), and the subjective sentiment of the need for relatedness not having been met (Deci & Ryan, 2000). Taken together, we conclude that the motives of social reputation, belongingness, level of identity, and selfuncertainty form important needs that guide the extent to which people attend to procedures; the reason being that procedural fairness is assumed to communicate this type of information, which people see as relevant to the social self. Further, with respect to the issue of cooperation, and following a motivational account, all these motives are believed to influence the goal and direction one will take during the process of self-regulation. That is, if procedures shape whether or not people experience belongingness, influence people’s satisfaction with one’s intragroup status, shape whether people define themselves in terms of their relationships with the group, and finally, shape whether or not they experience self-uncertainty, people’s goals will be shifted toward a group orientation, and cooperation is more likely to emerge (e.g., De Cremer & Van Dijk, 2002). To demonstrate that procedures indeed communicate information relevant to these social self-features, we describe below several of our own studies illustrating such a linkage between procedures and self-related consequences.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

163

IV. Overview of Procedural Fairness Research In this section, we describe research findings showing that procedures enacted by both authorities and fellow group members have the potential to significantly influence people’s social self by aVecting their belongingness needs, their concerns about social reputation, their type of self-definition, and their degree of self-uncertainty. Procedures are expected to communicate information relevant to these features of the social self. In demonstrating these relationships, our studies made use of the functional approach (Snyder, 1993), in which it is assumed that social situations (e.g., the relationship with the authority and group) can be seen as an input into one’s own personal values, goals, and agendas. This person-situation approach (e.g., Mischel, 1973; Snyder & Cantor, 1998) indicates that the psychological eVect of a situational feature such as (un)fair procedures depends on how people interpret the situation, and such interpretations may vary as a function of individual diVerence variables. For example, if I am someone who cares very much about inclusive relationships, then procedural fairness may influence my reactions if procedures indeed communicate information about belongingness and inclusiveness. Thus, by examining the moderating eVect of individual diVerences in motives on procedural fairness, this approach may enhance our insight into the type of social self information procedural fairness actually communicates. This interactional strategy between person and situation was first articulated in detail by Snyder and Ickes (1985). This strategy seeks to understand regularities and consistencies in social behavior in terms of the interactive influence of dispositional and situational features. In doing this, this research line uses measures of personality processes and constructs as moderator variables and, as such, can be referred to as a dynamic interactionist approach: a reciprocal interplay between person and situation. According to Snyder and Cantor (1998), this interactionistic approach (combining personality and social situations) is an eVective tool ‘‘because they focus on shared concerns with motivation’’ (p. 639). As mentioned earlier, it is exactly this focus on motives that is of particular interest when determining the concerns related to the social self that underlie people’s interest in and reactions toward procedural fairness. Thus, people will interpret any given social situation as a function of their own motivational orientations and, in turn, will be guided by the opportunities and specifications communicated by those situations. In other words, social situations can guide individuals in their actions and reactions as long as the information derived from this social situation fits well with the needs and motives of the specific individuals involved. That is, social situations can

164

DAVID DE CREMER AND TOM R. TYLER

be perceived as inputs to the personal agenda of the individual actor involved in the situation (Snyder, 1993).

A. BELONGINGNESS AND PROCEDURAL FAIRNESS According to the relational model of authority, procedures communicate messages of inclusion and belongingness (Cropanzano et al., 2001; Tyler & Smith, 1999; Tyler & Blader, 2002; Tyler & Lind, 1992). Thus, procedures carry with them information about how much the group considers you as belonging to the group. De Cremer and Blader (2004), therefore, argue that the extent to which procedural fairness influences people’s reactions should be a function of how strong people’s need to belong is. The theoretical analysis that these authors put forward is that people attend to procedural fairness because it is relevant to the evaluation of their belongingness in the group. They do so in their eVort to develop a sense of self vis-a`-vis the group. De Cremer and Blader (2004) argue that an implication of this theoretical assumption is that diVerences in the need to belong should aVect the extent to which people react to procedural fairness, which would provide support for the implicit assertion of the relational models that belongingness needs are the fundamental motive driving people’s reactions to justice. Lack of such evidence would challenge the validity of those models and the relevance of the fundamental psychological motives on which they are based. To test these predictions, three studies were conducted to examine the relationship between belongingness needs and procedural fairness. In the experimental study, these authors first assessed participants’ need to belong by means of a recently developed individual diVerence scale (Leary, Kelly, Cottrell, & Schreindorfer, 2001; see appendix I for example items). Thereafter, participants were given a booklet that informed them about the task that they would have to perform during the experimental session. It was explained that their performance on this task would be evaluated by means of a settled procedure (i.e., performance would be accurately assessed and the leader would take suYcient time to evaluate). Thereafter, the procedural fairness manipulation was introduced. For this purpose, De Cremer and Blader operationalized procedural fairness by applying the rule of consistency (Leventhal, 1980). That is, participants were told that the leader evaluating the performance was someone who always used the described procedure across persons and time (i.e., high consistency condition) or that the leader was someone who did not always follow up the procedures (i.e., low consistency condition). Results revealed that participants’ level of self-esteem was higher when the leader was consistent rather than inconsistent, but only when participants were high in need to belong. Thus, these findings provide

PROCEDURAL JUSTICE, SELF, AND COOPERATION

165

evidence that procedural fairness information carries with it important motivational information related to belongingness needs. In a related manner, De Cremer and Tyler (2004) showed that a relational component of procedures, that is, respect, is also related to this fundamental human motive of belongingness. In a laboratory study, they invited participants to the laboratory, where they had to perform a group task. Before starting with the experiment, participants’ need to belong was assessed (see also De Cremer & Blader, 2004). Thereafter, participants were told that they would work in a group of four and that they first had to complete a survival task. The responses of each group member on this task would be taken together to form the group response. Participants then completed this task, and thereafter responses were sent to the other group members. After this, the respect manipulation was introduced. Participants received a message from the other group members communicating either that they had much respect for this person and his or her responses or that they disrespected this person. Finally, participants were asked to what extent they wished to leave the group for the remainder of the study. As expected, results revealed an interaction showing that participants wanted to leave the group when they were disrespected rather than respected, but this pattern was only found among participants high in need to belong. Again, these findings indicate that procedural fairness and the associated relational components like respect have a strong influence on people’s belongingness needs. In addition, De Cremer and Blader (2004) further reasoned that to the extent that belongingness is indeed related to process fairness, those with a high need to belong should be more vigilant processors of fairness information. That is, their concern with accurately judging process fairness (and, more generally, with determining the extent to which they are included in the group) should lead them to process procedural fairness information more systematically, as compared with those with a low need to belong who are less concerned with the procedural evaluation and what it communicates. Therefore, in line with dual-process theories (e.g., Chaiken, Liberman, & Eagly, 1989; Eagly & Chaiken, 1993), they predicted that the processing of procedural information should depend on an individual’s processing motivation, which is expected to be shaped by an individual’s belongingness needs. These authors suggested that showing such a cognitive eVect would further indicate that procedures are recognized as sources of belongingness information and, because of this function, will attract the attention of those who are in need of this specific relational information. This line of reasoning was tested in a scenario study in which participants had to imagine that they were a member of a department within a larger organization. Before reading this scenario, however, participants first filled

166

DAVID DE CREMER AND TOM R. TYLER

out the need-to-belong scale. Thereafter, they started reading the scenario. Participants were told that because of bad economics, departments (including the participants’) had to be restructured. Therefore, the organization had hired an external manager to guide this process. In trying to guide this restructuring process, this manager decided to give the members of the department involved voice. Thus, each participant was told that the manager would listen to their ideas and suggestions. To examine ‘‘processing motivation,’’ a common method in social psychology was used by varying argumentation strength (i.e., the strength of the reasons provided why voice was given). Half of the participants were provided with compelling reasons for why a manager would be providing their employees with voice, whereas the other half of participants were provided with poor reasons.3 As predicted, participants’ reactions (in this case fairness and trustworthiness judgments) were more strongly influenced when a strong reason rather than a weak reason was given (i.e., fairness and trust judgments were higher and more positive when voice relative to no voice was given), and this was particularly the case for those participants who were high in need to belong. Taken together, these studies thus reveal consistent evidence that a motivational account to explain procedural fairness eVects has its validity, because procedures do seem to communicate information that satisfies people’s need to belong. Further, because procedures communicate information relevant to belongingness needs, it was also shown that procedures will attract more attention and matter more in influencing people’s reactions when one is in need of belongingness information. As such, the use of fair procedures satisfies belongingness needs; a process that in turn shapes people’s social self. The findings were consistent across research setting (i.e., field versus laboratory) and research paradigm and hold for relational information communicated by both authorities and equal-status group members.

B. SOCIAL REPUTATION AND PROCEDURAL FAIRNESS Another part of the social self that procedural fairness is assumed to influence is people’s sense of social reputation. As early as the writings of James (1890), researchers have acknowledged that one’s social self is at least 3

The strong reasons that were used included ‘‘the manager believes in democratic values indicating to people that they are important and valuable’’ and ‘‘the manager considers it important that all employees are heard and feel that they are part of the company,’’ whereas the poor reasons that were used included ‘‘the manager once heard about giving voice opportunities and thought it was a fun idea’’ and ‘‘giving voice just feels good at this moment.’’ The relative strength of the reasons has been used successfully in prior research (see De Cremer & Blader, 2004; De Cremer & Sedikides, 2004).

PROCEDURAL JUSTICE, SELF, AND COOPERATION

167

partly determined by one’s social reputation. Indeed, social evaluation is an important element in the process of constructing the self (cf. Tice, 1992). People are, by their very nature, very motivated to obtain a positive image or reputation (see, e.g., the extensive literature on people’s public self-presentation skills; Baumeister, 1982; Leary, 2001). In fact, social reputations largely determine how one’s behaviors are recognized and rewarded (e.g., Johnson, Erez, Kiker, & Motowidlo, 2002). As a result, one’s social self clearly includes a concern about social evaluation and, consequently, a concern about one’s reputation within the group. The powerful eVects of social reputation in groups have, for example, been demonstrated by recent research showing that people exhibit more cooperative behavior when their reputation is threatened, help others more easily when they have a positive image, and are more likely to develop positive and enjoyable relationships with others if they possess a positive reputation (e.g., De Cremer, Snyder, & Dewitte, 2001; Ga¨chter & Fehr, 1999; Milinski, Semmann, & Krambeck, 2002). All of these findings are consistent with the argument that people care about how they are evaluated by others. In a series of studies, De Cremer and Tyler (2004) examined the extent to which people high in concern for reputation were influenced by manipulations of respect (i.e., a relational feature of procedural fairness) and voice. In the first study, participants were required to fill out a recently developed questionnaire assessing their concern about their social reputation (for more details, see De Cremer & Tyler, 2004; see appendix for sample items). Participants scoring high on this scale are generally concerned and worried about how they are perceived by others, whereas those scoring low on this scale are generally less concerned. After filling out this scale, the participants were asked to recall a situation in their department in which they felt respected or disrespected. They were asked to write a short story about this specific situation (most stories described situations of respect or disrespect in the classroom). Then, participants’ level of self-esteem was assessed. Results revealed an interaction showing that participants’ level of self-esteem was higher when they recalled a situation of respect rather than disrespect, but this pattern was only found among participants high in concern for social reputation. To obtain more evidence for this relation between respect and concern for social reputation, De Cremer and Tyler (2004) also conducted a study in which concern for reputation was assessed first, before participants had to read a scenario about a situation in which they felt respected versus disrespected. This time, a broader range of dependent measures was used by asking participants for both their positive and negative emotional reactions. As in the previous study, results revealed significant interactions on both negative and positive emotions. Participants were more happy when they

168

DAVID DE CREMER AND TOM R. TYLER

received respect rather than disrespect and were more angry when they received disrespect rather than respect. As expected, these patterns only emerged among those high in concern for social reputation. Finally, De Cremer and Sedikides (2004) also conducted a series of studies in which procedural fairness was directly manipulated by providing participants the opportunity to voice their opinion (or not). As in the respect studies, concern for social reputation was assessed by using the same individual diVerence scale. Results of these studies revealed that procedural fairness and concern for social reputation interacted in a similar way respect and concern for social reputation did. Indeed, the procedure of voice influenced participants’ self-esteem reactions only when participants were high in concern for social reputation. Furthermore, the studies of De Cremer and Sedikides also showed that violations in procedural fairness (i.e., lack of voice in one study and communicating weak arguments for giving voice in another study) diminished the reported self-esteem of participants with a positive reputation but not of those with a negative reputation (i.e., valence of reputation was induced by asking participants to recall social situations in which their reputation was clearly positive versus negative). These findings thus demonstrate that fair procedures signal a positive contribution to one’s reputation, and therefore those with a positive reputation are most sensitive toward signs from their supervisor that carry self-evaluative information (cf. Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001). All these studies thus convincingly show that procedural fairness strongly influences people’s social self by aVecting people’s self-evaluative concerns. One such specific concern is the extent to which people care about their reputation among others in the group. This focus on the importance of people’s social reputation in accounting for procedural fairness eVects is closely related to Tyler’s (2001) notion of the reputational social self. This distinct aspect of the social self refers to people’s concern about their position and reputation within the group (Tyler & Smith, 1999). Tyler indeed argued that self-related information such as procedures has a major eVect on people’s reactions when their reputational social self is salient.

C. SELF-DEFINITION AND PROCEDURAL FAIRNESS Relational models of procedural fairness have in common that the relationship between procedures and features of the social self must directly influence people’s sense of identity or, in other words, one’s self-description in terms of their identity. As a result, using our moderator approach, the predictions of these models should be particularly applicable to those people

PROCEDURAL JUSTICE, SELF, AND COOPERATION

169

who assign importance to their relationships with others for self-definitional reasons. Such a focus on how the quality of relationships one has with other parties, such as authorities and groups, influences people’s type and level of identity fits well with the strong focus of contemporary social psychology on the importance of the social self and identity motives (see, e.g., Sedikides & Brewer, 2001). Which type of identity is likely to play a role in the process of using relational information such as fair procedures to construct one’s social self ? One psychological variable likely to reflect the degree of psychological significance that people assign to their relationship with another party is interdependent self-construal (ISC; Markus & Kitayama, 1991; Singelis, 1994). ISC refers to the extent to which people define themselves on the basis of their relationships with others. Markus and Kitayama originally conceptualized ISC in the context of trying to account for cross-cultural diVerences in people’s cognitions, emotions, and behaviors. Initially, researchers advocated the perspective that individuals in Western and Eastern countries diVer in how they perceive and define themselves in terms of their relationships with others. However, more recently it has been argued that ISC does not predominate in one specific culture but, rather, exists within the individual across cultures (Singelis, 1994). Thus, scales assessing individual diVerences in ISC were developed. Relatedly, Vandello and Cohen (1999) recently found evidence of meaningful within-culture variance in individualism–collectivism constructs that are closely related to ISC. In short, although ISC had its roots in cross-cultural theory and research, it is now coming to be seen as a meaningful variable even in studies conducted within a single culture. Thus, people with a more interdependent self-construal assign greater psychological significance to their relationships with other parties, relative to those with lower levels of ISC. It therefore stands to reason that the relational explanation of procedural fairness eVects would be particularly applicable to those with high levels of ISC. Put diVerently, the identity explanation posits that procedural fairness factors should be more likely to influence the attitudes and behaviors of those high, rather than low, in ISC (see also De Cremer & Van den Bos, 2003, for evidence that the cooperation level of those high in ISC as a function of voice is explained largely in terms of relational concerns). These predictions were examined by Brockner et al. (in press) in a series of three studies. In a first experiment, these authors asked participants to fill out the interdependent self-construal scale, as developed by Singelis (1994). After this, participants engaged in a task that they would perform with another participant present in the laboratory. The task consisted of identifying a number of squares that were highlighted in a figure at the upper left

170

DAVID DE CREMER AND TOM R. TYLER

of the computer screen (see Van den Bos & Lind, 2002, for more details of this specific experimental task). Participants had to try to finish as many tasks as possible within a period of 10 minutes. After they finished this task, they were told that they would receive lottery tickets that they could use to earn money. The better they performed, the more lottery tickets they could receive. Then the procedural fairness manipulation was introduced by telling participants that they would receive voice to articulate their opinion about how many tickets they should receive or that they would not receive such voice opportunity. Finally, participants’ positive aVect was measured. Results revealed that procedural fairness influenced participants’ positive aVect positively, but this was only the case among participants high in ISC and not those low in ISC. To increase the external validity of these findings, these authors also conducted a field negotiation study in which participants’ level of ISC was assessed before engaging in the negotiation task. During the negotiation task, participants’ perceptions about how fairly and respectfully they were treated was measured. Results again revealed an interaction between ISC and procedural treatment in a way that participants were more likely to continue the interaction between themselves and the negotiation partner when treatment was perceived as fair rather than as unfair. Again, this pattern was only found among those high in ISC. These studies show that the fairness of procedures is used for purposes of self-definition. Procedural fairness is assumed to indicate to group members that they are valued by the group and its authority, which enables them to use their group membership to define themselves (e.g., Lind & Tyler, 1988; Tyler & Blader, 2000, 2003b). Individuals high in ISC consider their social memberships (i.e., those within groups and relationships) as important because they perceive self and others to be intertwined, and therefore information about their relationship with the authority and the group (i.e., procedural fairness) influences their reactions.

D. SELF-UNCERTAINTY AND PROCEDURAL FAIRNESS According to the uncertainty management model (Van den Bos & Lind, 2002), feelings of uncertainty make procedural fairness more important, but in our opinion, as noted earlier, this model emphasizes particularly how people cognitively deal with uncertainty about outcomes. In this analysis, we suggest that people’s sense of uncertainty about their social self also makes people be more responsive and attentive toward procedural fairness issues. In fact, in the group-engagement model, Tyler and Blader (2003b) argue that the influence of outcomes occurs through the influence of outcomes on

PROCEDURAL JUSTICE, SELF, AND COOPERATION

171

identity judgments. In contrast, the influence of identity judgments on attitudes, values, and behaviors is direct. However, we argue, as does the uncertainty management model, that uncertainty about one’s social self (i.e., about reputation and status in a group) leads people to seek information. The information that people view as most relevant to that social self is information about procedural justice. More specifically, we suggest that people use variations in procedural justice as an informational source, as a critical link between the social environment and the self, and as a way to reduce uncertainty about the self (McGregor, Zanna, Holmes, & Spencer, 2001; Sedikides & Gregg, 2003; Sedikides & Strube, 1997). When self-uncertainty (e.g., doubts about one’s identity and inclusiveness) is high, people will be particularly sensitive and responsive to procedural justice variations (cf. Brockner, 1984; Campbell, 1990). On the one hand, such variations will be seen either as validating a positive but fragile self-aspect or as threatening a negative and fragile selfaspect. In the first case, aVective, cognitive, and behavioral reactions will be overly positive, and in the second case they will be overly negative. On the other hand, variations in procedural justice will be perceived as having selfclarification potential. Such information will enable the individual to draw inferences concerning their social inclusiveness, social acceptance, or social standing inferences (Tyler, 1989, 1999). To examine the potential influence of procedural fairness on people’s sense of self-uncertainty, De Cremer and Sedikides (2003) conducted several studies in which self-uncertainty was measured and operationalized by means of self-esteem stability (Kernis, Whisenhunt, Waschull, Greenier, Berry, Herlocker, & Anderson, 1998), self-concept clarity (Campbell, Trapnell, Heine, Katz, Lavallee, & Lehman, 1996; see appendix for sample items), and self-doubt (Oleson, Poehlmann, Yost, Lynch, & Arkin, 2000; see appendix for sample items). In a first study, participants’ stability of selfesteem (Dykman, 1998) was assessed to determine whether participants possessed stable or unstable self-esteem. Thereafter, they were required to read a scenario in which they had to imagine being an employee of a company. In this scenario they were eligible for a job promotion. Then the procedural fairness manipulation was introduced by informing participants that those eligible for a promotion had to participate in nine diVerent tests. In the fair procedure condition, participants read that the company rated all nine tests to select the best candidate, whereas in the unfair procedure condition, participants read that only one of the nine tests was rated. Finally, participants’ fairness judgments and positive aVect was assessed. In line with predictions of the uncertainty management model, results revealed that for both fairness judgments and aVect, participants were influenced by the procedural fairness manipulation in such a way that the

172

DAVID DE CREMER AND TOM R. TYLER

accurate treatment was judged to be more fair and elicited more positive aVect than the inaccurate treatment. However, this pattern only emerged for those with unstable self-esteem. To further corroborate these findings, another study was conducted in which participants first had to fill out the self-doubt scale (Oleson et al., 2000). Self-doubt captures more directly the sense of disbelief and distrust in one’s abilities or characteristics than self-esteem stability does. High scores on this scale indicate that people experience doubts (i.e., are less certain) about their self-concept. After this, participants read a story saying that they were employees at a high-flying company and that, because of the company’s strong profit, a bonus of 10,000 Dutch Guilders (approximately $4,500) would be passed on to them. The allocation decision would be made by the management. Thereafter, the procedural fairness manipulation was introduced by informing half of the participants that management decided to give them voice to make this decision, whereas the other half of the participants was informed that no such voice was granted to them. Finally, participants’ negative aVect was measured. Again, in line with predictions, results indicated that participants receiving no voice reported stronger negative aVect than those receiving voice, but this was only found among those high in self-doubt. Further evidence for the proposition that procedural fairness influences people’s social self via uncertainty related to the self was obtained in a study in which the authors asked participants to fill out the self-concept clarity scale. Self-concept clarity can be defined as ‘‘the extent to which the contents of an individual’s self-concept (e.g., perceived personal attributes) are clearly and confidently defined, internally consistent, and temporally stable’’ (Campbell et al., 1996, p. 141). After responding to this scale, participants were placed in separate experimental cubicles and told that they had been asked to help the university board of a Dutch university in designing electronic planners that students could use to organize their studies and preparations of assignments. Therefore, the board needed comments from students. To make this issue clear to the participants, the students were given information about these electronics planners, and they also read a newspaper article that addressed this issue. Thereafter, participants were required to write down their comments on a piece of paper that was lying in front of them. After finishing this assignment, the comments were picked up by the experimenter, and participants had to wait while a representative of the university board read the comments and decided who would be asked to voice their comments to the university board. After a few minutes, half of the participants were told that they were given voice, whereas the other half was told that they did not receive voice. Finally, participants’ negative aVect was measured.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

173

Again, results revealed a significant interaction indicating that more negative aVect was reported in the no-voice condition relative to the voice condition, but only among those low in self-concept clarity.

E. SUMMARY These studies consistently show that procedures influence people’s reactions because they directly aVect the social self. More precisely, all these studies indicate that the motivational implication of using procedural fairness when making decisions in a group is that they satisfy people’s social selfrelated needs by providing people with a positive reputation, a sense of belonging, a sense of social interdependence, and a decrease in feelings of self-uncertainty. As such, it can be argued with confidence that procedural fairness has pervasive implications for people’s evaluation of these aspects of the social self. Final evidence that satisfying these needs and motives by means of fair procedures reflects a process in which people’s social self-evaluation is positively reinforced is provided by a study of De Cremer and Van Hiel (2004). These authors assessed all four features of the social self (i.e., need to belong, concern for social reputation, interdependent self-construal, and self-uncertainty) together with the big five factors questionnaire. Results of this study showed that all four features of the social self loaded highly on one specific big five factor: neuroticism. Neuroticism is part of the big five model of personality that includes five broad factors (neuroticism, extraversion, conscientiousness, agreeableness, and openness to experience; see Digman, 1990). Previous research has shown that neuroticism is strongly (and negatively) correlated with measures of self-esteem (Watson, Suls, & Haig, 2002). Because self-esteem involves self-evaluation by means of assessing how well one is regarded and accepted by others (see sociometer theory; Leary & Baumeister, 2000), it can thus be concluded that our finding that concerns about one’s social self—need to belong, concern for social reputation, interdependent self-construal, and self-uncertainty—loaded highly on the factor neuroticism strongly indicates that these motives reflect people’s concerns about maintaining or establishing a positive self-evaluation within the social relationships they have with the group and its members (see, e.g., Judge, Locke, & Durham, 1997, for evidence that neuroticism is in fact a core selfevaluation). As we argued earlier, procedural fairness is one powerful tool within groups to manage such self-evaluation process. Although the above studies demonstrate that procedures have the potential to significantly aVect people’s social self by arousing the diVerent aspects

174

DAVID DE CREMER AND TOM R. TYLER

of this type of self, a motivational approach of procedural fairness still requires an explanation of how this social self then motivates people to act as good group members; that is, to engage in cooperation. In this chapter, it is assumed that the process of activating the social self will define the appropriate goals and that this, in turn, will guide people’s self-regulation process, in which action is undertaken (Carver & Scheier, 1998). In a group context, the goals that are most likely to be made salient will be group or collective-oriented goals. Indeed, procedures enacted by the group and its authority will contribute to the social self of its group members and will make them define themselves more as a part of the group or, using terms of self-categorization theory (Turner, 1987), will make them experience themselves as an embodiment of the group. This has as a consequence that the group’s goals are experienced as one’s own goals, and consequently, group goals will guide one’s actions and decisions. In terms of a motivational account, one could thus argue that the fairness of procedures will first motivate group members by significantly influencing their social self. A second step in such a motivational analysis requires us to explain how this motivation is directed toward action.

V. Social Dilemmas and Procedural Fairness: Trust and Goal-Transformation Now that we have established that procedural fairness significantly influences people’s motivation via the process of shaping the construction and maintenance of the social self, it becomes more understandable why it is that procedures should have the potential to motivate group members to cooperate for the group’s welfare. By satisfying people’s social motives and activating their social self, people will be motivated to accept the goals and beliefs of the group that treated them fairly. As such, group members will adopt the goal of promoting the group’s interest, and this will guide the process of self-regulation and, consequently, their decisions and their actions. This self-regulating process that is guided by the goals that are made salient by people’s social self (i.e., the group’s goal in this particular instance) links well with research that shows that how people define themselves (e.g., as a group member vs. an individual or as a leader vs. someone lower in power) influences their tendency to engage in action. For example, recent research by Galinski, Gruenfeld, and Magee (2003) showed that if people define themselves as someone high in power they are more motivated to take action as the behavior of taking action fits better with the goals and beliefs associated with power. Thus, powerful persons have the goal to take more

PROCEDURAL JUSTICE, SELF, AND COOPERATION

175

action, and this will shape their behavioral tendencies. In a similar vein, those whose social self is activated and positively reinforced by the use of fair procedures will be more likely to act in terms of their social self, which means that cooperation is more likely to emerge. A similar argument has been used in the social dilemma literature (Komorita & Parks, 1994). Social dilemmas are interdependence situations in which the personal interest is at odds with the collective interest, thus representing a mixed-motive situation (Kopelman, Weber, & Messick, 2002). In these situations, the dominant choice is to follow one’s own personal interest and not to contribute to the interest of one’s group. However, if all of those in a given situation adapt this economically rational strategy, the outcomes will be worse than if all members or both partners decide to cooperate. According to Dawes (1980), a social dilemma is thus characterized by two important features: first, defection will always be economically most beneficial, regardless of what others decide, but second, if all act this way, individual rationality (i.e., defecting) will lead to collective disaster. Within the social dilemma literature it is assumed that how the self is defined will determine the type of goal (personal or collective) that is pursued by group members. That is, people try to serve their own self-interest, but if the self is defined more socially, people’s level of interest may be situated at the level of the group. Thus, if people’s social self is positively promoted, then group members are more likely to cooperate in social dilemmas. Following this logic, Kramer and Brewer (1984) were the first to experimentally show that the extent to which individual decision-makers define themselves in terms of their group membership may function as a reference point for their decision behavior. In other words, when aYliation with the group or collective is reinforced, ‘‘the group is the basis of cooperation’’ (Turner, 1987; p. 34). Thus, how people’s self is defined (by means of procedures or other possible social cues) determines the extent to which people are motivated to pursue the group’s welfare. Why is it so important that people need to have a social self to promote cooperation? Recent research has argued that if group members have a positively reinforced social self, this psychological state will reduce the influence of two dominating motives in social dilemma situations; that is, fear and greed (De Cremer & Van Dijk, 2002; De Cremer & van Vugt, 1999). Thus, creating a strong sense of belongingness (i.e., reinforcing the social aspect of the self) diminishes people’s fear that others will exploit their cooperative eVorts and that they will end up as the ‘‘sucker’’ (Kerr, 1983). In addition, under these circumstances, people’s actions are also less likely to be guided by personal greed. Why? In the late 1970s, Pruitt and Kimmel (1977) developed their goal/ expectation theory to predict under which circumstances cooperation in

176

DAVID DE CREMER AND TOM R. TYLER

mixed-motive situations is most likely to emerge. According to these researchers, two conditions have to be fulfilled: group members need to have a cooperative goal and to trust the others in their cooperative intentions. In other words, ‘‘the goal of achieving mutual cooperation is insuYcient to elicit cooperative behavior. It must be accompanied by the expectation that the other will cooperate’’ (Pruitt & Kimmel, 1977, p. 375). Hence, trust or lack of trust in others is central to obtaining cooperation. A recent study by De Cremer and Stouten (2003) indeed showed that cooperation levels in a public good dilemma (a specific instance of a social dilemma) were highest when group members trusted the others and when they felt that the group and its interests were merged in their own selfdefinition (see self-other merging; Aron, Aron, Tudor, & Nelson, 1991). Thus, following the rationale of goal/expectation theory, it seems clear that enhancing trust in others and creating a collective or prosocial goal will reduce fear for exploitation and a tendency to pursue greed, respectively. That both having trust in others and having a prosocial goal are important determinants of cooperation has been demonstrated by a vast amount of empirical research. Trust has been argued to represent a psychological construct that may be beneficial in solving the conflict between one’s own interest and the interest of the others (Dawes, 1980; De Cremer & Van Vugt, 1999; Kramer, Brewer, & Hanna, 1996). Following Mayer, Davis, and Schoorman (1995), trust can be defined as ‘‘the willingness of a party to be vulnerable to the actions of another party based on the expectations that the other will perform a particular action important to the truster, irrespective of the ability to monitor or control that other party’’ (p. 712). Most studies of trust have used definitions similar to this one. In general, this perspective indicates that when trust is high, people will have confidence in another’s goodwill, will expect others to act in a moral way, and therefore engage in reciprocal cooperation (e.g., Granovetter, 1992; Ring & van de Ven, 1994). As such, people will experience less fear that others will exploit them, making it easier for them to justify their decisions to cooperate (Yamagishi & Sato, 1986). In line with this, research on social dilemmas has shown that a high level of trust influences expectations about other’s motives with respect to oneself (e.g., Brann & Foddy, 1987; De Cremer, 1999) and aVects behavior in interdependence situations (e.g., Messick, Wilke, Brewer, Kramer, Zemke, & Lui, 1983; Parks & Hulbert, 1995). Interestingly, these studies also showed that high trusters (who expect reciprocity) cooperated irrespective of whether others cooperated, indicating that trust is also linked to a sense of moral commitment (Kramer & Goldman, 1995). With respect to having a prosocial goal (see Pruitt & Kimmel, 1977), research on social value orientation, or preferences for particular patterns

PROCEDURAL JUSTICE, SELF, AND COOPERATION

177

of distributions of outcomes for self and others (Messick & McClintock, 1968; Van Lange, Otten, De Bruin, & Joireman, 1997), is particularly relevant. Research on this concept has revealed support for a three-category typology, including prosocial orientation (i.e., enhancing joint outcomes and enhancing equality in outcomes), individualistic orientation (i.e., enhancing outcomes for self with no or little regard to other’s outcomes), and competitive orientation (i.e., enhancing relative advantage over other’s outcomes). The latter two are often combined and form one specific group referred to as proselfs (Van Lange & Liebrand, 1991). Extant evidence indicates that individuals with prosocial orientations tend to act and think in a collectively rational manner (i.e., to think about what is good for the group), whereas individuals with proself orientation tend to act and think in an individually rational manner (i.e., thinking about what is good for themselves). Also, numerous studies indicate that, relative to proselfs, prosocials approach others more cooperatively, exercise more personal restraint in so-called resource dilemmas, and respond more cooperatively to a variety of strategies (e.g., Kramer, McClintock, & Messick, 1986; Liebrand & Van Run, 1985; McClintock & Liebrand, 1988). Such patterns have been consistently observed in several diVerent social dilemmas like give-some and take-some games (see Van Lange, 2000).4 All these findings provide support for the notion that diVerences between prosocials and proselfs represents diVerences in how they transform an objective social situation as a function of their own motives and goals (i.e., prosocials will focus on cooperative features in the interdependence situation, whereas proselfs will focus on noncooperative features; Kelley & Thibaut, 1978). Thus, if one’s motives are socially based (i.e., the social self is salient), then the interdependence situation will be perceived as a cooperative setting. Within social dilemma research it has been argued and demonstrated that reinforcing the social self (as procedural fairness does) may, in fact, have positive eVects on cooperation, because it positively influences beliefs of 4 In give-some and take-some dilemmas, individual decision makers’ level of cooperation is assessed by measuring how many chips (which all have a monetary value) each individual wishes to give to their partner (give-some) or take from a common resource (take-some). For example, the give-some dilemma task is presented in such a way that participants can make a choice between five options, varying systematically from most to least cooperative. Participants are told that they are paired with another person participating in the same study. They are informed that they are given four blue chips and that the partner with whom they will play the task (and who is unknown to them) is given four yellow chips. Each chip the participant keeps for himself or herself is worth 25 units, whereas a chip given to the partner is worth 50 units. The same logic applies for the partner: each chip the partner keeps is worth 25 units, whereas each chip the partner gives to the participant is worth 50 units.

178

DAVID DE CREMER AND TOM R. TYLER

trust and the elicitation of prosocial goals. Research on the self and relationships within groups has indeed made a similar claim. With regard to trust, research on intergroup relations has suggested that a salient social self may produce more positive expectations toward the other ingroup members, relative to outgroup members (i.e., ingroup favoritism; Tajfel & Turner, 1986). For example, in line with an ingroup favoring explanation, Brewer (1979) argued that ingroup members are judged as more trustworthy and honest than outgroup members, particularly when a socially based self or identity is salient (Brewer, 1979). As Brewer (1981, p. 356) puts it: ‘‘As a consequence of shifting from the personal level to the social group level of identity, the individual can adopt a sort of ‘depersonalized trust’.’’ With respect to adopting a prosocial goal, group dynamics and social dilemma research has argued that depending on the type of self (personal versus social), people pursue personal self-interest or the group’s interest. Indeed, Brewer (1991; p. 476) argued that, ‘‘when the definition of self changes the meaning of self-interest and self-serving motivations changes accordingly.’’ A salient social self thus implies that interdependent relationships (as observed within groups) will be more characterized by mutual concern for one’s own interest and the interest of the others. Some convincing support for this assertion that a salient and strong social self positively influences trust and prosocial goals is given by De Cremer and Van Dijk (2002), who conducted an experimental study in which participants were invited to the laboratory to play a public good dilemma game. Participants were business students and were given an endowment of 300 Dutch cents (1.50 US$) at the start of the first contribution session, and were free to contribute any amount of it to establish the common good. It was explained that the amount that had to be contributed was 1050 Dutch cents (in a group of seven individuals). If achieved, this amount would be split equally among all members regardless of their contribution. In this study, the introduction of an intergroup manipulation reinforced people’s social self. That is, participants were told in the high group identification condition that the decisions of groups of business students would be compared to the decisions made by students of the psychology department. In the low group identification condition, they were told that the individual decisions of each student within the group would be compared. Before the start of this experiment, participants’ social value orientation was also assessed (see Van Lange et al., 1997, for specific details about the measurement; see Appendix II for a sample item). Social value orientation is a stable individual diVerence variable, which refers to the value people assign to their personal welfare versus the collective welfare (De Cremer & Van Lange, 2001). The results of this study revealed an interaction showing that reinforcing identification with the group (i.e., a strong social self ) motivated

PROCEDURAL JUSTICE, SELF, AND COOPERATION

Fig. 1. (SVO).

179

Cooperation as a function of group identification and social value orientation

proselfs to contribute more, whereas prosocials were not aVected by this manipulation (see Fig. 1). Thus, identifying strongly with the group made people transform their goals and motives from the personal to the group level, consequently enhancing cooperation. Thereafter, participants were told whether their group succeeded or failed in obtaining the public good. In half of the conditions the group succeeded, whereas in the other half the group failed. Then, participants were asked again how much they were willing to contribute during a second session. Results revealed that after the group feedback a significant interaction between group identification and feedback emerged. When the group failed in establishing the public good, it was shown that those with a strong sense of group identification remained cooperative, whereas those with a low sense of group identification dropped their levels of cooperation (Fig. 2). No eVect of group identification was found when the group was successful. Also, no moderating eVect of social value orientation was observed anymore. This second analysis thus indicates that enhancing group identification must have reinforced perceptions of trust in others as well, but this only started to play a role in the second contribution session (when participants knew how the group as a whole performed and perceptions of trust could be inferred). Taken together, this important experiment makes clear that once people have a salient social self they use the group goals as their own goals and express more trust and confidence toward their fellow group members.

180

DAVID DE CREMER AND TOM R. TYLER

Fig. 2. Cooperation as a function of group identification and group feedback.

Because both the procedural fairness literature and the social dilemma literature have in common their reference to promoting people’s social self, we argue that procedural fairness is an important group tool that is able to promote cooperation because it should positively influence both psychological processes of prosocial goal activation and beliefs of trust as articulated in the social dilemma literature. In line with this suggestion, we will now review some recent studies that have indeed demonstrated that these processes underlie the positive relationship between procedural fairness and cooperation within groups.

VI. Overview of Cooperation Research In social dilemmas a strong research tradition exists examining the emergence and eVect of authorities on cooperation (Tyler & Degoey, 1995). In this literature, authorities have been referred to as structural solutions to social dilemma problems (e.g., Messick & Brewer, 1983), and the validity of this type of solution was based on the famous phrase of Hardin (1968), who argued for ‘‘mutual coercion, mutually agreed upon’’ solutions to solve the problem of cooperation. Because the interest in authorities emerged from this tradition, this type of solution has mainly been adopted to restrain people’s tendency to maximize self-interest. The reason for an instrumental perspective of cooperation is that many social scientists believe that voluntary cooperation in social dilemmas is a

PROCEDURAL JUSTICE, SELF, AND COOPERATION

181

‘‘fragile flower, easily crushed by self-interest’’ (Batson, O’Quin, Fultz, Vanderplas, & Isen, 1983), and therefore, structural leadership solutions require sanctioning of self-interested behavior. However, it has quite often been overlooked that the structural solution has to be accepted by all and, as such, needs a kind of regulation philosophy that all group members can buy into. Therefore, Hardin’s suggestion has to be made more transparent by pointing out that any coordination rules or procedures need to communicate that they are free of bias and appeal to the needs and concerns of all group members (see also Rawls, 1971). In this way, structural solutions like leaders and authorities need to incorporate procedural fairness if they are to be eVective in influencing others. Being unbiased and appealing to members’ needs fits what we have already described as procedural fairness (Lind & Tyler, 1988). As such, we argue that an accepted and legitimate authority will be perceived and evaluated as procedurally fair, which, in turn, will motivate group members to comply and cooperate with the group and its authority (see Tyler et al., 1997). Below we will describe several studies that have shown that authorities are eVective tools to promote cooperation and that in this process trust and prosocial goals play an important role.

A. PROCEDURAL FAIRNESS AND PROSOCIAL GOALS Does procedural fairness influence cooperation and does the social self play a role in this process? To answer this question, De Cremer and Van Knippenberg (2002, experiment 2) studied the eVects of leader procedural fairness (in addition to a manipulation of leader’s self-sacrifice; something we will not discuss here) on cooperation in a public good dilemma. They manipulated whether group members were given the opportunity to voice their opinion or not about how to allocate the public good. Because fair procedures are supposed to create a sense of belongingness among group members, they also predicted that these belongingness feelings would mediate the eVect of procedural fairness on cooperation. Results indicated that a procedurally fair leader indeed promoted contributions more than a procedurally unfair leader and that belongingness feelings mediated the relation between procedural fairness and cooperation. Another study by De Cremer (2003b) provided further evidence that the relational qualities of procedural fairness influence cooperation in social dilemmas. In this study, participants were told that they as a group would participate in a public good dilemma but that, before making the decision, they would be informed about the extent to which the other members in their group showed respect (i.e., respect is perceived as an important relational

182

DAVID DE CREMER AND TOM R. TYLER

quality of procedural fairness; De Cremer & Tyler, in press; Tyler & Smith, 1999). Respect versus disrespect was manipulated by giving participants the average score of the rest of the group members on a respect scale that was assessing how respectful they were toward the others (i.e., the focal participant). Results revealed that participants contributed more to the public good when the others were respectful rather than disrespectful (see also De Cremer & Tyler, 2004). Moreover, this eVect was mediated completely by participants’ feelings of belongingness. Because sense of belongingness was treated as a dependent variable rather than a moderator variable, these findings do not speak to the moderating role of the social self. Therefore, De Cremer and Van Vugt (2002, experiment 1) examined the extent to which leader’s procedural fairness influenced cooperation levels in social dilemmas as a function of group member’s sense of group belongingness. In this experiment, participants would play a public good dilemma game. After explaining the game, participants were told that a leader was appointed. Thereafter, information about leader characteristics was manipulated; that is, whether the level of group commitment of the leader was high or low (which we will not discuss here) and whether the leader was procedurally fair or not (operationalized as the opportunity to voice their opinion or not). In addition, by either introducing an intergroup comparison (between the participants’ own university and a rival university) or an intragroup comparison (between the individual group members; see also De Cremer & van Dijk, 2002) participants’ level of group identification was manipulated. On the basis of our prediction that procedural fairness communicates information relevant to people’s social self, it should be expected that particularly those individuals with a strong desire to belong (i.e., who identify strongly with the group) should care most about procedural fairness information. In line with this prediction, results revealed that fair (as compared to unfair) leaders were more eVective in enhancing contributions, particularly when group members exhibited strong group identification. Tyler and Degoey (1995) obtained similar results to De Cremer and Van Vugt (2002), but this time in a real-life setting. These authors examined people’s perceptions of the fairness of the legal authorities in California and their sense of identification with this state. At the time of the study, California was plagued by a severe drought, and as such, people had to try to maintain water resources; a situation that qualified as a social dilemma (see e.g., Kahrl, 1982; Ostrom, 1990). Results of this survey study in a real-life social dilemma situation revealed that procedural fairness perceptions significantly influenced people’s willingness to save and maintain water resources but particularly when they exhibited a strong sense of identification with the community (see also Huo, Smith, Tyler, & Lind, 1996; Tyler et al., 1996 for more evidence regarding

PROCEDURAL JUSTICE, SELF, AND COOPERATION

183

the moderating eVect of belongingness and identification on procedural fairness eVects). A final study by De Cremer (2002) showed the moderating eVect of belongingness in the context of respectful versus disrespectful treatment (see also De Cremer, 2003b for the notion that respect constitutes a relational aspect of procedural fairness). In this study a strong need to belong was installed by classifying participants as peripheral versus core group members (see Noel, Wann, & Branscombe, 1995, for a similar procedure). Peripheral group members were people that were just inside the personality category that characterized their group, whereas core group members were people that were clear examples of this personality category and thus of the group. EVorts were also made that membership in this group (which was characterized by the specific personality trait) was perceived as very attractive to all group members (regardless of the experimental condition they were in). Because of this attractive group membership, it was assumed that peripheral members should wish to gain inclusion in this attractive group (i.e., a high need to belong), whereas core group members would have not have such strong need because they were considered as very inclusive group members. Thereafter, respect was manipulated by sending participants an e-mail message that was written by the other group members and that was either disrespectful in content or respectful. The results showed that respect positively influenced contributions (i.e., contributions were higher in the respect than in the disrespect condition), but only among the peripheral group members. This result thus indicates that when people really wish to belong (i.e., because they are peripheral and wish to gain more inclusion in this attractive group) they care more about the quality of treatment with the group. Taken together, these studies show that procedural fairness and associated relational qualities of procedural fairness like respect indeed influence cooperation. In line with the goal/expectation theory (Pruitt & Kimmel, 1977), the reason for this behavioral eVect of procedural fairness to emerge was because fair procedures created a prosocial goal among followers, consequently promoting cooperation. Of course, this theory states that in addition to having a prosocial goal, beliefs about trust also need to be enhanced. In accordance with this assumption, several studies have provided evidence that procedural fairness also aVects this process of trust significantly.

B. PROCEDURAL FAIRNESS AND TRUST Tyler and Degoey (1995), in their real-life social dilemma study, also showed that trust influenced views about the legitimacy of authorities and

184

DAVID DE CREMER AND TOM R. TYLER

that these attitudes and perceptions influenced the willingness of community members to accept decisions voluntarily and to obey group rules. Thus, trust was communicated by the fairness of procedures enacted, and this process, in turn, aVected levels of cooperation. In a similar vein, De Cremer (2004a) conducted a series of studies to examine the extent to which trustworthiness mediated interactions between two elements of procedural fairness; that is, bias and accuracy (see Leventhal, 1980). Participants in this experiment were said to be part of a group led by a person known to be biased or unbiased. In addition, this leader would have to use a procedure to determine how the individual performances of each group member would be rated. In the accurate procedure condition, all the tasks that the group member performed were rated, whereas in the inaccurate condition only some of the tasks were rated. The results revealed that, if leaders were seen as unbiased, participants’ responses were influenced by the accuracy procedure (i.e., participants’ reactions were more positive when accurate relative to inaccurate procedures were used), but this was not the case when the leader was perceived as a biased person. More important, this interactive eVect between bias and accuracy on participants’ responses was largely explained by participants’ perceptions of the leader’s trustworthiness. Thus, information about leader’s procedural fairness in a group performance context clearly communicated to participants how trustworthy the authority enacting the procedures was. Van den Bos, Wilke, and Lind (1998) provided further evidence that trust and procedural fairness have a strong relation. These researchers demonstrated that if procedural fairness information is lacking (i.e., one does not know the exact procedures used), people use information about the authorities’ trustworthiness as a substitute to direct their reactions. Van den Bos and colleagues as such argue in their fairness heuristic theory that trust acts as a substitute for procedural fairness information. Furthermore, Van den Bos et al. (1998) also suggested that more research on the relation between trust and procedural fairness is required and that, in doing so, researchers should focus on types of trust studies in the context of social dilemmas. That is, ‘‘future studies may pay attention to other operationalizations of trust, and, more important, may want to try to integrate the various research domains in which trust has been studied, such as . . . social dilemmas’’ (p. 1456). This suggestion clearly demonstrates that in the justice literature, procedural fairness is seen as having clear relations with the concept of trust as it is approached and articulated in the social dilemma literature. In a related manner, Brockner et al. (1997) also showed in a series of organizational field studies that when information is given about the favorability of one’s own outcomes, procedural fairness information is translated

PROCEDURAL JUSTICE, SELF, AND COOPERATION

185

by employees in terms of trust. That is, according to Brockner and colleagues, procedural fairness acts as a source of trust. Finally, Korsgaard, Schweiger, and Sapienza (1995) examined the relation between procedural fairness and trust in another decision-making situation; that is, strategic decision-making teams. Their results revealed that providing consideration of each team member positively enhanced trust in the leader and that this relation was fully accounted for by procedural fairness perceptions. As such, this finding strongly indicates that the enactment of fair procedures promotes trust within groups, which, consequently, may promote decision making and, ultimately, cooperation.

C. CONCLUSION The above review of studies shows that the relation between procedural fairness and creating a prosocial goal and belief in the trustworthiness of group authorities is clearly supported. Thus, following Pruitt and Kimmel (1977), procedural fairness acts as an important group tool that is able to influence the two processes necessary to promote cooperation. The fact that procedural fairness is likely to influence cooperation has also been suggested by Lind (2001a), when he argues that ‘‘fair treatment leads to a shift from responding to social situations in terms of immediate self-interest, which might be termed the individual mode, to responding to social situations as a member of the larger social entity, which might be termed the group mode. . . . people in group mode are primarily concerned with what is good for the group and what they can do to reach group goals’’ (p. 67). Empirical support for this argument flows from studies indicating that experiencing fair procedures encourages commitment to groups (Tyler & Blader, 2000) and support for group authorities (Tyler, 1990; Tyler & Huo, 2002). These attitudes and values are found to shape cooperative behaviors, and such behaviors become less strongly linked to considerations of selfinterest. In other words, people’s motivations become autonomous and are linked to their own internal values. Until now, this chapter has shown that procedural fairness influences the social self of individual group members and that because of this eVect on their social self, two essential processes for promoting cooperation are likely to be influenced: trust and prosocial goals. Thus, it seems clear that the social self has taken an important place in the justice literature, and as our findings show, particularly in relation to cooperative group behavior. Before describing our motivational analysis of procedural fairness as a management tool of cooperation in terms of a SelfBased Model of Cooperation (SMC), we first elaborate further on this special focus of the social self in the justice field.

186

DAVID DE CREMER AND TOM R. TYLER

VII. The Self and Justice Research The arguments outlined in our motivational account about the psychology of cooperation do not only provide important insights into why procedural fairness relates positively to cooperation but, at the same time, help us to understand changes in the justice literature over the last several decades. In this section we focus primarily on these changes to arrive at a better understanding of how justice research has come to assign much weight to the importance of self- and identity-related concerns.

A. THE SHIFT FROM DISTRIBUTIVE TO PROCEDURAL JUSTICE Early research on justice focused on the argument that people’s feelings and behaviors in social interactions flow from their assessments of the fairness of their outcomes when dealing with others (distributive fairness). This hypothesis was widely supported. In particular, a series of experimental studies demonstrated that people were most satisfied when outcomes were distributed fairly (Walster, Walster, & Berscheid, 1978). What was most especially striking and provocative, because it contradicts self-interested predictions, were the adverse reactions by those who received more than they felt they deserved: people did not react well to being ‘‘overbenefited.’’ This finding indicates that people will give up resources and accept less when they believe doing so is fair. Other studies show that people will leave a beneficial, but unfair, situation to move to a fairer situation in which they receive fewer resources (see Tyler et al., 1997). Despite the impressive findings of early studies of distributive justice, the focus of attention among justice researchers has increasingly shifted away from studying only distributive justice to a combined focus on people’s distributive and procedural justice concerns. A number of factors have driven this shift. First, research shows that distributive justice judgments are often biased (e.g., Messick & Sentis, 1985; Ross & Sicoly, 1979; Thompson & Loewenstein, 1992). This limits the utility of distributive justice as a construct, because people will often see themselves as deserving more favorable outcomes than others see them as deserving. As a consequence, people frequently cannot be given what they feel that they deserve, and distributive justice has not proved as useful in resolving group conflicts as was initially expected (Leventhal, 1980). A greater focus on procedural justice issues was also driven by later studies that looked simultaneously at the eVect of distributive and procedural justice judgments and found a predominant influence of procedural justice

PROCEDURAL JUSTICE, SELF, AND COOPERATION

187

on people’s reactions to groups to which they belong (Alexander & Ruderman, 1987; Tyler & Caine, 1981). These studies, conducted in settings in which people had information about both distributive and procedural justice, found that procedural justice judgments typically play the major role in shaping people’s reactions to their personal experiences. More recent research echoes these findings about the relative eVect of procedural and distributive justice concerns (Tyler & Blader, 2000). In addition, when people are asked to talk about their personal experiences of injustice they are usually found to talk primarily about procedural issues, in particular about being treated with a lack of dignity and politeness when dealing with others (Messick, Bloom, Boldizar, & Samuelson, 1985; Mikula, Petri, & Tanzer, 1990). Justice research has followed the path outlined by this converging experimental and field study evidence indicating that the primary eVect on people comes from their judgments about the fairness of procedures (see Tyler et al., 1997; Tyler & Smith, 1998). This does not mean, of course, that people no longer consider issues of distributive justice but, rather, that there is a particularly strong focus on issues of procedural justice in recent research. This shift toward issues of procedure is consistent with the argument that people focus on those issues that provide them with the greatest amount of identity-relevant information because, as we have been outlining, procedural justice is heavily intertwined with issues of status, identity, and the self.

B. THE FOCUS ON ISSUES OF INTERPERSONAL TREATMENT IN PROCEDURAL JUSTICE FRAMEWORKS A second important shift in justice research has been a change in how procedural justice is conceptualized. Early work on procedural justice was guided by the influential research program of Thibaut and Walker (1975). Thibaut and Walker centered their procedural justice studies on procedures as mechanisms for making decisions about the allocation of outcomes in courtroom trials to resolve disputes about outcome fairness. In particular, they focused on formal procedures that related to decision-making processes in legal settings. So Thibaut and Walker linked their discussions of procedures primarily to issues of decision making and, in particular, to issues of decision making about allocation decisions. Because their procedural models were rooted in an era where distributive justice dominated, their focus on outcome fairness as the goal of procedures was natural. This context also influenced their theory development, because they linked people’s desire for fair procedures to their desire to achieve equitable (i.e., fair) outcomes.

188

DAVID DE CREMER AND TOM R. TYLER

They proposed that people valued procedural justice (operationalized in their research as voice or process control in the context of an adversary trial) because it facilitated the decision-makers’ ability to make equitable judgments. In other words, procedures were valued insofar as they aVected the outcomes that were associated with them. The focus on equity as a fair distributive justice principle is key, because a judge cannot make an equitable decision without having information about the parties’ contributions— information they can provide when they have their opportunity for voice. If, in contrast, equality were the principle of outcome fairness that was used, input would not be necessary, because everyone would receive the same amount. This is one reason that other justice theorists recognize the value of equality as a principle of outcome fairness (Messick, 1995). This exclusive focus on decision making in allocation contexts is no longer true of procedural justice research. Researchers have increasingly moved their attention away from a central focus on the decision-making function of procedures to include more attention to the interpersonal aspects of procedures. Those interpersonal aspects of procedures arise because procedures are settings within which people are involved in a social interaction with one another. This is true irrespective of whether the procedure involves economic bargaining, a market exchange over purchasing something, team interactions among equals, or a third-party procedure with a decision maker, such as mediation or a trial. Of course, in groups, interactions with authorities are central. In social interactions there is considerable variation in the manner in which people treat one another. They can act politely, respectfully, rudely, with hostility, and so on. These aspects of the interpersonal experience of a procedure—which occur in the context of an interaction whose overt purpose is to make a decision about how to allocate resources or how to resolve a conflict—may also influence those who are involved (which parallels our idea mentioned earlier on in this chapter that social interactions and its associated features like the use of procedures aVect the self of those involved, see e.g., Sedikides & Gregg, 2003). In other words, although people may go into an interaction with a task purpose, there are socioemotional aspects to interaction that also influence people. An example of this shift from an exclusive focus on decision-making to a focus that includes attention to the interpersonal quality of the interaction can be found in the literature on process control. In the early work of Thibaut and Walker (1975), as we have outlined, the opportunity to present evidence was linked to the desire to influence the decisions made by thirdparty decision makers. The value of the opportunity to speak was directly related to a person’s estimate of how much influence they had over the decision maker. As a consequence, in this research, people were not asked

PROCEDURAL JUSTICE, SELF, AND COOPERATION

189

about whether they were treated politely and with dignity by the decision maker: Quality of treatment was irrelevant. However, later studies of voice indicated that having the opportunity for ‘‘voice’’ had interpersonal or ‘‘value-expressive’’ worth that was not linked to any estimated influence over the decisions made (Tyler, 1987). These studies showed that people still rated a procedure to be more fair if they had voice than if they lacked voice, even if they estimated that what they said had little or no influence on the decisions made (Tyler, Rasinski, & Spodick, 1985). This was true even in the most extreme case that can be created for communicating that voice has no instrumental value—when the opportunity for voice came after the decision was already made (De Cremer & Van den Bos, 2003; Lind et al., 1990). These findings indicate that voice has value beyond its ability to shape decision making processes and outcomes. What factors drive the influence of voice, even when it clearly cannot aVect the eventual outcome or decision? If an authority listens to people’s arguments, we might hypothesize that people think that the authority was conferring interpersonal respect on that person. This argument was supported by the finding that people only value such voice opportunities if they feel that the authority is ‘‘considering’’ their arguments (Tyler, 1987). This indicates that people were focused on whether or not they had their concerns and were treated respectfully by being taken seriously by the decision maker, independent of whether or not the course of action they recommended to resolve those concerns was adopted. Other research on people’s procedural justice concerns directly measures people’s focus on the quality of their interpersonal treatment (‘‘standing’’ or ‘‘status recognition’’) and finds that it has an eVect that is distinct from their interest in the fairness of decision-making judgments. Drawing on these findings, the relational model of authority (Tyler & Lind, 1992) explicitly includes issues of interpersonal treatment within the framework of procedural justice concerns. The relational model, therefore, directly recognizes the importance of interpersonal treatment. Subsequent studies confirm that issues of interpersonal treatment or standing independently shape procedural justice judgments (De Cremer & Sedikides, 2004; De Cremer & Tyler, 2004; Tyler, 1994; Tyler & Huo, 2002; Van Prooijen, Van den Bos, & Wilke, 2002). In fact, issues of interpersonal treatment are often found to be the most important factors shaping procedural justice assessments. These interpersonal aspects of procedures have been found by recent studies to be so powerful in their eVect that some researchers have argued that they might potentially be treated as a separate type of ‘‘interactional’’ justice (Bies & Moag, 1986; Tyler & Bies, 1990). Irrespective of whether the quality of the treatment that people experience via procedures is actually considered a distinct form of justice (see Blader & Tyler, 2003a,b),

190

DAVID DE CREMER AND TOM R. TYLER

justice researchers have again followed their findings about what most strongly aVects the people they study by increasingly turning their research toward exploring interpersonal or interactional aspects of procedures. Again, this shift in focus is consistent with the arguments outlined here. Because people are focused on issues of the social self and identity, their primary concern is with the self-relevant implications of their experiences with others. Interpersonal respect is central to status messages and is therefore important to people in group settings. This focus can be contrasted to the focus of early research on decision making. That early focus assumed that the key issue that shaped reactions to procedures was the manner in which a decision was made (i.e., the ability of a procedure to produce a fair outcome). Here the focus is on the experience that people have when they are dealing with others (i.e., whether their identity and status are aYrmed or undermined by the treatment they receive from others).

C. MOVING FROM ANGER AND NEGATIVE BEHAVIORS TO POSITIVE ATTITUDES/VALUES AND COOPERATIVE BEHAVIORS Early research on justice was grounded in the literature on relative deprivation, a literature whose origins lie in eVorts to understand and explain riots, rebellion, and other forms of collective action (Crosby, 1976; Gurr, 1970). This focus on negative attitudes and behaviors continued in later eVorts to understand distributive influences on pay dissatisfaction, employee theft, sabotage, turnover, and resistance to third-party decisions and rules (Tyler & Smith, 1998). However, recent research on procedural justice increasingly focuses on more prosocial actions, such as how to build trust, encourage responsibility and obligation, generate intrinsic motivation and creativity, and stimulate voluntary cooperative behavior in interactions with others (Tyler & Blader, 2000). Similarly, there has been increasing attention to exploring when justice motivations encourage people to provide resources to the needy and disadvantaged (Montada & Schneider, 1989). Interestingly, this shift is consistent with a shift that has been taking place within psychological research more generally (Snyder & Lopez, 2002). In recent years the positive psychology movement has focused the attention of psychologists on the goal of stimulating desirable behavior and encouraging high self-esteem, happiness, and favorable mental health (Cameron, Dutton, & Quinn, 2003). This broadening of the focus of justice research is consistent with the argument that justice theories provide a basis for understanding the nature of people’s relationship to groups. That includes both people’s negative

PROCEDURAL JUSTICE, SELF, AND COOPERATION

191

reactions to injustice and the ability of experiencing justice to motivate their engagement and cooperation. Society, after all, does not just want people not to riot or destroy. It also wants them to be satisfied and cooperative. In particular, we are interested in cooperation of two forms—rule following and productive eVort on behalf of groups (Tyler & Blader, 2000, 2003b). While continually supporting the basic importance of people’s justice judgments, these shifts in focus have also resulted in a dramatic change in the character of justice research since the 1960s. In fact, early justice researchers might have trouble recognizing many recent justice studies as being about justice—at least as they originally understood that construct. Instead of viewing justice as residing in the rules used in the distribution of resources in a group (Leventhal, 1980; Thibaut & Walker, 1975), justice is more recently viewed as being strongly linked to issues related to the quality of interpersonal treatment, such as treating people with politeness and dignity in social interactions. It is also focused on stimulating commitment and cooperation, rather than minimizing anger and destructive behaviors. We argue that these changes—which were guided by the empirical results of justice research—can best be understood by considering the psychological dynamics underlying justice (i.e., we need to understand why people care about justice). These changes can be explained by considering the psychological processes that lead people to react to issues of justice or injustice when they are dealing with others. Much early justice research was focused on showing that justice matters; that is, in demonstrating that people’s thoughts, feelings, and behaviors are shaped by their justice judgments, indicating that information about justice is central to people’s evaluations of social situations (Tyler et al., 1997; Van den Bos & Lind, 2002). However, to develop a deeper understanding of why these eVects emerge—and why the shifts in research focus we have outlined have occurred—we need to pay attention to the psychology underlying justice. Several models have been proposed to understand the psychology underlying procedural justice. We focus here on a set of models that share an emphasis on the relational implications of justice evaluations—the groupvalue model (Lind & Tyler, 1988), the relational model (Tyler & Lind, 1992), and the group engagement model (Tyler & Blader, 2000, 2003b). These models represent a programmatic research program designed to understand the psychology of justice. The models diVer first in their focus. The group-value model focuses on the psychology underlying judgments of procedural justice. The relational model explores the factors shaping reactions to authorities and rules within groups, organizations, and societies. The group engagement model is concerned with psychological and behavioral engagement in groups (i.e., with cooperation). The models also diVer in their predictions. The

192

DAVID DE CREMER AND TOM R. TYLER

group-value model predicts that noninstrumental factors will influence procedural justice judgments, a prediction confirmed both by findings of noninstrumental voice eVects (Lind et al., 1990; Tyler, 1987), and by demonstrations that people care more about issues of procedural justice when dealing with members of their own groups (Tyler, 1999). The relational model predicts that procedural justice will influence reactions to authorities, as has been subsequently found by studies of legal, political, managerial, familial, and educational authorities (Tyler & Smith, 1998). It further predicts that relational concerns—in particular neutrality, trustworthiness, and status recognition—will influence procedural justice judgments, an argument supported by a number of studies (Tyler, 1989, 1994; Tyler et al., 1996). The group engagement model identifies the antecedents of attitudes, values, and cooperative behavior in groups, positing a general model of the relationship between people and groups (Tyler & Blader, 2000, 2003b). The group engagement model argues that identity judgments will be the primary factors shaping attitudes, values, and cooperative behaviors in groups. It further suggests that two factors shape identity judgments: evaluations of resources and assessments of procedural justice. This model is shown in Fig. 3. Each of these group engagement predictions has received empirical support (Tyler & Blader, 2000). Those members of work groups who identify more strongly with their groups are more psychologically and behaviorally engaged in those groups. They have more favorable attitudes and more supportive values and are more behaviorally engaged in voluntary cooperation—including both voluntary deference to rules and engagement in extrarole behaviors. Further, neither resource-based connections to groups nor procedural justice judgments directly shape engagement when identification

Fig. 3. The group engagement model.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

193

is controlled for. Identification, in turn, is influenced both by resource-based connections and by the procedural justice of the group. The support found for the group engagement model is important because it is consistent with the core argument we are making here. The key factor shaping cooperation is the self. The impact of both procedural justice and of resources on cooperation occurs because those organizational factors shape the connection between the individual and the group—they shape the self. Two aspects of the self are important. First, people’s feelings about the status in relationship to the group. One such feeling is their feeling about the status of the group (pride). People are more strongly invested in a high status group. Second, their feelings about their status in the group (respect). People are more strongly invested in a group in which they have high status.

D. SUMMARY To sum up, justice research has been around quite some time now and currently seems to have arrived at a new point in its existence that puts a focus and emphasis on understanding the psychology of the eVect of procedures on people’s thinking, feeling, and behavior within the context of social interactions. One important consequence of this focus is the observation that procedural fairness has major implications for positive group behavior like cooperation. As we have shown, the psychological underpinnings of this eVect on cooperation seems to reside within the concept of people’s social self—constituting a variety of self-related concerns regarding belongingness, reputation, identity, and uncertainty. In the following section, we will describe the social psychological model that we put forward in explaining why the fairness of procedures matters in promoting cooperation.

VIII. Summary and Conclusions Within groups, organizations, and societies, people have their own goals and behaviors that do not always align well with the aims of the higher collective, and as such constituting a mixed-motive situation. A consequence of this conflict in motives is that people do not always consider the collective welfare, the obligation to follow rules, acts of voluntarism—which can all be classified as acts of cooperation—to be important to their own values, motives, and self-definitions. In other words, cooperation is often thought of as conflicting with people’s views of their own self-interest. Thus, for social psychologists the important question becomes understanding how

194

DAVID DE CREMER AND TOM R. TYLER

cooperation with others can be viewed by people as self-relevant and selfmotivating and creating conditions under which people will more easily internalize the value and importance of cooperative acts. Earlier research mainly focused on the eVects of punishment and rewards to increase the attractiveness of cooperation. For example, studies have shown that cooperation increases if the payoV for the cooperative choice becomes more attractive and the payoV for the defecting choice less attractive (see Van Lange, Liebrand, Messick, & Wilke, 1992). Despite the fact that such monitoring and sanctioning systems may work to some extent, the potential side-eVects are that they undermine intrinsic motivation (Deci & Ryan, 2000). As a consequence, once such instrumental (or self-interest based) incentives are removed, the level of cooperation is likely to drop again and will fall to even lower levels if intrinsic motivation has been undermined (Mulder, van Dijk, De Cremer, & Wilke, 2003). Thus, sanctioning or monitoring systems targeted at individuals’ selfinterest do not motivate people to value cooperation for intrinsic reasons. As a result, their use commits a group to a never-ending need to have available the resources to reward desirable behavior and to create and maintain a credible sanctioning system that will discourage rule breaking through the provision of sanctions. These mechanisms are costly and ineYcient and constitute a constant drag on the ability of groups to deploy their resources in ways that are most adaptive for the group. Further, groups are least likely to have the ability to provide incentives or deploy sanctions during times of crisis or change, when cooperation is most needed for the group to survive and flourish. During a crisis, for example, groups need for all members to focus on group needs and group survival. Yet during wars, economic downturns, or natural disasters, it is least possible to provide immediate incentives for cooperation. If everyone pursues their own immediate self-interest in such times the viability of the group is threatened. Even when societies are not threatened by immediate events, the motivation of cooperation by incentives and sanctions is a costly and ineYcient mechanism. Incentives are problematic because of their ineYciency. To provide incentives to cooperate it is necessary to specify desired behavior in advance, so that people will know what to do and authorities will know what to reward. However, many of the behaviors that are desirable in groups require people to use their discretion—acting as is appropriate to a unique situation. In work settings we value employees who engage in voluntary behaviors that solve immediate problems; in communities we value people who step forward to act in resolving the immediate issues within the community. These behaviors are hard to specify and incentivize in advance. Because it is hard to know what to tell people they will be rewarded for doing in advance, incentive systems are generally not a good way to motivate

PROCEDURAL JUSTICE, SELF, AND COOPERATION

195

behavior. They work when the behavior desired can be specified in advance (i.e., with clearly specified and repetitive jobs such as work on assembly lines). Sanctions are ineYcient because their use requires the authorities to constantly maintain the credible threat of sanctioning. Research suggests that it is the likelihood of punishment, rather than the anticipation of severity in punishment, that most strongly shapes behavior (Doob & Webster, 2003; Tyler, 2003). Hence, sanctioning requires widespread surveillance. In the modern world people are familiar with surveillance cameras in parks, at entrances to buildings, and even in bathrooms, just as there is a ubiquitous presence of police oYcers and private security personnel in many communities and organizations. Such eVorts are costly for groups, organizations, and societies and produce only minimal decreases in the level of undesirable behavior (Tyler, 2003). Finally, as we have noted, the use of incentives and sanctions has the longterm eVect of undermining internal motivations for cooperation. When people work for incentives, their intrinsic motivation (i.e., the degree to which they work out of enjoyment of work) is undermined. When people comply out of fear of punishment, their feelings of responsibility and obligation to obey are undermined. Hence, it is not only that these approaches do not take advantage of the motivating power of internal motivations but they actually undermine these other sources of motivation for cooperation. Over time the use of incentives and sanctions can create a situation in which, however costly or unwiedly they may be, it is not possible to manage groups in any other way, because other motivations for cooperation have been diminished or extinguished. To be able to change the approach taken to motivating cooperation it is necessary to increase our understanding of particular group means that directly aVect people’s needs and values. We argue that these are linked to the concept of the social self (Tyler & Smith, 1999). If the social self is positively influenced, then people will be motivated to cooperate for the group. One relevant group means that we have discussed for shaping the social self is the fairness of procedures that are enacted by the group and its representative authorities. In the prior sections, the studies discussed have provided clear evidence that fair procedures aVect people’s social self by influencing concerns of social reputation, need for belongingness, feelings of self-uncertainty, and finally, self-construals (i.e., the self as being interdependent or not). In other words, fair treatment makes a positive contribution to people’s sense of themselves as social beings, thereby installing a salient and positive social self. Further, if people evaluate their relationships, interdependence, and relatedness with others as positive and rewarding, those others are likely to

196

DAVID DE CREMER AND TOM R. TYLER

Fig. 4. A presentation of a ‘‘self-based model of cooperation.’’

be perceived as positive and more attractive (Brewer, 1979). Consequently, cooperative behavior toward others will be facilitated. Our description of social dilemma research showed indeed that if a positive attitude and attachment toward the group and its members exists, cooperation rates increase significantly. The reason for this eVect is that a salient social self motivates group members to merge the personal self into the group, consequently transforming personal goals into collective goals (De Cremer & Van Vugt, 1999; De Cremer et al., in press), and to increase trust in the group (and its members; Brewer, 1979; De Cremer & Van Dijk, 2002). In addition, these transformation processes may also be associated with an increase of trustworthiness and feelings of morality (e.g., Kramer & Goldman, 1995; Rusbult & van Lange, 1996). To conclude, fairness of procedures can be seen as a useful group tool to manage the social self and, consequently, facilitating the psychological processes that are believed to underlie the emergence of cooperation, that is, trust and goal-transformation (see De Cremer & Stouten, 2003; Pruitt & Kimmel, 1977). Therefore, in this chapter, we refer to the psychology of this relation between procedural fairness and cooperation as a ‘‘Self-Based Model of Cooperation (SMC)’’ (see Fig. 4).

A. THIS CONTRIBUTION TO THE LITERATURE Our approach, as presented in this chapter, contributes in a variety of ways to the existing social psychology literature on justice and cooperation. First of all, as we mentioned earlier, many of the psychological models addressing justice have a strong cognitive flavor (Lind, 2001a). Although these models have been very eVective in unraveling justice problems and stimulating exciting experimental and field studies, we suggest that a cognitive framework is too limited and that motivational accounts of justice issues also need to be examined. After all, both cognition and motivation act

PROCEDURAL JUSTICE, SELF, AND COOPERATION

197

in tandem when it comes down to predicting behavior and people’s aVective life (Higgins & Kruglanski, 2001), and as such it is clear that a complete understanding of the psychology of procedural justice involves an emphasis on both the dimensions of cognition and motivation. Moreover, our proposition that the fairness of procedures intrinsically motivates people to engage in cooperative behavior by its influence on the social self integrates well with other motivational analyses as we know them in the field of social psychology. For example, self-determination theory recognizes that if people assimilate identities into the self they are intrinsically motivated to pursue the associated goals (Ryan & Deci, 2003). This perspective argues that motivation flowing from self-conception is intrinsic to the individual, thereby suggesting that the eVectiveness of procedural fairness influencing group members’ motivation and behavior through their social self can be assumed to be less contingent on monitoring and external rewards (cf. Deci & Ryan, 2000). In a similar vein, our studies have shown that making the social self salient (by means of procedural fairness) motivates people to adopt collective goals, consequently eliciting voluntary cooperation (e.g., De Cremer et al., in press). Further, research on the social self has been well-grounded in the motivational tradition by its explicit foci on processes like self-enhancement, self-accuracy, and self-verification (Sedikides & Strube, 1997), both on the individual and collective level (Tesser, Stapel, & Wood, 2002). Finally, as mentioned earlier, our present conception of procedural fairness also emphasizes the importance of social interaction by linking the concept of justice to issues like treatment and social behavior. Along related lines, researchers on social interaction recently argued that more motivation and less cognition is required to examine the determinants and consequences of social interactions (see Kelley, Holmes, Kerr, Reis, Rusbult, & van Lange, 2003). As a matter of fact, Zajonc (1998) even concluded that social psychologists ‘‘need to look less at the mind and more at interactions.’’ Taken together, the present motivational account can thus be seen as an important addition to the existing justice literature. In our opinion the next logical step to take seems then to be to integrate both motivation and cognition into one dynamic social model of fairness. Second, by focusing on the relation between procedural fairness and cooperation, this chapter presents a theoretical and conceptual integration between two domains of scientific research that have existed independent from one another; that is, procedural fairness and social dilemmas. In their review, Tyler and Dawes (1993) already made the suggestion that the fields of procedural fairness and social dilemmas should be urged to communicate as they both have a common focus on the role of instrumental and noninstrumental concerns. That is, in both literatures it is assumed and shown that

198

DAVID DE CREMER AND TOM R. TYLER

not only economic but also relational concerns drive people’s preferences for fair procedures and cooperative behavior. Also, Lind (2001a) suggested that the noninstrumental or relational component of procedures will be particularly important when people are concerned about potential problems with social interdependence; a situation that he referred to as the fundamental social dilemma (see also De Cremer & Van Knippenberg, 2003). This theoretical integration thus is important because it backs up recent suggestions that ‘‘fair treatment leads to a shift from responding to social situations in terms of immediate self-interest, which might be termed the individual mode, to responding to social situations as a member of the larger social entity, which might be termed the group mode. . . . people in group mode are primarily concerned with what is good for the group and what they can do to reach group goals’’ (Lind, 2001a, p. 67). Further, it provides a simple and elegant answer to recent questions highlighted in the justice literature regarding the predictive role of procedural fairness with respect to behavioral responses. That is, recently, Greenberg (2001, p. 254) argued that the justice literature is ‘‘hard-pressed to tell exactly what form a response might take.’’ This model provides an answer to this concern by showing that fair procedures motivate group members to behaviorally respond in a cooperative manner and by identifying the psychological processes underlying this response to justice. Finally, this model can also be considered as parsimonious because it reduces complexity surrounding the relation between procedural fairness and positive behaviors like cooperation by highlighting the role of similar psychological processes across diVerent research traditions examining cooperation (a task much needed in the social justice field; see Colquitt & Greenberg, 2001, for a comment).

B. IMPLICATIONS AND SUGGESTIONS FOR FUTURE RESEARCH 1. Social Self, Procedures, and Emotions Our motivational approach to understanding the relation between procedural fairness and cooperation can also be seen as having implications for the role that emotions may play in this process. In fact, Tesser, Wood, and Stapel (2002, p. 6) recently argued that ‘‘emotion is an integral facet of motivation.’’ How could emotions play a role in our line of thinking? In this chapter we started from the assumption that the self is socially constructed by, among other things, the fairness of procedures. Relevant to this point of view is Haidt’s (2001, p. 197) recent argument—derived from his social intuitionist model in the justice area—in which he stresses ‘‘the importance of social

PROCEDURAL JUSTICE, SELF, AND COOPERATION

199

interaction as the best way to trigger new appraisals.’’ From this, it follows that our motivational approach of understanding the positive eVect of procedures on cooperation by focusing on the regulating role of the social self may best be served by also including a focus on the intuitive or emotional processes accompanying this regulating process. Including an emotional component in procedural fairness models is also important, as it responds to recent claims that justice research has largely overlooked the importance of emotions in understanding consequences of justice procedures (Bies & Tripp, 2002; Weiss, Suckow, & Cropanzano, 1999). Indeed, Weiss et al. (1999) even argue that ‘‘the relative lack of empirical research assessing emotional reactions to conditions of unfairness is a serious omission’’ (p. 786). In a similar vein, Bies and Tripp (2002) state that ‘‘to understand justice in organizations, one must understand the events that arouse the sense of injustice—the emotions of injustice’’ (pp. 204–205). Thus, clearly more justice research is required that includes aVective reactions (Tyler & Smith, 1998). Responding to this call, we therefore find it important to point out the importance of emotions in motivational accounts of procedural fairness. More precisely, we support the notion that the social self—as it is situated in the context of social interaction—has a strong aVective component that, in turn, may significantly aVect self-regulation and behavior. For example, research on emotions has suggested that emotions indeed are ‘‘intimately connected to our relationships with others’’ (Tangney, 1999, p. 543). As a matter of fact, Barrett (1995) regularly refers to self-conscious emotions like pride, shame, guilt, and embarrassment as ‘‘social emotions’’ because they are elicited within the context of social interactions. Relating this to our present model, it may be easy to see that emotions such as, for example, embarrassment can be linked to our social self concern of social reputation (cf. Miller, 1995). Thus, it seems clear that a wide range of emotions exist involving the concept of the self (Lewis, 1992, 1995), both emotions reflecting the global self-concept (e.g., shame; Leith & Baumeister, 1998; Tangney, 1995)—which are associated with processes like lower self-esteem and enhanced feelings of social exclusion—and self-evaluative emotions (e.g., guilt) that are related to more specific behaviors, like, for example, cooperating and even overcompensating after having defected in a cooperative environment and feeling guilty about it. Finally, a focus on emotions is something that is incorporated by other motivational models in social psychology as well. For example, regulatory focus theory (Higgins, 1998) assumes that people have the need to strive for gains and safety. As a consequence, people may have a promotion focus aimed at pursuing hopes and aspirations (ideal self) or a prevention focus aimed at pursuing duties and obligations (ought self). Studies (e.g., Higgins,

200

DAVID DE CREMER AND TOM R. TYLER

Shah, & Friedman, 1997) have shown that a promotion focus is associated with dejection-related emotions (cheerfulness) and a prevention focus with agitation-related emotions (quiescence). Thus, ‘‘regulatory focus oVers a motivational distinction with specific evaluative, emotional, and behavioral consequences’’ (Shah, Brazy, & Higgins, 2004), and therefore, we argue that a motivational integration of the concepts of procedures, social self, and cooperation inevitably may also benefit by linking it to an emotional component. As such, future justice research is, more than ever, urged to include such aVective dimensions. 2. Social Self, Procedures, and Leadership Our analysis also has the potential to contribute significantly to the leadership literature. Indeed, in many fairness studies (including our own) procedures are enacted by an authority. However, it is striking that the justice literature has hardly devoted any attention to how insights of leadership may help unraveling further the psychology of procedural fairness. In fact, within social sciences both literatures hardly refer to one another, whereas it is clear that leaders are expected to act fairly and that enactment of procedures often includes a form of leadership style (De Cremer & Alberts, 2004; De Cremer et al., in press). This lack of integration is, for example, illustrated by statements such as that ‘‘social justice research typically has not been thought of as being research about social influence’’ (Tyler, 2001, p. 69) and that ‘‘the role that justice plays . . . in paradigms of leadership . . . has only recently begun to receive research attention’’ (Pillai, Scandura, and Williams, 1999, p. 763). Because leaders are expected to motivate groups and their members to go beyond their own self-interest (Bass, 1985; Burns, 1978), procedural fairness could be an important tool to manage this aim. In light of this, it is noteworthy that the assumption of our SMC model that procedures aVect people’s social self indeed parallels very recent thinking in the leadership literature that eVective leadership depends strongly on how leaders regulate and shape followers’ self (Tyler, 2004a,b; Van Knippenberg, Van Knippenberg, De Cremer, & Hogg, in press). For example, recent articles in Advances of Experimental Social Psychology (Hogg & Van Knippenberg, 2003) and Research in Organizational Behavior (Van Knippenberg & Hogg, 2003) present evidence that leaders often take the positions of entrepreneurs of self and identity (cf. Reicher & Hopkins, 2001). Moreover, in an attempt to combine both the leadership and procedural fairness literatures, De Cremer (2004b) recently introduced Leader Fairness Theory (LFT). According to LFT, the important question in the procedural

PROCEDURAL JUSTICE, SELF, AND COOPERATION

201

fairness literature is to ask when procedures actually exert their strongest influence. One obvious way to examine this is to look at the behavior that is involved when enacting the procedures; that is, to identify the influence of leader behavior (see Tyler, 2004a,b). LFT makes the argument that diVerent leadership styles influence followers’ social self and, consequently, because of the activation of people’s social self, procedural fairness is believed to influence followers’ reactions more strongly. For example, De Cremer (2004c) showed that self-confident leaders install a sense of self-eYcacy and autonomy, and that by activation of this social self–related dimension (cf. Deci & Ryan, 2000), participants’ emotional reactions (e.g., disappointment) were more strongly influenced by procedural fairness (i.e., a voice manipulation) (Folger, 1977). Further, another study by De Cremer et al. (in press) also showed that a leader aimed at rewarding group members for a job well done installed a sense of competence and autonomy—all believed to enhance one’s social self-esteem. As a consequence, social self-esteem was promoted and procedural fairness exerted stronger influence (see De Cremer, 2003a; De Cremer, Van Knippenberg, van Dijke, & Bos, 2004; Vermunt, Van Knippenberg, Van Knippenberg, & Blaauw, 2001, for evidence that procedural fairness eVects are stronger among individuals with high social self-esteem). All of this indicates that our findings demonstrating that procedural fairness has implications for one’s social self and as such promotes cooperation also have implications for leadership and management. Leaders can thus be seen as important organizational and group tools that can aVect social self in such a way that procedures will reveal stronger eVects on valued group and organizational outcomes like cooperation. This approach leads to a strategy of leadership explicitly designed to encourage voluntary cooperation. This strategy is process-based leadership (Tyler, 2004a,b). It is based on the argument that people’s willingness to work on behalf of leaders is directly connected to their assessments of the fairness by which those leaders exercise their authority. If leaders lead in ways that followers experience as being fair, those followers become motivated to accept the leader’s directions and to work on behalf of the group in ways that go beyond those motivated by their own self-interest. Hence, leaders can lead by the manner in which they exercise their authority. 3. Procedural Justice, Self, and Regulation The other aspect of cooperation involves the ability of authorities to gain voluntary cooperation with rules and decisions that restrict and regulate the ability of individuals to act in self-interested ways. Whether the

202

DAVID DE CREMER AND TOM R. TYLER

issue is consuming resources (Tyler & Degoey, 1995) or ignoring rules and the directives of authorities (Tyler, 1990; Tyler & Huo, 2002), being an eVective authority requires the ability to gain ‘‘buy in’’ even when decisions are not in the person’s immediate self-interest. As we have outlined, procedural justice is central to gaining voluntary cooperation, and that centrality is mediated by legitimacy. When leaders are viewed as acting fairly, people’s self-conception involves feelings of obligation and responsibility to defer to those authorities because they are legitimate and entitled to be obeyed (see Sunshine & Tyler, 2003). Consistent with this identity-based view of legitimacy, studies indicate that people are more influenced by their social values when identity issues are more salient (Skitka, 2003; Tyler, 1997). The importance of this perspective lies in the possibility of engaging people’s sense of self in the task of regulation. Rather than regulating by the threat or application of sanctions, people can be motivated to self-regulate if their personal values and identity are linked psychologically to the needs and concerns of the group. When this happens, people take personal responsibility for supporting the rules of the group, and it is not necessary to threaten punishment for noncompliance. Because the self is activated procedurally, this leads to a strategy of regulation similar to the model of leadership we have already outlined. That strategy is process-based regulation. Tyler and Huo (2002) explore the possibility of process-based regulation in a sample of people with recent personal experiences with the police and the courts. Their findings suggest that reactions to authorities are strongly based on the procedures by which they exercise their authority. Further, as hypothesized by an identity-based model, people were more strongly influenced by procedural judgments when they were more highly identified with the group that the authorities represent—in this case American society. When people are more identified with American society their identity is more strongly intertwined with status in that society, and their treatment is more relevant to their identity. In this situation, as we would predict, people are more strongly influenced by the nature of their treatment by authorities. Conversely, those whose identities are less strongly intertwined with the group focus more heavily on the favorability or fairness of their outcome when deciding whether or not to defer to legal authorities.

C. CONCLUSION Social psychology shares with other social and policy sciences an interest in understanding how to motivate cooperative behavior on the part of the

PROCEDURAL JUSTICE, SELF, AND COOPERATION

203

people within groups, organizations, and societies. This concern is reflected in social psychology in the literature on social dilemmas—a literature focused on how to obtain cooperation in social settings. In particular, the social sciences are united by an interest in organizational design. We all want to understand how to structure social situations so as to most eVectively promote cooperative behavior among the people within them. Within social psychology, one of the most useful literatures has been the procedural justice literature, because cooperation has been found to be linked to procedural justice judgments, and procedures are a core element in the design of organizations. The value of securing cooperation is important in groups, organizations, and societies because when interacting with others people often find that they are in a mixed-motive situation in which, to some degree, their selfinterest is consistent with the interests of others, leading to the motivation to cooperate, and that, to some extent, their interests diVer from those of others, leading to the motivation to compete. As a result, people are motivated both to act in ways that also benefit others and to act in ways that maximize their own self-interest at the expense of the interests of others. People must balance between those two conflicting motivations when shaping their cooperative behavior. Social psychologists have examined how people manage this motivational conflict by exploring the psychological dynamics underlying cooperative behavior in interpersonal situations that range from dyadic bargaining to long-term relationships (Rusbult & Van Lange, 2003; Thibaut & Kelley, 1959). The problem of cooperation in mixed motive dyads also lies at the root of classical problems in economics, problems such as the prisoner’s dilemma game and the ultimatum game (Poundstone, 1992). Mixed-motive conflicts within groups and societies have also been studied by social psychologists within the literature on social dilemmas. This literature asks how people deal with situations in which the pursuit of short-term self-interest by all of the members of a group leads, in the long run, to damage to the self-interest of all (see Kopelman et al., 2002). 1. Real-World Cooperation The issue of cooperation is not confined to games and experiments. It is also central to many of the problems faced by real-world groups, organizations, and societies (Van Vugt, Snyder, Tyler, & Biel, 2000). As a result, the fields of law, political science, and management all seek to understand how to most eVectively design institutions that can best secure cooperation from

204

DAVID DE CREMER AND TOM R. TYLER

those within groups. Their eVorts to address these issues are informed by the findings of social psychological and economic research on dyads and small groups. Within law a central concern is with how to eVectively regulate behavior so as to prevent people from engaging in actions that are personally rewarding but destructive to others and to the group—actions ranging from illegally copying music and movies to robbing banks (Tyler, 1990; Tyler & Huo, 2002). In addition, the police and courts need the active cooperation of members of the community to control crime and urban disorder by reporting crimes and cooperating in policing neighborhoods (Tyler & Huo, 2002). Hence, an important aspect of the study of law involves seeking to understand the factors shaping cooperation with law and legal authorities. Government also wants people to cooperate by participating in personally costly acts ranging from paying taxes to fighting in wars. Further, it is also important for people to actively participate in society by voting, working to maintain their communities by working together to deal with community problems, and otherwise helping the polity to thrive. For these reasons, understanding how to motivate cooperation is central to political scientists. Work organizations seek to prevent personally rewarding, but destructive, acts such as sabotage and stealing oYce supplies by creating and encouraging deference to rules and policies. They also encourage positive forms of cooperation like working hard at one’s job and contributing extra role and creative eVorts to one’s work performance (Tyler & Blader, 2000). For these reasons a central area of research in organizational behavior involves understanding how to motivate cooperation in work settings. 2. Motivating Cooperation The literature on cooperation suggests that the use of incentives and sanctions can eVectively shape cooperative behavior. However, although eVective, rewards and punishments are not a particularly eYcient mechanism for shaping behavior for reasons that we have already outlined. First, their eVect on behavior is marginal. Further, these eVects are costly to obtain because organizations must commit considerable resources to the eVective deployment of incentive and sanctioning systems. For these reasons, the adequacy of instrumental approaches to motivating cooperation has been questioned within law (Tyler, 1990; Tyler & Huo, 2002), political science

PROCEDURAL JUSTICE, SELF, AND COOPERATION

205

(Green & Shapiro, 1994), and management (PfeVer, 1994; Tyler & Blader, 2000). A key contribution of social psychology is the suggestion that there are social motivations that can supplement instrumental motivations in securing cooperation within organizations. Our goal is to explore one of these social psychological mechanisms—procedural justice. Our concern is with why it has the important cooperation-enhancing properties we have outlined. In other words, we want to provide a psychological framework through which it is possible to understand why procedural justice plays such a central role in shaping people’s psychological and behavioral engagement in groups, organizations, and societies. People could potentially pay attention to a wide variety of aspects of their social experiences when reacting to others, and it is striking that they focus their attention on the fairness of the procedures they experience when dealing with other people, groups, and organizations. We might intuitively think that people would focus on their outcomes, and such an intuition would be consistent with early psychological models from both the cooperation and the justice literatures. However, recent work emphasizes the degree to which people focus on issues that are at best indirectly linked to their outcomes. Central to our argument is the suggestion that people seek and are influenced by information that speaks to their identities. This is the case because one key function that groups provide is to help people to develop and maintain an identity, with information from others having an important influence on the favorability of that identity. When dealing with others, people are sensitive to information that is relevant to aYrming or undermining their identities, and procedural information is widely found to be central to people’s assessments of their self-worth and, consequently, to their self-esteem. Hence, procedures are linked to the self, and the self is linked to cooperation.

Acknowledgments The first author was supported by a fellowship of the Netherlands Organization for Scientific Research (NWO, Grant No. 016.005.019).

206

DAVID DE CREMER AND TOM R. TYLER

Appendix I

Type of scale Need to belong (Leary et al., 2001)

Concern for reputation (De Cremer & Tyler, 2004)

Interdependent self-construal (Singelis, 1994)

Self-doubt (Oleson et al., 2000)

Self-concept clarity (Campbell et al., 1996)

Number of items 10

7

12

8

12

Sample items

Range of scale

‘‘If other people don’t seem to accept me, I don’t let it bother me’’ (reverse-scored) and ‘‘I do not like being alone’’

1 ¼ not at all characteristic of me, 5 = extremely characteristic of me

‘‘I find it important that others consider my reputation as a serious matter,’’ ‘‘I try hard to work on my reputation (in my relationships with others),’’ and ‘‘I am rarely concerned about my reputation’’ (reverse scored)

1 ¼ not at all characteristic of me, 5 = extremely characteristic of me

‘‘I often have the feeling that my relationships with others are more important then my own accomplishments,’’ and ‘‘My happiness depends on the happiness of those around me’’

1 ¼ strongly disagree, 7 ¼ strongly agree

‘‘When engaged in an 1 = not at all important task, most of characteristic of my thoughts turn to bad me, 5 = extremely things that might happen characteristic of me (e.g., failing) than to good’’ and ‘‘I sometimes find myself wondering if I have the ability to succeed at important activities’’ ‘‘My beliefs about myself often conflict with one another’’ (reverse score) and ‘‘In general, I have a clear sense of who I am and what I am’’

1 ¼ not at all characteristic of me, 5 ¼ extremely characteristic of me

PROCEDURAL JUSTICE, SELF, AND COOPERATION

207

Appendix II Participants are asked to imagine that they have to allocate points (which have a monetary value like, for example, 1 US$) between themselves and an anonymous other. They have a choice between three alternatives (A, B, or C) and have to make this decision nine times. In the social psychology literature these choices are framed in terms of decomposed games. Below, an example of a decomposed game is presented. A

B

C

You

560

500

500

Other

300

500

100

References Alexander, S., & Ruderman, M. (1987). The role of procedural and distributive justice in organizational behavior. Social Justice Research, 1, 177–198. Andersen, S. M., Chen, S., & Miranda, R. (2002). Significant others and the self. Self and Identity, 1, 159–168. Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close relationships as including other in the self. Journal of Personality and Social Psychology, 60, 241–253. Baden, J. A. (1998). A new primer for the management of common resources and public goods. In J. A. Baden & D. S. Noonan (Eds.), Managing the common (2nd ed., pp. 51–62). Bloomington: Indiana University Press. Baron, R. S., & Kerr, N. L. (2003). Group process, group decision, group action. Buckingham: Open University Press. Barrett, K. C. (1995). A functionalist approach to shame and guilt. In J. P. Tangney & K. W. Fisher (Eds.), Self-conscious emotions: Shame, guilt, embarrassment, and pride (pp. 25–63). New York: Guilford. Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press. Batson, C. D., O’Quin, K., Fultz, J., Vanderplas, M., & Isen, A. (1983). Self-reported distress and empathy and egoistic versus altruistic motivation for helping. Journal of Personality and Social Psychology, 45, 706–718. Baumeister, R. F. (1982). A self-presentational view of social phenomena. Psychological Bulletin, 91, 3–26. Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5, 323–370. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529. Bies, R. J., & Moag, J. S. (1986). Interactional justice: Communication criteria for fairness. In B. Sheppard (Ed.), Research on negotiation in organizations (pp. 43–55). Greenwich, CT: JAI Press.

208

DAVID DE CREMER AND TOM R. TYLER

Bies, R. J., & Tripp, T. M. (2002). ‘‘Hot flashes, open wounds’’: Injustice and the tyranny of its emotions. In D. D. Steiner, D. P. Skarlicki, & S. W. Gilliland (Eds.), Emerging perspectives on managing organizational justice (pp. 203–221). Greenwich, Connecticut: Information Age Publishing. Blader, S. L., & Tyler, T. R. (2003a). A four component model of procedural justice. Personality and Social Psychology Bulletin, 29, 747–758. Blader, S. L., & Tyler, T. R. (2003b). What constitutes fairness in work settings?. Human Resource Management Review, 12, 107–126. Blau, G. (1993). Operationalizing direction and level of eVort and testing their relationships to individual and job performance. Organizational Behavior and Human Decision Processes, 55, 152–170. Brann, P., & Foddy, M. (1987). Trust and the consumption of a deteriorating common resource. Journal of Conflict Resolution, 31, 615–630. Brewer, M. B. (1979). In-group bias in the minimal group intergroup situation: A cognitivemotivational analysis. Psychological Bulletin, 86, 307–324. Brewer, M. B. (1981). Ethnocentrism and its role in interpersonal trust. In M. B. Brewer & B. E. Collins (Eds.), Scientific inquiry and the social sciences (pp. 345–360). New York: JosseyBass. Brewer, M. (1991). The social self: On being the same and diVerent at the same time. Personality and Social Psychology Bulletin, 17, 475–482. Brockner, J. (1984). Low self-esteem and behavioral plasticity: Some implications for personality and social psychology. In L. Wheeler (Ed.), Review of personality and social psychology (Vol. 4, pp. 237–271). Beverly Hills, CA: Sage. Brockner, J., De Cremer, D., Van den Bos, K., & Chen, Y. (in press). The influence of interdependent self-construal on procedural fairness eVects. Organizational Behavior and Human Decision Processes. Brockner, J., Siegel, P. A., Daly, J. P., Tyler, T. R., & Martin, C. (1997). When trust matters: The moderating eVect of outcome favorability. Administrative Science Quarterly, 42, 558–583. Brockner, J., & Wiesenfeld, B. M. (1996). An integrative framework for explaining reactions to decisions: Interactive eVects of outcomes and procedures. Psychological Bulletin, 120, 189–208. Burns, J. M. (1978). Leadership. New York: Harper & Row. Cameron, K. S., Dutton, J. E., & Quinn, R. E. (2003). Positive organizational scholarship. San Francisco: Berrett-Koehler Publishers. Campbell, J. D. (1990). Self-esteem and clarity of the self-concept. Journal of Personality and Social Psychology, 59, 538–549. Campbell, J. D., Trapnell, P. D., Heine, S. J., Katz, I. M., Lavallee, L. F., & Lehman, D. R. (1996). Self-concept clarity: Measurement, personality correlates, and cultural boundaries. Journal of Personality and Social Psychology, 70, 141–156. Caporeal, L. R., & Brewer, M. B. (1991). Reviving evolutionary psychology: Biology meets society. Journal of Social Issues, 47, 187–195. Cartwright, D., & Zander, A. (1968). Group dynamics. New York: Harper and Row. Carver, C. S. (2001). Self-regulation. In A. Tesser & N. Schwarz (Eds.), Blackwell handbook of social psychology: Intraindividual processes (pp. 307–328). Oxford: Blackwell. Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge: Cambridge University Press. Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic information processing within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 212–252). New York: Guilford Press.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

209

Colquitt, J. A., & Greenberg, J. (2001). Doing justice to organizational justice: Forming and applying fairness judgments. In S. Gilliland, D. Steiner, & D. Skarlicki (Eds.), Theoretical and cultural perspectives on organizational justice (pp. 217–242). Greenwich: Information Age Publishing. Crosby, F. (1976). A model of egoistic relative deprivation. Psychological Review, 83, 85–112. Cropanzano, R., & Schminke, M. (2001). Using social justice to build eVective work groups. In M. E. Turner (Ed.), Groups at work (pp. 143–172). Mahwah, NJ: Lawrence Erlbaum Associates. Cropanzano, R., Byrne, Z. S., Bobocel, D. R., & Rupp, D. E. (2001). Moral virtues, fairness heuristics, social entities and other denizens of organizational justice. Journal of Vocational Behavior, 58, 164–209. Dawes, R. M. (1980). Social dilemmas. Annual Review of Psychology, 31, 169–193. Deci, E. L., & Ryan, R. M. (2000). The what and why of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11, 227–268. De Cremer, D. (1999). Trust and fear of exploitation in a public goods dilemma. Current Psychology, 18, 153–163. De Cremer, D. (2002). Respect and cooperation in social dilemmas: The importance of feeling included. Personality and Social Psychology Bulletin, 28, 1335–1341. De Cremer, D., & Van Knippenberg, D. (2003). Cooperation with leaders in social dilemmas: On the eVects of procedural fairness and outcome favorability in structural cooperation. Organizational Behavior and Human Decision Processes, 91, 1–11. De Cremer, D. (2003a). Why inconsistent leadership is regarded as procedurally unfair: The importance of social self-esteem concerns. European Journal of Social Psychology, 33, 535–550. De Cremer, D. (2003b). Non-economic motives predicting cooperation in public good dilemmas: The eVect of received respect on contributions. Social Justice Research, 16, 367–377. De Cremer, D. (2004a). The influence of accuracy as a function of leader’s bias: The role of trustworthiness in the psychology of procedural justice. Personality and Social Psychology Bulletin, 30, 293–304. De Cremer, D. (2004b). Leader fairness theory: The interplay between procedural fairness, leader behavior and self. Invited talk at the Pre-Justice conference, Society for Personality and Social Psychology, Austin, TX, January 31. De Cremer, D. (2004c). EVects of pre-decision and post-decision voice as a function of leader’s confidence. Unpublished manuscript, Tilburg University, the Netherlands. De Cremer, D., & Alberts, H. (2004). When procedural fairness does not influence how good I feel: The eVects of voice and leader selection as a function of belongingness needs. European Journal of Social Psychology, 34, 333–344. De Cremer, D., & Blader, S. (2004). Why do people care about procedural fairness? The importance of belongingness in responding and attending to procedures. Manuscript submitted for publication. De Cremer, D., & Sedikides, C. (in press). Self-uncertainty and responsiveness to procedural justice. Journal of Experimental Social Psychology. De Cremer, D., & Sedikides, C. (2004). Where do I stand? Reputational implications of procedural justice. Manuscript submitted for publication. De Cremer, D., Snyder, M., & Dewitte, S. (2001). The less I trust, the less I contribute (or not?): EVects of trust, accountability and self-monitoring in social dilemmas. European Journal of Social Psychology, 31, 91–107. De Cremer, D., & Stouten, J. (2003). When do people find cooperation most justified? The eVect of trust and self-other merging in social dilemmas. Social Justice Research, 16, 41–52.

210

DAVID DE CREMER AND TOM R. TYLER

De Cremer, D., Tyler, T. R., & den Ouden, N. (in press). Managing cooperation via procedural fairness: The mediating role of self-other merging. Journal of Economic Psychology. De Cremer, D., & Tyler, T. R. (in press). A matter of intragroup status: The importance of respect for the viability of groups. In B. Mannix, M. Neale, & M. Thomas-Hunt (Eds.), Research on managing group and teams. Greenwich, CT: Elsevier Science Press. De Cremer, D., & Tyler, T. R. (2004). Am I respected or not? Inclusion and reputation as issues in group membership. Manuscript submitted for publication. De Cremer, D., & Van den Bos, K. (2003). Instrumental and non-instrumental fairness considerations: EVects of pre-decision and post-decision voice moderated by interdependent self-construal. Manuscript submitted for publication. De Cremer, D., & Van Dijk, E. (2002). Reactions to group success and failure as a function of identification level: A test of the goal-transformation hypothesis in social dilemmas. Journal of Experimental Social Psychology, 38, 435–442. De Cremer, D., & Van Hiel, A. (2004). Procedural justice eVects on cooperation as a function of individual diVerences: Neuroticism as an organizing dimension. Manuscript submitted for publication. De Cremer, D., Van Knippenberg, B., Van Knippenberg, D., Mullenders, D., Stinglhamber, F. (in press). EVects of procedural fairness as a function of leadership style: Self-reward leadership and fair procedures as determinants of self-esteem. Journal of Applied Psychology. De Cremer, D., & Van Knippenberg, D. (2002). How do leaders promote cooperation? The eVects of charisma and procedural fairness. Journal of Applied Psychology, 87, 858–866. De Cremer, D., & Van Knippenberg, D. (2003). Cooperation with leaders in social dilemmas: On the eVects of procedural fairness and outcome favorability in structural cooperation. Organizational Behavior and Human Decision Processes, 91, 1–11. De Cremer, D., Van Knippenberg, D., van Dijke, M., & Bos, A. (2004). How self-relevant is fair treatment? Social self-esteem moderates the eVect of interactional justice. Social Justice Research. De Cremer, D., & Van Lange, P. A. M. (2001). Why prosocials exhibit greater cooperation than proselfs: The role of responsibility and reciprocity. European Journal of Personality, 15, S1, 5–18. de Cremer, D., & Van Vugt, M. (1999). Social identification eVects in social dilemmas; A transformation of motives. European Journal of Social Psychology, 29, 871–893. De Cremer, D., & Van Vugt, M. (2002). Intergroup and intragroup aspects of leadership in social dilemmas: A relational model of cooperation. Journal of Experimental Social Psychology, 38, 126–136. Digman, J. M. (1990). Personality structure: Emergence of the five-factor model. Annual Review of Psychology, 41, 417–440. Doob, A. N., & Webster, C. M. (2003). Sentence severity and crime: Accepting the null hypothesis Crime and Justice (Vol. 30, pp. 143–196). Chicago: University of Chicago Press. Dykman, B. J. (1998). Integrating cognitive and motivational factors in depression: Initial tests of a goal-oriented approach. Journal of Personality and Social Psychology, 74, 139–158. Eagly, A. E., & Chaiken, S. (1993). The psychology of attitudes. Orlando, FL: Harcourt Brace & Company. Folger, R. (1977). Distributive and procedural justice: Combined impact of ‘‘voice’’ and improvement of experienced inequity. Journal of Personality and Social Psychology, 35, 108–119. Folger, R. (1986). Rethinking equity theory: A referent cognitions model. In H. W. BierhoV, R. L. Cohen, & J. Greenberg (Eds.), Justice in social relations (pp. 145–162). New York: Plenum Press.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

211

Folger, R., & Cropanzano, R. (1998). Organizational justice and human resource management. Thousand Oaks, CA: Sage. Ga¨chter, S., & Fehr, E. (1999). Collective action as a social exchange. Journal of Economic Behavior and Organization, 39, 341–369. Galinski, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of Personality and Social Psychology, 85, 453–466. Granovetter, M. (1992). Problems of explanation in economic society. In N. Nohria & R. G. Eccles (Eds.), Network and organizations (pp. 22–55). Boston: Harvard Business Press. Greenberg, J. (2001). The seven loose can(n)ons of organizational justice. In J. Greenberg & R. Cropanzano (Eds.), Advances in organizational justice (pp. 245–272). Stanford, CA: Stanford University Press. Green, D. P., & Shapiro, I. (1994). Pathologies of rational choice theory: A critique of applications in political science. New Haven, CT: Yale University Press. Gurr, T. R. (1970). Why men rebel. Princeton, NJ: Princeton University Press. Haidt, J. (2001). The emotional dog does learn new tricks: A reply to Pizarro and Bloom (2003). Psychological Review, 110, 197–198. Hardin, G. (1968). The tragedy of the commons. Science, 162, 1243–1248. Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. Advances in Experimental Social Psychology, 46, 1–46. Higgins, E. T., & Kruglanski, A. W. (2001). Motivational science: The nature and functions of wanting. In E. T. Higgins & A. W. Kruglanski (Eds.), Motivational science: Social and personality perspectives (pp. 1–20). New York: Psychology Press. Higgins, E. T., Shah, J. Y., & Friedman, R. (1997). Emotional responses to goal attainment: Strength of regulatory focus as moderator. Journal of Personality and Social Psychology, 72, 515–525. Hogg, M. A. (2003). Social identity. In M. R. Leary & J. P. Tangney (Eds.), Handbook of self and identity (pp. 462–479). New York: Guilford. Hogg, M. A., & Van Knippenberg, D. (2003). Social identity and leadership processes in groups. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 35). New York, USA: Elsevier Science. Huo, Y. J., Smith, H. J., Tyler, T. R., & Lind, E. A. (1996). Superordinate identification, subgroup identification, and justice concerns: Is separatism the problem; is assimilation the answer?. Psychological Science, 7, 40–45. James, W. (1890). The principles of psychology. New York, NY: Holt. Johnson, D. E., Erez, A., Kiker, D. S., & Motowidlo, S. J. (2002). Liking and attributions of motives as mediators of the relationships between individuals’ reputations, helpful behaviors, and raters’ reward decisions. Journal of Applied Psychology, 87, 808–815. Judge, T. A., Locke, E. A., & Durham, C. C. (1997). The dispositional causes of job satisfaction: A core evaluations approach. Research in Organizational Behavior, 19, 151–188. Kahrl, W. L. (1982). Water and power. Berkeley: University of California Press. Kelley, H. H., Holmes, J. G., Kerr, N. L., Reis, H. T., Rusbult, C. E., & van Lange, P. A. M. (2003). An atlas of interpersonal situations. Cambridge: Cambridge University Press. Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal relations: A theory of interdependence. New York: Wiley and Sons. Kernis, M. H., Whisenhunt, C. R., Waschull, S. B., Greenier, K. D., Berry, A. J., Herlocker, C. E., & Anderson, C. A. (1998). Multiple facets of self-esteem and their relations to depressive symptoms. Personality and Social Psychology Bulletin, 24, 657–668. Kerr, N. (1983). Motivation losses in small groups. A social dilemma analysis. Journal of Personality and Social Psychology, 45, 819–828.

212

DAVID DE CREMER AND TOM R. TYLER

Komorita, S. S., & Parks, C. D. (1994). Social dilemmas. Dubuque, IA: Brown & Benchmark. Kopelman, S., Weber, J. M., & Messick, D. M. (2002). Factors influencing cooperation in commons dilemmas: A review of experimental psychological research. In E. Ostrom, T. Dietz, N. Dolsak, P. C. Stern, S. Stonich, & E. U. Weber (Eds.), The drama of the commons (pp. 113–156). Washington, DC: National Academy Press. Korsgaard, M. A., Schweiger, D. M., & Sapienza, H. J. (1995). The role of procedural justice in building commitment attachment and trust in strategic decision-making teams. Academy of Management Journal, 38, 60–84. Kramer, R. M., & Brewer, M. B. (1984). EVects of group identity on resource use in a simulated commons dilemma. Journal of Personality and Social Psychology, 46, 1044–1057. Kramer, R. M., Brewer, M. B., & Hanna, B. A. (1996). Collective trust and collective action: The decision to trust as a social decision. In R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations: Frontiers of theory and research (pp. 357–389). Thousands Oaks, CA: Sage. Kramer, R. M., & Goldman, L. (1995). Helping the group or helping yourself? Social motives and group identity in resource dilemmas. In D. A. Schroeder (Ed.), Social dilemmas: Perspectives on individuals and groups (pp. 49–67). London: Praeger. Kramer, R. M., McClintock, C. G., & Messick, D. M. (1986). Social values and cooperative response to a stimulated resource conservation crisis. Journal of Personality, 54, 101–117. Kunda, Z. (1999). Social cognition. Cambridge, MA: MIT Press. Leakey, R. E. (1978). People of the lake: Mankind and its beginnings. New York: Avon. Leary, M. (2002). The self as a source of relational diYculties. Self and Identity, 1, 137–142. Leary, M. R., & Baumeister, R. F. (2000). The nature and function of self-esteem: Sociometer theory. In M. P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 32, pp. 1–62). San Diego, CA: Academic Press. Leary, M. R., Kelly, K. M., Cottrell, C. A., & Schreindorfer, L. S. (2001). Individual diVerences in the need to belong. Unpublished manuscript, Wake Forest University, Winston-Salem, NC. Leith, K. P., & Baumeister, R. F. (1998). Empathy, shame, guilt, and narratives of interpersonal conflicts: Guilt-prone people are better at perspective taking. Journal of Personality, 66, 1–37. Leventhal, G. S. (1980). What should be done with equity theory?. In K. J. Gergen, M. S. Greenberg, & R. H. Willis (Eds.), Social exchange: Advances in theory and research. New York: Plenum Press. Lewin, K. (1948). Resolving social conflicts: Selected papers on group dynamics. New York: Harper and Row. Lewis, M. (1992). Shame: The exposed self. New York: Free Press. Lewis, M. (1995). Embarrassment: The emotion of self-exposure and evaluation. In J. P. Tangney & K. W. Fisher (Eds.), Self-consious emotions: Shame, guilt, embarrassment, and pride (pp. 198–218). New York: Guilford. Liebrand, W. B. G., & Van Run, G. J. (1985). The eVects of social motives on behavior in social dilemmas in two cultures. Journal of Experimental Social Psychology, 21, 86–102. Lind, E. A. (2001a). Fairness heuristic theory: Justice judgments as pivotal cognitions in organizational relations. In J. Greenberg & R. Cropanzano (Eds.), Advances in organizational justice (pp. 56–88). Stanford, CA: Stanford University Press. Lind, E. A. (2001b). Thinking critical about justice judgments. Journal of Vocational Behavior, 58, 220–226. Lind, E. A., Kanfer, R., & Earley, P. C. (1990). Voice, control, and procedural justice: Instrumental and noninstrumental concerns in fairness judgments. Journal of Personality and Social Psychology, 59, 952–959.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

213

Lind, E. A., & Tyler, T. R. (1988). The social psychology of procedural justice. New York: Plenum Press. Lord, R. G., Brown, D. J., & Freiberg, S. J. (1999). Understanding the dynamics of leadership: The role of follower self-concepts in the leader/follower relationship. Organizational Behavior and Human Decision Processes, 78, 167–203. Markus, H., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and involvement. Psychological Review, 98, 224–253. Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20, 709–734. McClintock, C. G., & Liebrand, W. B. G. (1988). Role of interdependence structure, individual value orientation, and another’s strategy in social decision making: A transformational analysis. Journal of Personality and Social Psychology, 55, 396–409. McGregor, I., Zanna, M. P., Holmes, J. G., & Spencer, S. J. (2001). Compensatory conviction in the face of personal uncertainty: Going to extremes and being oneself. Journal of Personality and Social Psychology, 80, 472–488. Messick, D. M. (1995). Equality, fairness, and social conflict. Social Justice Research, 8, 153–174. Messick, D. M., Bloom, S., Boldizar, J. P., & Samuelson, C. D. (1985). Why we are fairer than others. Journal of Experimental Social Psychology, 21, 480–500. Messick, D. M., & Brewer, M. B. (1983). Solving social dilemmas. In L. Wheeler & P. R. Shaver (Eds.), Review of Personality and Social Psychology (Vol. 4, pp. 11–44). Beverly Hills, CA: Sage Publications. Messick, D. M., & McClintock, C. G. (1968). Motivational basis of choice in experimental games. Journal of Experimental Social Psychology, 4, 1–25. Messick, D. M., Wilke, H. A. M., Brewer, M. B., Kramer, R. M., Zemke, P. E., & Lui, L. (1983). Individual adaptations and structural change as solutions to social dilemmas. Journal of Personality and Social Psychology, 44, 294–309. Messick, D. M., & Sentis, K. P. (1985). Estimating social and non-social utility functions from ordinal data. European Journal of Social Psychology, 15, 389–399. Mikula, G., Petri, B., & Tanzer, N. K. (1990). What people regard as unjust: Types and structures of everyday experiences of injustice. European Journal of Social Psychology, 20, 133–149. Milinski, M., Semmann, D., & Krambeck, H.-J. (2002). Reputation helps solve the ‘‘tragedy of the commons.’’ Nature, 415, 424–426. Miller, D. T. (2001). Disrespect and the experience of injustice. Annual Review of Psychology, 52, 527–553. Miller, R. S. (1995). Embarrassment and social behavior. In J. P. Tangney & K. W. Fisher (Eds.), Self-conscious emotions: Shame, guilt, embarrassment, and pride (pp. 322–339). New York: Guilford. Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80, 252–283. Montada, L., & Schneider, A. (1989). Justice and emotional reactions to the disadvantaged. Social Justice Research, 3, 313–344. Mulder, L., van Dijk, E., De Cremer, D., & Wilke, H. A. M. (2003). Undermining trust and cooperation: The paradox of sanctioning systems in social dilemmas. Manuscript under review. Noel, J. G., Wann, B. L., & Branscombe, N. R. (1995). Peripheral ingroup membership status and public negativity toward outgroups. Journal of Personality and Social Psychology, 68, 127–137.

214

DAVID DE CREMER AND TOM R. TYLER

Oleson, K. C., Poehlmann, K. M., Yost, J. H., Lynch, M. E., & Arkin, R. M. (2000). Subjective overachievement: Individual diVerences in self-doubt and concern with performance. Journal of Personality, 68, 491–524. Ostrom, E. (1990). Governing the commons: The evolution of institutions for collective action. New York: Cambridge University Press. Parks, C. D., & Hulbert, L. G. (1995). High and low trusters’ responses to fear in a payoV matrix. Journal of Conflict Resolution, 39, 718–730. Pervin, L. A. (1982). The stasis and flow of behavior: Toward a theory of goals. In M. M. Page & R. Dienstbier (Eds.), Nebraska symposium on motivation, (Vol. 30, pp. 1–53). Lincoln: University of Nebraska Press. PfeVer, J. (1994). Competitive advantage through people. Cambridge, MA: Harvard. Pillai, R., Scandura, T. A., & Williams, E. A. (1999). Leadership and organizational justice: Similarities and diVerences across cultures. Journal of International Business Studies, 30, 763–779. Pittman, T. S. (1998). Motivation. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 549–590). New York: McGraw-Hill. Poundstone, W. (1992). Prisoner’s dilemma. New York: Doubleday. Posner, E. A. (2000). Law and social norms. Cambridge, MA: Harvard University Press. Pruitt, D. G., & Kimmel, M. (1977). Twenty years of experimental gaming: Critique, synthesis, and suggestions for the future. Annual Review of Psychology, 28, 363–392. Rawls, J. (1971). A theory of justice. Cambridge, MA: Harvard University Press. Reicher, S. D., & Hopkins, N. (2001). Self and nation. London: Sage. Ring, P. S., & van de Ven, A. H. (1994). Development processes of cooperative interorganizational relationships. Academy of Management Review, 19, 90–118. Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of Personality and Social Psychology, 37, 322–336. Rusbult, C. E., & van Lange, P. A. M. (1996). Interdependence processes. In G. Lindzey & E. Aronson (Eds.), Social psychology: Handbook of basic principles (pp. 564–596). Guilford Press. Rusbult, C. E., & van Lange, P. A. M. (2003). Interdependence, interaction, and relationships. Annual Review of Psychology, 54, 351–375. Ryan, A. (1993). Justice. Oxford: Oxford University Press. Ryan, R. M., & Deci, E. L. (2003). On assimilating identities to the self: A self-determination theory perspective on internalization and integrity within cultures. In M. R. Leary & J. P. Tangney (Eds.), Handbook of self and identity (pp. 253–274). New York: The Guilford Press. Schroeder, D. A., Steel, J. E., Woodell, A. J., & Bembenek, A. F. (2003). Justice within social dilemmas. Personality and Social Psychology Review, 7, 374–387. Sedikides, C. (2002). Self and identity: Social psychological perspectives. Revue Internationale De Psychologie Sociale, 15, (Issues 3–4 ). Sedikides, C., & Brewer, M. B. (2001). Individual self, relational self, collective self. Philadelphia, PA: Psychology Press. Sedikides, C., & Gregg, A. (2003). Portraits of the self. In M. A. Hogg & J. Cooper (Eds.), Sage handbook of social psychology. London: Sage Publications. Sedikides, C., & Strube, M. J. (1997). Self-evaluation: To thine own self be good, to thine own self be sure, to thine own self be true, and to thine own self be better. In M. P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 29, pp. 209–269). New York: Academic Press. Shah, J. Y., Brazy, P. C., & Higgins, E. T. (2004). Promoting us or preventing them: Regulatory focus and manifestations of intergroup bias. Personality and Social Psychology Bulletin, 30, 433–446.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

215

Sherman, S. J., Hamilton, D. L., & Lewis, A. C. (1999). Perceived entitativity and the social identity value of group memberships. In D. Abrams & M. A. Hogg (Eds.), Social identity and social cognition. Oxford: Blackwell. Singelis, T. M. (1994). The measurement of independent and interdependent self-construal. Personality and Social Psychology Bulletin, 20, 580–591. Skitka, L. J. (2003). Of diVerent minds: An accessible identity model of justice reasoning. Personality and Social Psychology Review, 7, 286–297. Snyder, M. (1993). Basic research and practical problems: The promise of a ‘‘functional’’ personality and social psychology. Personality and Social Psychology Bulletin, 19, 251–264. Synder, M., & Cantor, N. (1998). Understanding personality and social behavior: A functionalist strategy. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.), Handbook of social psychology (Vol. 1, pp. 635–679). New York: McGraw-Hill. Synder, M., & Ickes, W. (1985). Personality and social behavior. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (Vol. 2, pp. 883–948). New York: Random House. Snyder, C. R., & Lopez, S. J. (2002). Handbook of positive psychology. New York: Oxford University Press. Sunshine, J., & Tyler, T. R. (2003). The role of procedural justice and legitimacy in shaping public support for policing. Law and Society Review, 37, 555–589. Tangney, J. P. (1995). Shame and guilt in interpersonal relationships. In J. P. Tangney & K. W. Fisher (Eds.), Self-conscious emotions: Shame, guilt, embarrassment, and pride (pp. 114–139). New York: Guilford. Tangney, J. P. (1999). The self-conscious emotions: Shame, guilt, embarrassment, and pride. In T. Dalgleish & M. Power (Eds.), Handbook of cognition and emotion (pp. 541–568). West Sussex: John Wiley & Sons. Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel (Ed.), The psychology of intergroup relations. Chicago: Nelson Hall. Tesser, A., Stapel, D. A., & Wood, J. V. (2002). Self and motivation: Emerging psychological perspectives. Washington, DC: APA. Tesser, A., Wood, J. V., & Stapel, D. A. (2002). Introduction; An emphasis on motivation. In A. Tesser, D. A. Stapel, & J. V. Wood (Eds.), Self and motivation: Emerging psychological perspectives (pp. 3–11). Washington, DC: APA. Tenbrunsel, A. E., & Messick, D. M. (1999). Sanctioning systems, decision frames, and cooperation. Administrative Science Quarterly, 44(4), 684–707. Thibaut, J., & Kelley, H. H. (1959). The social psychology of groups. New York: John Wiley. Thibaut, J., & Walker, L. (1975). Procedural justice. Hillsdale, NJ: Erlbaum. Thompson, L., & Loewenstein, G. (1992). Egocentric interpretations of fairness and interpersonal conflict. Organizational Behavior and Human Decision Processes, 51, 176–197. Tice, D. M. (1992). Self-presentation and self-concept change: The looking glass self as magnifying glass. Journal of Personality and Social Psychology, 63, 435–451. Tice, D. M., & Wallace, H. M. (2003). The reflected self: Creating yourself as (you think) others see you. In M. R. Leary & J. P. Tangney (Eds.), Handbook of self and identity (pp. 91–105). New York: Guilford Press. Turner, J. C. (1987). A self-categorization theory. In J. C. Turner, M. A. Hogg, P. J. Oakes, S. D. Reicher, & M. Wetherell (Eds.), Rediscovering the social group: A self-categorization theory (pp. 42–67). Oxford: Basil Blackwell. Tyler, T. R. (1987). Conditions leading to value-expressive eVects in judgments of procedural justice: A test of four models. Journal of Personality and Social Psychology, 52, 333–344. Tyler, T. R. (1989). The psychology of procedural justice: A test of the group value model. Journal of Personality and Social Psychology, 57, 333–344.

216

DAVID DE CREMER AND TOM R. TYLER

Tyler, T. R. (1990). Why people obey the law. New Haven, CT: Yale. Tyler, T. R. (1994). Psychological models of the justice motive. Journal of Personality and Social Psychology, 67, 850–863. Tyler, T. R. (1997). The psychology of legitimacy. Personality and Social Psychology Review, 1, 323–344. Tyler, T. R. (1999). Why people cooperate with organizations: An identity-based perspective. Research in Organizational Behavior, 21, 201–246. Tyler, T. R. (2000). Social justice: Outcome and procedure. International Journal of Psychology, 35, 117–125. Tyler, T. R. (2001). Cooperation in organizations: A social identity perspective. In M. A. Hogg & D. J. Terry (Eds.), Social identity processes in organizational contexts (pp. 149–165). Philadelphia: Psychology Press. Tyler, T. R. (2003). Procedural justice, legitimacy, and the eVective rule of law. Crime and justice (Vol. 30, pp. 283–358). Chicago: University of Chicago Press. Tyler, T. R. (2004a). Process based leadership: How do leaders lead?. In D. Messick & R. Kramer (Eds.), The psychology of leadership: New Perspectives and Research. Hillsdale, NJ: Erlbaum. Tyler, T. R. (2004b). Justice, identity, and leadership. In D. van Knippenberg & M. Hogg (Eds.), Leadership and power: Identity processes in groups and organizations. London: Sage. Tyler, T. R., & Bies, R. J. (1990). Interpersonal aspects of procedural justice. In J. S. Carroll (Ed.), Applied social psychology in business settings. Hillsdale, NJ: Erlbaum. Tyler, T. R., & Blader, S. L. (2000). Cooperation in groups: Procedural justice, social identity, and behavioral engagement. Philadelphia: Taylor & Francis. Tyler, T. R., & Blader, S. L. (2002). Autonomous versus comparative status: Must we be better than others to feel good about ourselves. Organizational Behavior and Human Decision Processes, 89, 813–838. Tyler, T. R., & Blader, S. L. (2003a). Social identity and fairness judgments. In S. W. Guilliland, D. D. Steiner, & D. P. Skarlicki (Eds.), Emerging perspectives on values in organizations. Greenwich, CT: Information Age Publishing. Tyler, T. R., & Blader, S. L. (2003b). Procedural justice, social identity, and cooperative behavior. Personality and Social Psychology Review, 7, 349–361. Tyler, T. R., Boeckmann, R. J., Smith, H. J., & Huo, Y. J. (1997). Social justice in a diverse society. Boulder, CO: Westview Press. Tyler, T. R., & Caine, A. (1981). The role of distributional and procedural fairness in the endorsement of formal leaders. Journal of Personality and Social Psychology, 41, 642–655. Tyler, T. R., & Dawes, R. M. (1993). Fairness in groups: Comparing the self-interest and social identity perspectives. In B. A. Mellers & J. Baron (Eds.), Psychological perspectives on justice: Theory and applications (pp. 87–108). Cambridge, NY: Cambridge University Press. Tyler, T. R., & Degoey, P. (1995). Collective restraint in social dilemmas: Procedural justice and social identification eVects on support for authorities. Journal of Personality and Social Psychology, 69, 482–497. Tyler, T. R., Degoey, P., & Smith, H. J. (1996). Understanding why the justice of group procedures matters: A test of the psychological dynamics of the group-value model. Journal of Personality and Social Psychology, 70, 913–930. Tyler, T. R., & Huo, Y. J. (2002). Trust in the law. New York: Russell-Sage. Tyler, T. R., & Lind, E. A. (1992). A relational model of authority in groups. In M. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 25, pp. 115–191). New York: Academic Press.

PROCEDURAL JUSTICE, SELF, AND COOPERATION

217

Tyler, T. R., Rasinski, K., & Spodick, N. (1985). The influence of voice on satisfaction with leaders: Exploring the meaning of process control. Journal of Personality and Social Psychology, 48, 72–81. Tyler, T. R., & Smith, H. J. (1998). Social justice and social movements. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (Vol. 2, pp. 595–629). New York: McGraw-Hill. Tyler, T. R., & Smith, H. J. (1999). Justice, social identity, and group processes. In T. R. Tyler, R. M. Kramer, & O. P. John (Eds.), The psychology of the social self (pp. 223–264). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Van den Bos, K. (2001). Uncertainty management: The influence of uncertainty salience on reactions to perceived procedural fairness. Journal of Personality and Social Psychology, 80, 931–941. Van den Bos, K. (2003). On the subjective quality of social justice: The role of aVect as information in the psychology of justice judgments. Journal of Personality and Social Psychology, 85, 482–498. Van den Bos, K. (2004). Personal communication. Van den Bos, K., & Lind, E. A. (2002). Uncertainty management by means of fairness judgments. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 34, pp. 1–60). San Diego, CA: Academic Press. Van den Bos, K., Lind, E. A., & Wilke, H. A. M. (2001). The psychology of procedural and distributive justice viewed from the perspective of fairness heuristic theory. In R. Corpanzano (Ed.), Justice in the workplace: From theory to practice (pp. 49–66). Mahwah, NJ: Erlbaum. Van den Bos, K., Lind, E. A., Vermunt, R., & Wilke, H. A. M. (1997). How do I judge my outcome when I do not know the outcome of others? The psychology of the fair process eVect. Journal of Personality and Social Psychology, 72, 1034–1046. Van den Bos, K., Wilke, H. A. M., & Lind, E. A. (1998). When do we need procedural fairness? The role of trust in authority. Journal of Personality and Social Psychology, 75, 1449–1458. Vandello, J. A., & Cohen, D. (1999). Patterns of individualism and collectivism across the United States. Journal of Personality and Social Psychology, 77, 279–292. Van Knippenberg, D., & Hogg, M. A. (2003). A social identity model of leadership eVectiveness in organizations. Research in Organizational Behavior, 25, 243–296. Van Knippenberg, D., Van Knippenberg, B., De Cremer, D., & Hogg, M. A. (2004). Leadership, self, and identity: A review and research agenda. Leader Quarterly, 15, 825–856. Van Lange, P. A. M. (2000). Beyond self-interest: A set of propositions relevant to interpersonal orientations. In M. Hewstone & W. Stroebe (Eds.), European Review of Social Psychology (Vol. 11, pp. 297–331). London: Wiley. Van Lange, P. A. M., & Liebrand, W. B. G. (1991). Social value orientation and intelligence: A test of the goal prescribes rationality principle. European Journal of Social Psychology, 21, 273–292. Van Lange, P. A. M., Liebrand, W. B. G., Messick, D. M., & Wilke, H. A. M. (1992). Introduction and literature review. In W. B. G. Liebrand, D. M. Messick, & H. A. M. Wilke (Eds.), Social dilemmas: Theoretical issues and research findings (pp. 3–28). Oxford: Pergamon Press. Van Lange, P. A. M., Otten, W., De Bruin, E. N. M., & Joireman, J. A. (1997). Development of prosocial, individualistic, and competitive orientations: Theory and preliminary evidence. Journal of Personality and Social Psychology, 73, 733–746.

218

DAVID DE CREMER AND TOM R. TYLER

Van Prooijen, J. W., Van den Bos, K., & Wilke, H. A. M. (2002). Procedural justice and status: Status salience as antecedent of procedural fairness eVects. Journal of Personality and Social Psychology, 83, 1353–1361. Van Vugt, M., Snyder, M., Tyler, T. R., & Biel, A. (2000). Cooperation in modern society. London: Routledge. Vermunt, R., Van Knippenberg, D., Van Knippenberg, B., & Blaauw, E. (2001). Self-esteem and outcome fairness: DiVerential importance of procedural and outcome considerations. Journal of Applied Psychology, 86, 621–628. Walster, E., Walster, G. W., & Berscheid, E. (1978). Equity: Theory and research. Boston: Allyn and Bacon. Watson, D., Suls, J., & Haig, J. (2002). Global self-esteem in relation to structural models of personality and aVectivity. Journal of Personality and Social Psychology, 83, 185–197. Weiss, H. M., Suckow, K., & Cropanzano, R. (1999). EVects of justice conditions on discrete emotions. Journal of Applied Psychology, 84, 786–794. Young, P. T. (1961). Motivation and emotion: A survey of the determinants of human and animal activity. New York: Wiley. Yamagishi, T., & Sato, K. (1986). Motivational bases of the public goods problem. Journal of Personality and Social Psychology, 50, 67–73. Zajonc, R. B. (1998). As quoted in the APA monitor. (September, 1998). Washington, DC: APA. Zander, A. (1971). Motives and goals in groups. New York: Academic Press.

SO RIGHT IT’S WRONG: GROUPTHINK AND THE UBIQUITOUS NATURE OF POLARIZED GROUP DECISION MAKING Robert S. Baron

A review of the research and debate regarding Janis’s groupthink model leads to the conclusion that after some 30 years of investigation, the evidence has largely failed to support the formulation’s more ambitious and controversial predictions, specifically those linking certain antecedent conditions with groupthink phenomena. Moreover, research in the years since the theory’s inception indicates that most of the ‘‘groupthink’’ phenomena described by Janis occur in a far wider range of group settings than he originally envisioned. Collectively, these data strongly suggest that Janis erred when identifying the necessary and suYcient antecedent conditions for groupthink. A ubiquity model of groupthink is introduced that specifies a revised set of antecedent conditions to explain why groupthink-like behavior occurs in mundane, temporary, and even minimal groups and yet is not an invariant feature of group decision making. ‘‘I come to bury Caesar, not to praise him.’’ —Marc Anthony, in Julius Caesar (Shakespeare)

I. The History of a Fine Idea Janis’s model of groupthink is arguably the most widely publicized application of psychological principles to high-level military, political, and technical group decision making in the history of experimental psychology. This had to have pleased its author, who oVered this formulation as a compelling 219 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 37

Copyright 2005, Elsevier Inc. All rights reserved. 0065-2601/05 $35.00

220

ROBERT S. BARON

bridge between principles documented by laboratory research and ‘‘real-life’’ problems (e.g., Janis, 1971). As such the model ‘‘legitimized’’ the importance of decades of academic research on social influence and group process, much of which focused on perceptual and attitudinal judgments having little or no material consequence for participants.1 The model has been widely cited (cf. Fuller & Aldag, 1998), and is still described as a valid model in most texts within social psychology (e.g., Aronson, Wilson & Akert, 2003; Baron & Byrne, 2000; Forsyth, 1999; Lord, 1997) and in many introductory psychology texts as well (e.g., Myers, 2004). The longevity of this broad coverage both reflects and contributes to the common acceptance of groupthink as a valid and verified phenomenon not only by the lay public but by many academic psychologists as well (Fuller & Aldag, 1998; Turner & Pratkanis, 1998). As a result, it is disconcerting to find that there is substantial skepticism regarding this model among those involved scholars who have oVered detailed reviews of the groupthink and group decisionmaking literature (e.g., Aldag & Fuller, 1993; Esser, 1998; Hogg, 1992; Kerr & Tindale, 2004; Kramer, 1998; Longley & Pruitt, 1980; McCauley, 1989; Park, 1990; Paulus, 1998; Whyte, 1998). This chapter examines a possible explanation for this paradoxical state of aVairs, in which a model is widely accepted as valid despite the deep concerns of those most familiar with the research literature. One particular strength of the original groupthink model was that it oVered an array of testable assumptions regarding antecedent conditions, symptoms, deficient decision-making processes, and outcome variables. According to Janis, very strong group cohesion was the primary antecedent condition for groupthink, provided that it was complemented by several other group and situational antecedent conditions (see just below) (Janis, 1982; Janis & Mann, 1977; cf. Hogg & Hains, 1998). Janis felt that such intense cohesion was likely to be found in very high-echelon decision groups such as Kennedy’s cabinet or Nixon’s inner circle of advisors, where the perks of membership are at intoxicating levels and the credibility of fellow members is extremely high. The logic here is that in such group contexts, normative and informational social influence should both be very powerful forces. As noted, however, Janis assumed that strong group cohesion was likely to evoke groupthink only when supported by certain secondary antecedent conditions. These conditions referred to the nature of the group and the situation. In terms of group characteristics the secondary antecedent

1

For example, Baron et al. (1996) found only three studies in the conformity literature manipulating judgment importance some 60 years after Sherif’s (1935) initial report of manipulated social influence.

GROUPTHINK

221

conditions were the insulation of the group from outside influences, the lack of a tradition of impartial leadership (i.e., directive leadership), a lack of group norms favoring methodical search procedures, and homogeneity of member attitude or ideology. The secondary situational antecedent conditions included high stress from an external threat, group insulation from critics, and low situational member self-esteem either because of recent failure or the complexity of the current decision problem. Collectively these antecedent conditions were thought to provoke a tendency for concurrence seeking among members. This concurrence seeking was presumed to provoke (or manifest itself as) a series of groupthink symptoms. These symptoms fell into three clusters; overestimation of the ingroup (as strong, smart, invulnerable, morally superior), with corresponding negative stereotyping regarding the outgroup (as weak, immoral, vulnerable, stupid, and wrong), close-mindedness (e.g. rationalization of doubt), and pressures for uniformity (via mindguards, self censorship, illusion of unanimity) (Janis & Mann, 1977). These symptoms in turn were thought to lead to a number of defective decision-making processes. This list included: inadequate contingency plans for failure; inadequate information search; biased assessment of risks, costs, benefits, and moral implications (e.g., inadequate consideration of worst-case scenarios); incomplete consideration of the full range of decision options; and failure to reconsider the extent to which original/fundamental objectives were served by the advocated action. These flawed decision-making processes were then hypothesized to lead to the type of grossly inadequate, polarized (i.e., extreme) and premature group solutions that often, in retrospect, astound us by their hubris, absence of insight, and lack of concern with the consequences and likelihood of failure. The classic example here, of course, was the Kennedy group’s decision to support the Bay of Pigs invasion. A more recent example is the collective assumption of the Bush White House that the U.S. invasion and rebuilding of Iraq would be a relatively painless and rapid aVair, with U.S. troops greeted with cheers and flowers, and the costs of nation building minimized by sales of Iraqi oil and the cooperation of a grateful nation. Janis’s careful dissection of the group decision process, particularly his insightful specification of defective decision procedures and dysfunctional group reactions (symptoms), revitalized discussion and consideration of group decision-making as a theoretical topic. As such, it undoubtedly contributed to continued interest on topics such as group polarization and minority influence among others. Moreover, given the careful specification of antecedents, symptoms, and consequences within this model, it seemed amenable to a variety of empirical tests. Unfortunately, on closer inspection, recreating many of the antecedent conditions described by Janis (e.g., intense

222

ROBERT S. BARON

ingroup cohesion, meaningful threat, homogeneity of values, group insulation, etc.) proved diYcult under laboratory conditions. As a result, controlled lab studies are relatively scarce in this literature (cf. Mullen, Anthony, Salas, & Driskell, 1994), with the majority of reports taking the form of group decision case studies (e.g., Esser & Lindoerfer, 1989) or historical sampling studies (e.g., Herek, Janis, & Huth, 1987; Tetlock, Peterson, McGuire, Chang, & Feld, 1992). In short, as various writers have noted (e.g., Aldag & Fuller, 1993; Esser, 1998; Mullen et al., 1994) we have far fewer empirical tests of this formulation than one would expect given its widespread eVect. For example, Mullen et al. (1994) conducted an exhaustive literature search for their meta-analytic review of groupthink-cohesion research and uncovered only nine independent empirical studies examining Janis’s key prediction regarding the negative relation between group cohesion and decision quality. Likewise, Esser’s more general review (1998) identified 11 laboratory ‘‘groupthink’’ studies and some 17 historically based reports. This stands in stark contrast to the number of studies on other group-related topics that became popular at roughly the same time. For example, some 10 years ago a meta-analysis of the minority group influence literature identified 97 relevant studies (Wood, Lundgren, Ouellette, & Busceme, 1994), and a PsychINFO search by the current author identified 99 studies from 1974 until the present that listed group polarization or a related synonym (i.e., risky shift or choice shift) in the title. What is even more disconcerting is that the results reported in the groupthink literature are not particularly encouraging. Certain variables (e.g., homogeneity of background, group member insecurity, threat) have been largely ignored by the laboratory literature. Only one or two studies manipulate factors such as threat level (Turner, Pratkanis, Probasco, & Leve, 1992) or time pressure (Courtright, 1978; Neck & Moorhead, 1995), and Moorhead and Montanari (1986) report the only lab study that examines the eVect of in-group homogeneity on decision quality. In these particular cases, certain results are, in fact, congruent with the model. For example, Turner et al. found that threat of public scrutiny, coupled with high cohesion, did in fact lower decision quality. However, it seems premature to draw any confident conclusions regarding the eVect of threat or time pressure on the basis of these isolated studies. The variable of cohesion has drawn noticeably more research attention. The majority of the laboratory studies on this topic examine the key prediction that high group cohesion will impair group decision quality (assuming the other antecedent conditions are met). Sadly, the results regarding the eVects of group cohesion are inconsistent and often contradictory to prediction (see Esser, 1998; Fuller & Aldag, 1998; Hogg & Hains, 1998;

GROUPTHINK

223

McCauley, 1989; Mullen et al. 1994; Paulus, 1998).2 The results of historical sampling studies are also disappointing. For example, Tetlock et al. (1992) found that neither degree of crisis nor cohesion was reliably related to decision quality of major national policy decisions. The results remain just as disappointing when one considers alternative conceptions of cohesion. Hogg and Hains (1998) contrasted measures of cohesion based on group identification and ‘‘social attraction’’ (for the group as a whole) against a measure of cohesiveness based on individualized assessments of interpersonal attraction. In this laboratory study, groups role-played a group decision about closing a popular on-campus theater. Although cohesion based on group identification was positively correlated with five symptoms of groupthink at significant (or near significant) levels after covarying out other possible confounds, it was also negatively correlated with five other symptoms of groupthink at a significant level (23 symptoms were assessed in total; see Table I, p. 335). Very similar data patterns were observed on the social attraction and personal friendship measures of cohesion (see Table I, below). In short, regardless of measure, cohesion was positively related to only a few symptoms of groupthink as a rule and was negatively related to roughly the same or greater number of symptoms within the very same discussion groups. Given that Hogg and Hains did their best to create time pressure and directive leadership within these groups, this set of outcomes oVers little support for the ‘‘cohesion hypothesis.’’3 The historical case studies do suggest some support for the prediction that philosophical homogeneity and ingroup insulation will impair decision quality when one conducts comparisons across case history examples (McCauley, 1989). In addition, directive leadership has also been linked to poor decision-making in both laboratory groupthink research and historical reports (e.g. Esser, 1998; McCauley, 1989; but see Peterson, Owens, Tetlock, Fan, & Martorana, 1998). These results, however, do not represent

2 In fairness, Mullen et al. (1994) does find some marginally significant meta-analytic support for the prediction that cohesion impairs decision quality among those three experimental tests (of 17) in which certain conditions specified by Janis (1972) were experimentally activated (directive leadership, weak exploration of alternatives). Note however, that the reported eVect size was low r ¼ .176, as were the number of key tests. This perhaps explains the continued skepticism of various reviewers regarding this factor (e.g., Fuller & Aldag, 1998; Hogg & Hains, 1998; Paulus, 1998). 3 Note that these data also do not provide much support for Hogg and Hains’s contention (1998, p. 323) that cohesion based on either social attraction or group identification is more predictive of groupthink than cohesion based on friendship or interpersonal attraction despite the statements made in their abstract (p. 323).

224

ROBERT S. BARON

TABLE I Number and Direction of Significant and Marginally Significant Correlations Between Three Measures of Cohesion and 23 Symptoms of Groupthink in Hogg and Hains (1998) Measure

Number of positive r’s

Number of negative r’s

Group identification Mean rc Social attraction Mean r Personal friendship Mean r

4a[1]b .34 2 [1] .39 2 [6] .15

5 .28 4 [1] .31 2 [5] .21

Adapted from Table I, Hogg M.A. and Hains, S.C. (1998). Friendship and group identification: A new look at the role of cohesiveness in groupthink. European Journal of Social Psychology, 28, 323–341. Copyright # 1998 by John Wiley & Sons Limited. Adapted with permission. a Numbers without brackets are the number of correlations reported with p values of .05 or lower. Maximum possible number of correlations ¼ 23. b Numbers inside brackets are the number of marginally significant correlations (all reported with p values of .08 or lower, following the reporting convention adopted by Hogg & Haines, 1998). c Mean r computed over significant and marginally significant correlations.

particularly strong support for the groupthink model given that such factors would seem likely to produce defective decision-making even in the total absence of ‘‘groupthink reactions.’’ A directive leader who oVers a preferred solution and pushes for a rapid decision while also discouraging debate is also likely to increase the likelihood of a premature and incomplete group solution. In contrast, a leader who instead urges a group to follow a prescribed set of decision-making procedures that encourage debate and the free expression of ideas is likely to elevate decision quality and divergence of opinion (Peterson, 1997). Similar comments can be made about those studies, documenting the fact that variables such as ingroup insulation, or time pressure impair group decision quality (Courtright, 1978, Neck & Moorhead, 1995). In contrast, support for the more innovative predictions linking crisis, group member insecurity, and intense cohesion to groupthink outcomes are rare. On the basis of this disappointing lack of support for these key predictions, the great majority of reviews cited above recommend revisions, replacement, or even outright rejection of the model (see Table II). Other reviewers share the concerns we express above regarding the uncritical acceptance of the model (e.g., Fuller & Aldag, 1998). For example, Turner

GROUPTHINK

225

TABLE II Sample of Reviewers’ Comments Aldag and Fuller (1993): ‘‘The groupthink model has served a valuable role in generating interest in group problem-solving . . . however, the model has not incorporated two decades of research and has received limited empirical support and is restrictive in scope’’ (p. 549). Brown (2000): ‘‘It clearly is not the case, as Janis had surmised, that cohesion leads to poor decision making. Indeed, all the evidence suggests that it is unrelated to decision quality or may even be associated with better decision processes’’ (p. 219). Esser (1998): ‘‘The small number laboratory tests of groupthink theory conducted in the 25 years since Janis first presented the theory has not been suYcient to provide an evaluation of each of the antecedents of groupthink let alone an overall evaluation of the complete theory’’ (p. 133). Fuller and Aldag (1998): ‘‘In our view, groupthink is a compelling myth. Like other myths it tells of things that never were but always are. . . . How did we come to so widely and gladly accept it in the absence of compelling evidence?’’ (p. 177). Kramer (1998): ‘‘New evidence including recently declassified documents, rich oral histories, and informative memoirs by key participants in these decisions have become available for scholars, casting new light on the decision-making process behind both the Bay of Pigs and Vietnam. Much of this new evidence does not support Janis’s original characterization of these processes’’ (p. 236). McCauley (1989): ‘‘There has been surprisingly little research aimed at Janis’s hypotheses. . . . The results of manipulating cohesion are relatively weak and uncertain’’ (pp. 258–9). Paulus (1998): ‘‘There’s little evidence for the negative role of cohesion in group decisionmaking. There’s fairly consistent evidence for the role of directive leadership. For most of the other elements of the research literature, the evidence is rather limited. There certainly is not compelling support for the full model. . . . The impact of the antecedent conditions in the existence of various symptoms of groupthink is likely to depend upon a variety of task and contextual features’’ (p. 366). Raven (1998): ‘‘Even if most of us can identify some flaws in the groupthink analysis, we would still give Janis a lot of credit for his careful and scholarly analysis, his relating a broad body of literature and group processes and group dynamics to the understanding of the new series of very significant social political events’’ (p. 359). Turner and Pratkanis (1998): ‘‘The unconditional acceptance of the groupthink phenomenon without due regard to the body of scientific evidence surrounding it leads to unthinking conformity to a theoretical standpoint that may be invalid for the majority of circumstances’’ (p. 112).

and Pratkanis (1998) state, ‘‘The unconditional acceptance of the groupthink phenomenon without due regard to the body of scientific evidence surrounding it leads to unthinking conformity to a theoretical standpoint that may be invalid for the majority of circumstances’’ (p. 112). Actually, this skepticism regarding the groupthink model surfaced as early as 1980 when Longley and Pruitt (1980) raised several sensible criticisms

226

ROBERT S. BARON

of Janis’s groupthink analysis. Among these were the dangers posed by selective historical analysis, the possibility that the groupthink symptoms in Janis’s historical examples (particularly self censorship of dissent) might be more a result of group stage (early formation) than a function of Janis’s antecedent conditions (crisis, cohesion, directive leadership, etc.), and the argument that suppression of dissent might be functional in certain group settings. Added to these concerns was the conceptual ambiguity regarding antecedents that could be viewed as symptoms (e.g., directive leadership, cohesion, pressure for consensus) and symptoms that could be viewed as antecedents (e.g., outgroup stereotypes, illusion of consensus).4 However, Longley and Pruitt’s critical analysis did little to stem the excitement and attention directed at the groupthink model despite the lack of support reported from the earliest lab-simulation studies (e.g., Flowers, 1977; Fodor & Smith, 1982). Primary support for the groupthink model in these years stemmed from a number of sophisticated and ambitious historical sampling analyses (cf. Esser, 1998). Thus, for example, Tetlock (1979) conducted a content analysis of statements and speeches made by policy makers and verified that these statements were more simplistic and defensive (i.e., protective of the in-group) in cases that Janis had identified as instances of groupthink (see Tetlock et al., 1992 for a related study employing Q sort procedures). In another historical sampling study, Herek et al. (1987) had historical experts rate the quality of 19 randomly chosen U.S. policy decisions while other trained raters scanned records for procedural symptoms of groupthink. As predicted, the results indicated a negative relationship between symptoms and decision quality (see also Hensley & GriYn, 1986). The problem here is that there is not a lot of controversy regarding these specific predictions. It is not surprising that symptoms of groupthink (e.g., self censorship, rejection of criticism) or the defective decision processes these symptoms are thought to produce (e.g., poor information search, inadequate risk assessment) lead to low-quality decisions. Indeed, it would be remarkable if they did not. Rather, the excitement of the groupthink model lies in 4 There is also some degree of ambiguity regarding the extent to which certain symptoms of groupthink (e.g., illusions of consensus) cause or are caused by various characteristics of defective decision making (e.g., failure to fully evaluate the likelihood and costs of failure). A related point is that the model depicted by Janis and Mann (1977) implies that the desire for concurrence seeking mediates most of the symptoms. However, an alternative view is that concurrence seeking itself may just serve as one of several of these symptoms. Finally, there is ambiguity regarding whether the term groupthink refers only to concurrence seeking tendencies, the set of symptoms originally specified by Janis, or the entire process of antecedent conditions, symptoms, defective decision processes, and polarized, defective decision-making. This chapter uses the last interpretation.

GROUPTHINK

227

the prediction that the antecedent conditions (e.g., cohesion and crisis combined with directive leadership, insulation, a shared ideology, insecurity of members, etc.) would generally produce these symptoms and their consequences. Sadly, the historical studies, like their laboratory counterparts, did little to verify such relationships. As noted above, Tetlock et al. (1992) found that in the historical cases they examined, neither situational urgency (crisis) nor social cohesion, two of the major antecedent conditions, had much eVect on decision quality, whereas the historical analysis oVered by Herek et al. (1987) for some reason makes no mention of these antecedent conditions at all, leading the reader to suspect that in this study, findings regarding these antecedent variables were disappointing, null, or incoherent. In light of these results, the similar failure of the laboratory studies to verify the cohesion predictions is particularly troubling. Equally disconcerting is the fact that despite the chorus of criticism from reviewers, the groupthink model continues to be widely accepted as originally described, be it in textbooks, educational videos (e.g. Timmons, 1991), research articles, Web sites, or headlines (e.g. Fuller & Aldag, 1998). How might we explain the resilience of this model in the face of sparse, uneven, and contradictory findings? One answer is that, despite the existence of discordant data as far back as Flowers’s initial report in 1977, this model has a certain ‘‘ring of truth’’ that resonates with readers (McCauley, 1998). The symptoms and mechanisms described by the model seem familiar to us. They echo group processes we have experienced in our own social interactions. As a result, we are predisposed to accept the validity of such a formulation given only a modicum of supporting data. This tacit acceptance perhaps explains the rapid dissemination of the groupthink notion, at first within academic psychology and related decision sciences and soon beyond.

II. The Ubiquity of Groupthink My contention is that we are familiar with ‘‘groupthink’’ symptoms and processes because the concurrence-seeking, illusion of consensus, self censorship, and ingroup defensiveness described by Janis are far more widespread phenomena than he envisioned. After all, few of us have been to lunch at Hyannisport or Camp David, but most of us, I suspect, have been in settings in which our private reservations regarding some group option have been assuaged by a seeming consensus of our group mates or where our concerns about having pleasant social interactions and our own social acceptance take precedence over any need to explore every last objection and nuance to a collective decision.

228

ROBERT S. BARON

Put diVerently, the premise I oVer in this chapter is that Janis’s probing and insightful analysis of historical decision making was correct about the symptoms of groupthink and their relationship to such outcomes as the suppression of dissent, polarization of attitude, and poor decision quality and yet wrong about the antecedent conditions he specified. I contend that not only are these conditions not necessary to provoke the symptoms of groupthink but that they often will not even amplify such symptoms given the high likelihood that such symptoms will develop in the complete absence of intense cohesion, crisis, group insulation, and so on.

A. HAIL CAESAR In short, I argue that the frequent failure to verify the more ambitious of Janis’s predictions regarding the causal role played by the model’s antecedent conditions stem from the general prevalence of consensus seeking, group polarization, outgroup stereotyping, and the suppression of dissent in a wide array of group settings (see Levine & Thompson, 1996). As a result, such phenomena will often be at close to ceiling levels even in the absence of intense cohesion, crisis, homogeneous values, and so forth. Although such ceiling eVects may not be always present (see below), they may be common enough to blunt the eVect of Janis’s antecedent conditions, especially when relatively weak laboratory manipulations of threat and cohesion are employed. From this perspective, then, Janis was describing group processes that are likely to occur in ‘‘everyday,’’ mundane group decision settings as well as in ‘‘rarified’’ high-status groups. In fairness, Janis (1971) did acknowledge that on occasion, groupthink processes might occur in ‘‘everyday’’ groups, but the clear implication was that such events would be rare. My contention is that the implications of Janis’s model are far more sweeping than he envisioned.

B. CONFORMITY What evidence is there in defense of these assertions? A variety of findings now exist that support the view that many of the group symptoms and defective decision characteristics associated with groupthink are often found in ‘‘ordinary’’ groups. First, there is the fact that strong conformity eVects have been amply documented in laboratory experiments almost since the inception of modern experimental social psychology despite the fact that in almost all cases, the groups created in such studies operate in the absence of crisis, pronounced cohesion, philosophical homogeneity, and so on. As just

GROUPTHINK

229

one example, in Sherif’s (1935) classic report, social influence in the autokinetic paradigm was noted on 80% of recorded trials (Baron & Kerr, 2003). Although conformity is typically lower in the Asch paradigm (e.g., approximately 33% in Asch, 1956), several studies report stronger conformity eVects over trials using the Asch line–matching paradigm, provided that the judgment is characterized by a moderate degree of ambiguity. Thus for example, Deutsch and Gerard (1955), using the Asch paradigm, reported conformity on 57% of the 12 critical trials (i.e., mean conformity score ¼ 6.87) when participants had their group membership stressed and worked from memory (see Fig. 1). Indeed, even in nonambiguous conditions, Asch (1957) reported that 76% of his participants conformed on at least one critical trial. Of course conformity in such studies is not completely analogous to the social influence eVects characterizing groupthink. Simple conformity studies generally do not entail active discussion and provide no direct data regarding the private suppression of dissent or participants’ degree of belief in the group norm. In addition, the judgments in question (light movement, line matching, etc.) do not have any ‘‘real-life’’ consequences or importance.

Fig. 1. Number of conforming responses in anonymous conditions as a function of task diYculty and group salience. Adapted from Table I, Deutsch, M., and Gerard, H. (1955). A study of normative and informational social influence upon individual judgment. Journal of Abnormal and Social Psychology, 51, 629–636. In the public domain.

230

ROBERT S. BARON

Fig. 2. Conformity as a percentage of critical trials. Adapted from Fig. 2, Baron, R. S., Vandello, J., and Brunsman, B. (1996). The forgotten variable in conformity research: The impact of task importance on social influence. Journal of Personality and Social Psychology, 71, 915–927. Copyright # 1996 by the American Psychological Association. Adapted with permission.

However, a study by Baron, Vandello, and Brunsman (1996) does examine conformity on consequential judgments. Baron et al. (1996) modified the Asch paradigm. They manipulated judgment importance by oVering participants a $20 reward for superior performance on an eyewitness identification (face-matching) task. This manipulation increased conformity from the standard 33% of trials (Asch, 1956) to 55% of trials provided that the judgment was modestly ambiguous (see Fig. 2).5 In addition, in a second study, when confederates appeared confident and united, the conformity they evoked was correlated with participants’ feelings of confidence in this (incorrect) judgment as well. In short, exposure to a unified consensus provoked substantial and confident social influence on a consequential judgment despite the absence of many of the antecedent processes specified by 5 Baron et al. viewed these results as an instance in which motivated participants were more likely to use social cues to reach a decision because of their inability to employ a more systematic individualistic process under conditions of moderate task diYculty. Note that under low task diYculty, the $20 payment in this study should serve as an added inducement to resist group influence assuming that here, participants were relatively certain of their opinion. This is, in fact, what occurred in a low-ambiguity condition, where conformity occurred on only 16% of trials. This error rate, however, was still a significantly larger error score than obtained in the absence of confederates. Thus, even when participants were extremely confident of their initial judgment and oVered a substantial reward to be accurate, there was still evidence of concurrence seeking described by Janis as a hallmark of groupthink.

GROUPTHINK

231

the groupthink model (e.g., cohesion, homogeneity, crisis, etc.). Thus, these data provide good evidence that in the complete absence of groupthink antecedent conditions, individual (correct) opinions will often be verbally suppressed when a unified consensus voices an opposing point of view. Complementing these data, Wood, Pool, and their associates found that when individuals report identification with a membership group (Texas Aggies) they show several related reactions if they learn that they disagree with that group. First, they are more likely to change their opinion than control participants. Second, they are more likely to selectively interpret key words in attitude statements endorsed by the group so that they can ‘‘explain away’’ or minimize any potential disagreement between themselves and the group or alternatively justify changing their own position (Wood, Pool, Leck, & Purvis, 1996). Finally, if this ‘‘reinterpretation’’ option is not made salient, such participants are likely to show lower levels of self-esteem than individuals in comparison conditions (Pool, Wood, & Leck, 1998). Note these reactions are congruent with the notion that individuals are uneasy disagreeing with groups that they identify with and that such tendencies exist in the absence of directive leadership, group insulation, time pressure, a sense of crisis, or even direct contact with a highly attractive working group.

C. SUPPRESSION OF DISSENT Research as far back as Festinger and Thibaut’s (1951) classic study of written messaging within groups (re: a football and a delinquency problem) indicates that group members who express deviate opinions get initially pressured and then ignored and occasionally punished for failing to conform to salient group norms (e.g., Schachter, 1951; see Levine, 1989 for a review). Moreover, the recent literature on ostracism indicates that social rejection or even trivial exclusion within ad hoc groups or minimal groups is generally psychologically punishing for the targets of rejection (Williams & Sommer, 1997). For example, Williams, Cheung, and Choi (2000) had participants interact in ‘‘cyber-groups’’ where one participant was given the impression that they were being excluded during a game of Internet ‘‘ring toss.’’ This trivial form of rejection in unseen (minimal) Internet groups lowered mood and self-esteem while increasing the tendency of individuals to agree with the judgments of others. Thus, the research provides evidence of the social censorship described by Janis as well as ample documentation of the power of social rejection to deplete self-esteem and elevate conformity in the complete absence of the antecedent conditions outlined by the groupthink model.

232

ROBERT S. BARON

D. GROUP DISCUSSION AND DECISION POLARIZATION One caveat here is that the studies discussed above do not entail the active form of group discussion that Janis was referring to in his theoretical statements. However, there are now several lines of research on group discussion indicating that concurrence-seeking and intensification of attitude do occur reliably in groups despite the absence of the antecedent conditions specified by Janis. The literature on group polarization documents that discussion within like-minded groups reliably results in an intensification of attitude and judgment on a wide array of issues and decisions (see Baron & Kerr, 2003; Myers & Lamm, 1976 for reviews). The crucial antecedent condition for group polarization to occur is the presence of a like-minded group; that is, individuals who share a preference for one side of the issue. For example, Myers and Bishop (1970) found that groups of racial liberals became more liberal on race-related issues following discussion, whereas groups of racial conservatives polarized in the opposite direction. However, cohesion, crisis, threat, directive leadership, time pressure, and so on (i.e., Janis’s antecedent conditions) are rarely present in this research and, when manipulated, have not been found to heighten group polarization (Dion, Miller, & Magnan, 1971). Moreover, the explanations that have been oVered to account for such polarization eVects have referred to such processes as competitive social comparison (i.e., normative social influence) (Goethals & Zanna, 1979; Sanders & Baron, 1977), a biased flow of information and arguments (i.e., informational social influence) (Burnstein & Vinokur, 1977), social corroboration (Baron, Hoppe, Linneweh, & Rogers, 1996), and social identity concerns (Hogg, Turner, & Davidson, 1990). These processes are all closely related to explanatory mechanisms alluded to by Janis in his general discussion of groupthink. As noted above, however, Janis’s array of antecedent conditions, with the notable exception of group homogeneity, are not specified as necessary to provoke group polarization.

E. GROUP DISCUSSION AND SELF CENSORSHIP Stasser and his associates’ research on hidden profile eVects (e.g., Stasser, Vaughn, & Stewart, 2000) is particularly relevant to groupthink phenomena given the focus within this work on concurrence seeking, judgment polarization, and heightened confidence. Stasser’s research demonstrates that when members of a discussion group all share a number of positive bits of information regarding a decision alternative while simultaneously each holding an unshared (or unique) reservation regarding that alternative, a number

GROUPTHINK

233

of eVects emerge. Group members begin discussion favoring the decision option supported by the shared information. As discussion unfolds, the shared information gets disproportionately mentioned and reconsidered, whereas the unshared information tends to be ignored or given short shrift in the discussion. This results in the group becoming more positively (and confidently) disposed to the initially favored decision alternative following discussion. This occurs despite the fact that the full set of facts clearly favors the other (nonchosen) decision alternative. This hidden profile paradigm seems to provide a precise demonstration of the concurrence-seeking tendency specified by Janis. Individual, uniquely held reservations are ignored, repressed or deemphasized during discussion, whereas the initially favored decision becomes more polarized during discussion. Several processes appear to contribute to this hidden profile eVect. First, given that more people have access to the shared information, such information has a numerical advantage in terms of the probability that it gets mentioned by someone. Second, it appears that normative concerns are important as well. Thus, even when unshared information does get mentioned during discussion, it is less likely to provoke continued debate or to be repeated or reconsidered during discussion. Moreover, it appears that final attitudes and judgments of group members are more a function of the distribution of prediscussion individual member preferences than of the informational bits that surface in the group discussion (Gigone & Hastie, 1993). Gigone and Hastie conclude that in this research setting, group members appear to initially engage in some ‘‘averaging rule’’ to establish a consensual group position, which in turn aVects discussion content as well as the final group position (Gigone & Hastie, 1993). Indeed Gigone and Hastie suggest that group discussion generally serves to justify this consensual group position rather than to establish it (see p. 973). One likely explanation for such a ‘‘controlled discussion’’ focusing disproportionately on shared information is that group members have greater concerns for establishing and maintaining harmonious relations with fellow group members than they do in fully exploring the various facets of the decision problem. In accord with this view, Wittenbaum, Hubbell, and Zuckerman (1999) find that individuals prefer to both oVer and receive shared versus unshared information during discussion, and judge others to be more competent, knowledgeable, and credible when they endorse shared perspectives. A second reason group discussion may disproportionately favor shared information is that if there is pressure for rapid closure because of time urgency, a strong initial consensus driven by shared information can provide the justification for an attenuated and biased discussion (Kerr & Tindale, 2004) It is noteworthy that the hidden profile eVect is often a dramatic one. Thus, Stasser, and Titus (1985) reported that if participants’ were given the

234

ROBERT S. BARON

impression that they (as a group) might not have all the required information needed to solve a mystery, only 35% of the groups successfully unearthed the key (unshared) clues despite the fact that 100% of all groups had the complete set of clues. Indeed, even when groups were flatly told that they did have enough collective information to solve the mystery, only 67% managed to uncover the key bits of unshared information needed for solution. Similarly, Stasser, Taylor, and Hanna (1989) reported that, on average, discussion groups mentioned only 18% of the unshared information known to individual group members. In contrast, on average, some 46% of the shared information was discussed (i.e., at least a two-to-one ratio). Moreover, this tendency for the group to disproportionately discuss and consider the material that they all agree on and share is not easy to eradicate. Providing participants with explicit instructions to explore fully all decision alternatives and to try to uncover as much information as possible does not moderate this tendency (Stasser et al., 1989). Nor is the ‘‘hidden profile eVect’’ reduced by telling participants that they do not yet have all the information they will receive (Stasser, Stewart, & Wittenbaum, 1995) or that they will be held publicly accountable for the quality of their decisions (Stewart, Billings, & Stasser, 1998). Indeed, even privately informing participants that group member X has proportionally more information than the other members does not always lessen this eVect (Stasser et al., 2000 but see Stasser et al., 1995). In fairness, several manipulations have proven eVective at weakening the hidden-profile eVect (see Kerr & Tindale, 2004 for a review). If participants are given the clear expectation that their decision problem (e.g., a murder mystery) has a verifiable correct answer, they are less prone to the bias favoring shared information (Stasser & Stewart, 1992). This manipulation is likely to have elevated participants’ confidence about their collective ability to solve the problem. If so, these data indicate that low participant confidence contributes to the concurrence seeking that tends to occur during group discussion. Some other means of weakening the hidden profile eVect involve extending the time available for discussion (Larson, Foster-Fishman, & Keys, 1994) and giving one group member access to both shared and unshared information (Stewart & Stasser, 1998). Interestingly, Brodbeck, Kerschreiter, Mojzisch, Frey, and Schulz-Hardt (2002) found that assigning one group member the job of advocating the option supported by the unshared information (Janis’s ‘‘devil’s advocate strategy’’) also increases discussion of unshared information. However, although there are some means of moderating the hidden profile eVect, it is a pervasive and wellreplicated phenomenon. Again, few of the antecedents specified by the Janis model, with the exception of homogeneity of group members’ initial attitudes (created by the initial distribution of shared/unshared information), appear necessary to provoke the hidden profile bias. Groups in this research

235

GROUPTHINK

are not likely to be highly cohesive given their temporary and ad hoc nature; threat is not present given the hypothetical or trivial nature of the decision problems, and directive leadership is not encouraged. McLeod, Baron, Marti, and Yoon (1997) report a hidden profile study that also supports the view that group cohesion in particular is not a necessary condition for substantial suppression of dissent to occur during group discussion. In this study, business students discussed a business decision problem (which of three firms to acquire) either in face-to-face ad hoc lab groups or over a computer network. As in Stewart and Stasser (1998), one group member was fully informed whereas other participants shared partial information that favored a suboptimal choice. McLeod et al. (1997) found that the fully informed ‘‘expert’’ participants were likely to suppress their unshared information during discussion even in computer communication conditions in which they never saw each other face to face (and where, presumably, cohesion was minimal). In this study, the suppression eVect was substantially and significantly lessened only in one condition in which participants were allowed to participate in computer groups anonymously. Under these conditions the informed expert not only mentioned more of their ‘‘unshared’’ facts than in other conditions but also was more likely to remention the facts during discussion as well (see Table III). This liberating eVect of anonymity suggests, of course, that normative pressure (a presumed mediator of groupthink eVects) was at least partially responsible for the suppression of unshared information in the remaining treatments. Thus, we have evidence of self censorship and selective group attention (i.e., groupthink characteristics) occurring despite the general absence of most of the antecedent conditions specified by the groupthink model.

TABLE III Expression of Minority Views by Condition in McLeod et al., 1996

Face to face Nonanonymous computer groups Anonymous computer groups

Number of minority facts expressed

Number of times minority facts repeated

3.97 3.04 6.10

6.74 3.95 10.53

Adapted from Table II, McLeod, P., Baron, R. S., Marti, M. W., and Yoon, K. (1997). The eyes have it: Minority influence in face to face and computer mediated group discussion. Journal of Applied Psychology, 82, 706–718. Copyright # 1997 by the American Psychological Association. Adapted with permission.

236

ROBERT S. BARON

F. PLURALISTIC IGNORANCE AND THE ILLUSION OF CONSENSUS The studies reviewed above focus primarily on opinion polarization, concurrence seeking, the suppression of dissent, and selective group attention, all key characteristics of groupthink. Other work provides data regarding yet another characteristic mentioned by Janis (1972)—the illusion of consensus. The research on pluralistic ignorance is relevant here. This research indicates that individuals often publicly endorse decisions and attitude positions that they view as normative for their membership group despite having private reservations regarding such views or holding less extreme positions than those endorsed by the group. Moreover, in such settings, the individuals involved assume that similar (extreme) endorsement from other group members reflects their true feelings. Stated diVerently, pluralistic ignorance describes a situation in which each ‘‘member of a group or society privately rejects a belief, opinion, or practice, yet believes that virtually every other member privately accepts it’’ (Prentice & Miller, 1996, p. 162; see also Allport, 1924). As a result, each individual assumes that the private group consensus is more united and extreme than it actually is (see Miller & McFarland, 1991 for a review). For example, Korte (1972) completed a series of studies indicating that students felt that the dominant political climate on campus was decidedly more radical than it actually was (and more radical than their own). Similarly, Prentice and Miller (1993) reported that college students assumed that the other students on their campus and in their friendship network privately held far more tolerant attitudes regarding alcohol abuse than their own (see also Suls & Green, 2003). In addition, in public statements these students conformed to this illusion of consensus. Indeed, male students slowly relinquished their private reservations regarding excessive drinking over time. Prentice and Miller interpret these instances of public compliance and eventual agreement in terms of the participants’ desires to be accepted by groups with which they are aYliated. In a second report, Miller and Nelson (2002) found evidence of pluralistic ignorance in voting decisions made in the 2000 presidential election. In this case the typical voter cast a ballot using a ‘‘lesser of two evils’’ strategy while believing that others who endorsed the same candidate did so out of sincere attraction for the candidate. Several follow-up studies in this report replicated this eVect with respect to more prosaic choices regarding candies and soft drinks. In short, this research indicates that when group norms are salient, the public behavior of fellow members is attributed to an internal cause—their attitudinal endorsement of this norm. Stated diVerently, although one might acknowledge their own private reservations regarding a group norm, the group as a whole is viewed as privately agreeing with this

GROUPTHINK

237

point of view. These results are consistent with the general theme that we have been emphasizing above; that is, we find characteristics of groupthink (here the illusion of consensus) aVecting the decisions and attitudes of individuals despite the general absence of the key antecedent conditions specified by the groupthink model.

G. OUTGROUP VILIFICATION IN ORDINARY AND MINIMAL GROUPS An additional symptom of groupthink entails ingroup/outgroup stereotyping that disparages the enemy as weak or unworthy and extols the invulnerability and moral virtues of the ingroup. Research on both prejudice and social identity theory have documented a good number of such eVects in both ordinary and even minimal groups, and these have been noted by numerous commentators over the years (e.g., Hogg & Abrams, 2001; Murphy, 1953; Tajfel & Turner, 1979). Thus, Sherif and his team commented on such tendencies when observing the judgments and evaluations of preadolescent Rattlers and Eagles in the Robbers Cave Study (Sherif, Harvey, White, Hood, & Sherif, 1961). Similarly, Long and Spears (1998) reported ingroup/outgroup biases when members of brainstorming groups evaluated ingroup/outgroup solutions. In the same vein, Wang and McKillip (1978) reported that both Asian and American respondents made excessive judgments regarding the responsibility of outgroup members for traYc accidents (‘‘they’’ are bad, and irresponsible drivers/pedestrians). As an added example, Taylor and Jaggi (1974) found distinct self-serving stereotypes between Hindus and Moslems in India and then documented that when Hindu participants read about vignettes describing admirable or reprehensible behavior, Moslem failings were consistently attributed to internal causes whereas Hindu failings were conveniently attributed to external causes (see also Duncan, 1976). The reverse pattern occurred when positive behaviors were described. These data were cited by Pettigrew (1978) as classic examples of the ultimate attributional error but, more to the point, are congruent with the notion that various membership and ethnic groups frequently generate disparaging stereotypes and conclusions regarding outgroups while generating positive stereotypes and defensive rationalizations for the actions of the ingroup. Indeed, the existence of invidious stereotypes, at least between ethnic groups, is so pervasive that the process of activating them is thought by many to be automatic in its nature (Bargh, 1990; Devine, 1989). Moreover, such diVerential evaluations and interpretations are even reported between minimal groups (e.g., Brewer, 1979; Brown, Tajfel, & Turner, 1980;

238

ROBERT S. BARON

Sachdev & Bourhis, 1987). Thus for example, Sachdev and Bourhis (1987) assigned Canadian students to groups based presumably on the manner in which they completed a ‘‘creativity test’’ and found that ingroup members disparaged the creativity of outgroup members’ problem-solving solutions. Thus, again we have group eVects that ‘‘mimic’’ standard groupthink characteristics in settings that, for the most part, lack the antecedent conditions specified by Janis (although we would agree that cases of ethnic/racial stereotyping are most likely to involve a good deal of social identification and at least a moderate degree of cohesion). These results are consistent with our contention that groupthink is a far more pervasive phenomenon than Janis’s model presumes.

III. Rethinking Groupthink: The Ubiquity Model The one diYculty with the analysis oVered above is that if it is true, one must wonder why groups ever reach rational informed decisions and why so many group discussions are marked by acrimony, divisiveness, vituperative debate, turf battles, and so forth. Where is the concurrence seeking and suppression of dissent in these situations (where we can only pray for it)? And how can we contend that ‘‘groupthink [like Chickenman] is everywhere’’ given the incontrovertible evidence of such divisive group behavior in a range of public settings. It is apparent that there must be some limiting (or ‘‘antecedent’’) conditions aVecting the symptoms (and consequent defective decision processes) that characterize groupthink. The challenge is to reconsider what they must be in light of the lack of support regarding such variables as crisis, cohesion, insulation from critics, and so on. To this end I oVer a ubiquity model of groupthink, suggesting that three key conditions may serve as antecedents. The first is that the individuals in question must feel a sense of social identification with the collection of individuals they are among. This of course requires the preliminary perception among members that this collection of individuals comprises a group. This perception of entitativity (Campbell, 1958) will depend on the extent to which the individuals in question are linked by some common purpose, history, or shared fate (cf. Campbell, 1958, Hamilton & Sherman, 1996). Deciding that one is part of a ‘‘group’’ is assumed to generally provoke feelings of allegiance and social identity (cf. Tajfel & Turner, 1979), which we feel is a key antecedent condition for groupthink-like phenomena. This stipulation is based on the assumption that normative and informational social influence mediate groupthink phenomena and that both processes are dramatically limited unless there is some ‘‘minimal’’ degree of social

GROUPTHINK

239

identification (David & Turner, 1996, Hogg et al., 1990), even if this identification derives only from sharing a salient social designation or working collectively on some common if transient problem (say in a laboratory study).

A. SOCIAL IDENTIFICATION In accord with the view that informational social influence requires a sense of social identification, a growing list of studies indicate that instances of indirect6 minority influence are limited to entreaties and messages from ingroup members and that such ‘‘ingroup messages’’ provoke closer scrutiny and elaboration than those attributed to outgroup members (e.g., Alvaro & Crano, 1996, 1997; David & Turner, 1996). Given that cases of indirect influence are not likely to be caused by compliance processes (i.e., normative social influence), such data support the view that informational influence has a more pronounced eVect (i.e., is trusted and attended to more; Alvaro & Crano, 1996) when it stems from an ingroup source. Indeed, Deutsch and Gerard (1955) in their classic replication of Asch (1957) reported that even when participants were responding anonymously (and normative social influence should be minimal), conformity rates were approximately doubled in those conditions in which group identity had been emphasized to participants (see Fig. 1). The group identity manipulation in this case involved recurrently mentioning the ‘‘group nature’’ of the research and oVering a reward for group accuracy. Similarly, several studies (e.g., Mackie, 1986; Mackie & Cooper, 1984; Mackie, Gastardo-Conaco, & Skelly, 1992; McGarty, Haslam, Hutchinson, & Turner, 1994) find that group polarization eVects are limited to cases in which individuals are exposed to the taped arguments of ingroup members. Moreover, in these ingroup conditions, participants are more aVected by variations in message quality (Mackie et al., 1992; McGarty et al., 1994); a sign of greater message elaboration. These results are consistent with the proposed importance of social identification as a moderating variable in informational social influence. Heightened ingroup informational social influence appears to be mediated by several processes. First, as just noted, several studies find that ingroup messages elicit more attention and elaboration. This is likely because of 6 Indirect social influence refers to persuasion assessed on delayed measures or on those that assess persuasion on topics that are related to but separate from the focal topic of social influence. Thus, indirect social influence describes someone who after being urged to oppose sex education shows opposition to free distribution of condoms to sexually active students.

240

ROBERT S. BARON

the audience’s presumption that ingroup members share their vested interests, values, limitations, and frames of reference. If so, the views of these ingroup members should be more crucial for purposes of social comparison. Second, for this same reason, ingroup input may also be viewed as more trustworthy. Social identification is also assumed to significantly amplify normative social influence (Deutsch & Gerard, 1955). Here one can adopt a social identity perspective as one avenue of explanation, arguing that self-definition and self-esteem are strongly aVected by one’s social allegiances. This social identity view provides a ready explanation for why social deviance might be punishing for the deviant individual. Such deviance may threaten one’s self categorization as an ingroup member, thereby heightening uncertainty while also exposing the group member to expulsion from an admired group. As noted above, Wood et al. (1996) and Pool et al. (1998) provide good evidence that when people identify with a group, they will use cognitive distortion and semantic reinterpretation in an eVort to minimize perceptions of social deviance and will show drops in self-esteem when such cognitive avoidance is diYcult. These data are quite congruent with the notion that social deviance is aversive for individuals and has negative implications for self conceptions. Prentice and Miller (1996) oVer a related view when discussing pluralistic influence phenomena. They suggest that such eVects are caused at least in part by group identification—‘‘that individuals often act out of a desire to be good group members.’’ A second explanation for why social identification may moderate the eVects of normative social influence is based on conditioning principles. From our earliest moments, social acceptance and rejection from ingroup members are associated with a wide array of rewards and punishments be it food, freedom from discomfort, pleasing tactile stimulation, and so on. This contiguous pairing occurs with a wide range of exemplars (i.e., various forms of social acceptance/rejection), a multiplicity of primary reinforcers and punishments, in a variety of situations, and across a wide range of intimate individuals. This is the just the type of associative pattern that Skinner (1956) outlines when describing the development of generalized reinforcers and punishers; that is, conditioned stimuli paired with a varied array of primary reinforcers or punishers (respectively). According to Skinner, these conditioned stimuli come to be particularly potent sources of reinforcement and punishment, with the unique feature of being exceptionally resistant to extinction. Money is a common Skinnerean example of such a generalized reinforcer. Similarly, social rejection is a classic example of a generalized punishment. Stated diVerently, aversive reactions to social rejection can be viewed as an extremely enduring classically conditioned response that occurs

GROUPTHINK

241

reflexively7 even in situations in which actual, primary punishment is unlikely, inconsequential, or even impossible. In this view, deviation from ingroup standards should serve as a discriminatory stimulus that signals the possibility of such rejection. Moreover, given the reflexive nature of classically conditioned responses, the mere thought of deviating from ingroup members is likely to evoke this form of social anxiety. This conditioning view has the advantage of explaining the power of normative social influence in situations involving temporary and minimal groups. The social identity view becomes particularly plausible when considering more meaningful groups such as reference groups and groups with some shared history (e.g., Wood et al., 1996), because deviation here can aVect self image and self-esteem. That is, deviation undermines entitativity and the social identity derived from it. In contrast, the conditioning view outlined above seems particularly applicable to minimal groups and ad hoc laboratory decision groups where even low levels of entitativity should be suYcient to trigger these conditioned concerns regarding one’s deviation and potential social rejection even though the implications of such deviations for self image and self esteem are weak. Marques and Abram’s recent work on the ‘‘blacksheep eVect’’ (e.g., Abrams, Marques, Bown, & Dougill, 2002; Marques, Abrams, & Serodio, 2001) supports the view that individuals take comfort from ingroup unanimity (as suggested just above) and are likely to punish those who disrupt or prevent it. In a series of studies these investigators document that individuals who deviate from group norms are particularly likely be derogated if they are ingroup members as opposed to outgroup members (Abrams, Marques, Bown, & Henson, 2000). Moreover, this derogation is more pronounced among those individuals who identify most with the group (Abrams et al., 2002) and when individuals have doubts about the superiority of their ingroup (Marques et al., 2001). This work complements the general line of reasoning discussed just above by indicating that when ingroup deviance does occur, it is likely to trigger distinct social sanctions, particularly in cases in which social identity is highly salient or problematic. Although we feel a sense of group entitativity and social identification are crucial antecedent conditions for groupthink eVects, it is important to note that such feelings can be superceded by subgroup or coalition formation in which case the subgroup is far more likely to be the most accessible social 7 Although our discussion is primarily concerned with how generalized reinforcers and punishers contribute to the operant conditioning of conforming responses, the creation of such reinforcers (and the source of their ‘‘power’’) is a function of classical conditioning (i.e., pairing an initially neutral event with a variety of positive or negative unconditioned stimuli [UCSs]). As such, the aversive reaction to potential social rejection is thought to be evoked in a reflexive or ‘‘automatic’’ fashion, as is the case with any conditioned response.

242

ROBERT S. BARON

category. The members of the U.S. Senate certainly have a basis for selfidentification as ‘‘members of the Senate’’ given the club-like climate within this group, but this is often less salient to these individuals than their party aYliations or their social identity as Liberals, Conservatives, Presidential Loyalists, Southerners, Friends of Labor, or prolife/prochoice advocates. If so, any normative and informational social influence aVecting such individuals will be subgroup specific and likely to generate intergroup debate and acrimony. Under these circumstances any symptoms of groupthink that emerge will also be subgroup specific. Thus, the concurrence seeking, illusions of consensus, and defensive rationalization that occur will not necessarily result in absence of debate between opposing subgroups but rather should fuel such debate as subgroups become more polarized and more confident of their own subgroup position. This, then, explains why group discussions will often be contentious and divisive in contrast to the prototypic groupthink pattern.

B. SALIENT NORMS The second antecedent condition I suggest is that group interaction and discussion must produce or reveal an emerging or dominant group norm if the symptoms and defective decision processes of groupthink are to occur. The philosophical and attitudinal homogeneity cited by Janis as an antecedent condition will often create or influence such norms. Thus, for example, key members of the Bush Administration (among them Vice President Richard Cheney, Secretary of Defense Donald Rumsfeld, and Assistant Secretary of Defense Paul Wolfowitz) long shared a philosophy regarding the strategic wisdom of using unilateral, preemptive military interventions (or their threat) as a key aspect of U.S. foreign policy. This view (a.k.a. the Bush Doctrine) came to be normative within the Bush White House following the World Trade Center attack of 9/11 and precipitated several dramatic administrative decisions including the sequential invasions of both Afghanistan and Iraq, despite protestations from many of our allies. It is important to note in this regard that most of the experimental demonstrations mentioned above involve some form of attitudinal/normative homogeneity. Thus, the research on social influence as well as that regarding pluralistic ignorance document the eVect of emergent or preexisting social norms. Similarly, group polarization occurs only among like-minded groups who tend to value one side of the issue or the other regardless of whether the issue involves risk taking during a football game, caution regarding the selection of a marriage partner, racial equality, or Parisian students’ dislike of Americans.

GROUPTHINK

243

Stasser’s hidden profile paradigm has a related characteristic. Participants in this procedure receive shared information that suggests to all (or almost all) group members that one decision option is most sensible. This option almost always serves as the initial group norm. Moreover, research indicates that this initial norm plays a primary role in mediating the nature of the eventual group decision (e.g., Gigone & Hastie, 1993). Similarly, Raven (1998) cites an historical precursor to the hidden profile research (and the groupthink notion in general) that highlights the eVect of initial norms on discussion outcome. In this early study, Maier and Solem (1952) reported that when groups discussed the Horse Trader problem to consensus, their performance on this eureka task exceeded the individual solver baseline provided that a majority of participants began the discussion favoring the correct solution. If, however, the majority favored an incorrect solution, group discussion lowered performance despite the presence of minority members having insight into the correct solution. This is quite congruent with the view that initial or emergent norms within the group serve to bias discussion and disenfranchise those who hold dissenting or minority views regardless of their correctness. Note that in this study too, one observes a groupthink-like process in the absence of the antecedent factors listed by Janis (e.g., cohesion, crisis, directive leadership, etc.). Finally, the one antecedent factor that has been linked to signs of groupthink in both laboratory and historical studies (e.g., Baron, Crawley, & Paulina, 2003), directive leadership, has a component that generally suggests a preferred solution to group members early in the discussion (e.g., Hodson & Sorrentino, 1997).

C. LOW SELF-EFFICACY The third antecedent condition I propose for groupthink phenomena is low situational self-eYcacy, in which group members generally lack confidence in their ability to reach satisfactory resolution of the conundrum facing them. The emphasis here is on situation-specific (i.e., state) conceptions of self-eYcacy that might be aVected by such things as decisional complexity, fatigue, priming, low self-confidence, or negative social feedback. Low situational self-eYcacy actually was mentioned briefly by Janis and Mann (1977) as one means of lowering group member self-esteem, one of their specified antecedent conditions. I see it however as a more fundamental condition even in cases where self-esteem is unaVected. As pointed out above, several studies indicate that social influence, and suppression of dissent are either dramatically lessened or completely eliminated in conditions in which self-eYcacy is likely to be high because of low task diYculty (e.g., Deutsch & Gerard, 1955), absence of time pressure (Baron et al., 1996),

244

ROBERT S. BARON

or manipulations of perceived self-eYcacy (Stasser & Stewart, 1992; Stasser et al., 2000). In contrast, when situational self-eYcacy is likely to be low whether because of impaired cognitive capacity, recent failure, task diYculty, time pressure, fear, or lack of confidence, social influence tends to be elevated particularly when such influence depends on flawed argumentation, normative pressure, or heuristic message processing (e.g., Baron, 2000). Thus, for example, Kelly, Jackson, and Hutson-Comeaux (1997) found that when group members were given the impression that a rank ordering task had a correct solution and were not pressured to work quickly (high self-eYcacy), the group discussion was less likely to be characterized by attempts at normative social influence. Rather, under these conditions, the group discussions contained considerable reasoning, sharing of facts, and argumentation (i.e., informational social influence), which in turn was associated with greater solution accuracy. If we shift our focus to the eVects of fear, Darley (1966) reported that fear of shock elevated standard conformity eVects, whereas Baron, Inman, Kao, and Logan (1992) found that dental fear increased how persuaded participants were by heuristic cues such as audience approval of a flawed message. One reason low self-eYcacy has not been emphasized as a necessary antecedent condition in prior discussions of groupthink is because the overwhelming majority of studies on this topic hold this feature constant. That is, most research on groupthink focuses on decisions that entail a good deal of ambiguity and decisional conflict. Thus, the international crises examined in historical case studies as well as the complex decision problems examined in laboratory research are both likely to challenge feelings of self-eYcacy in participants. However, once we begin to consider the likelihood that groupthink processes may occur in mundane group contexts we must acknowledge the obvious fact that many problems faced by such groups will often lack the complexity of (say) whether or not to underwrite the Bay of Pigs invasion. Here I would argue that past research on social influence, suppression of dissent, and group decision making (e.g., Deutsch & Gerard, 1955; Kelly et al., 1997; Stasser & Stewart, 1992) strongly suggests that self-eYcacy is likely to be an important antecedent condition for groupthink phenomena particularly with respect to the extent to which group members are willing to risk oVering a dissenting view and likely to internally accept the validity of the group solution. The logic here is that group members often risk (or, at least, fear) serious sanctions when challenging group norms. They are unlikely to take such a risk unless they feel extremely confident regarding their own preferred solution. In contrast, a lack of confidence provides them with more reason to both suppress their (tentative) dissent and internally accept the solution favored by the group (‘‘all those folks can’t be wrong’’).

GROUPTHINK

245

IV. Strong Versus Moderate Versions of the Ubiquity Model It should be apparent by now that the ubiquity model represents more a revision of Janis’s model than a repudiation. The social identification variable modifies Janis’s emphasis on intense-high-status group cohesion as an antecedent condition for groupthink. Similarly, low self-eYcacy amplifies Janis’s prior consideration of this factor. The one major shift is that the ubiquity model assumes that when combined, social identification, salient norms, and low self-eYcacy are both necessary and suYcient to evoke ‘‘groupthink reactions.’’ Such reactions include Janis’s array of defective decision processes as well as suppressed dissent, selective focus on shared viewpoints, polarization of attitude and action, and heightened confidence in such polarized views. Note that such elevated confidence will often evoke the feelings of ingroup moral superiority and invulnerability alluded to by Janis (1972, 1982). As this implies, the more restrictive (i.e., less common) conditions specified by Janis (e.g., crisis, intense cohesion, insulation, member insecurity, directive leadership) are not deemed to be necessary. As a result, one would expect groupthink reactions in a far wider array of group settings than those originally envisioned by Janis. As we have seen above, there is now ample evidence for the relative ubiquity of such defective decision processes even in temporary and trivial groups. A ‘‘strong’’ version of the ubiquity model would hold that the three antecedent conditions specified above are not only necessary and suYcient but exhaustive as well, with other factors such as crisis or cohesion adding nothing as predictor variables. Although a case can be made for this strong position (especially given the conflicting data regarding cohesion) it seems more likely that a more moderate version is likely to be true. This ‘‘moderate’’ version of the ubiquity model leaves open the possibility that many of the antecedent conditions specified by Janis might still, under certain circumstances, heighten the likelihood or intensity of groupthink phenomena. Thus, if directive leadership heightens the salience and nature of group norms in a setting in which such norms may be otherwise vague, or not yet obvious, it is plausible that this might amplify groupthink reactions. In the same vein there is reason to suspect that crisis may serve as a similar ‘‘amplifying condition’’ elevating the intensity of groupthink. Theorists in political sociology have long considered the likelihood that crisis situations created by intergroup conflict heightens the likelihood that directive ‘‘oligarchical’’ leadership will overshadow or replace more democratic processes of decision-making (Michels, 1962). The logic underlying this prediction is that during a crisis, the need for rapid and decisive action makes time-consuming democratic processes too costly or dangerous. An additional mechanism

246

ROBERT S. BARON

that in all likelihood applies here is that the threat of a crisis may lower members’ feelings of self-eYcacy while simultaneously heightening their dependency needs. Such eVects should increase members’ susceptibility to both informational and normative social influence from directive leaders or dominant ruling coalitions. One caveat to this analysis, however, is that threat/crisis level must be substantial to provoke such reactions. Thus, laboratory simulation studies that depend upon participants role-playing or recalling a crisis related setting seem poorly suited to testing the eVects of threat and crisis. Similar criticisms could be made regarding prior lab manipulations of cohesion. These manipulations only rarely test participants in the presence of preexisting cohesive others (see as exceptions Flowers, 1977; Hogg & Hains, 1998; Leana, 1985) relying instead on less powerful manipulations based on such factors as bogus personality feedback or task importance. One of the few studies that does provide support for the role of threat (Turner et al., 1992) exposes participants to a manipulation (public scrutiny) that has a direct and immediate eVect on actual outcomes experienced by these individuals. Interestingly, cohesion in this study was manipulated with a social identity manipulation (e.g., social categorization coupled with giving participants time to get to know each other). Research on the eVect of emotions on stereotyping is also congruent with the argument that meaningful threat and crisis should be capable of amplifying groupthink-related phenomena. This research quite consistently indicates that emotions associated with crisis (i.e., fear and anger), amplify various forms and measures of stereotyping (e.g., Baron et al., 1992; Bodenhausen, Sheppard, & Kramer, 1994; Friedland, Keinan, & Tytiun, 1999; Wilder & Shapiro, 1988). Given that stereotyping is one of the classic symptoms of groupthink (at least when such stereotyping focuses upon ingroup/outgroup attributes), these data point to the potential importance of crisis and threat as amplifying variables. The relation between fear, anger, and stereotyping is thought to be mediated, in part, by a diminution of available attentional capacity under these particular emotions (Baron, 1986, 2000; Bodenhausen 1993, Wilder & Shapiro, 1988), an eVect that should both lower self-eYcacy and heighten participants’ reliance on heuristic cues in problem solving. In accord with the admonition above, the emotional manipulations used in this research are substantial, involving such things as threat of electric shock, threat of public embarrassment (Wilder & Shapiro, 1988), dental surgery (Baron et al. 1992), or the stress of flight training (Friedland et al., 1999). In short there is reason to think that crisis, like directive leadership, may be capable of amplifying groupthink eVects provided that the manipulations are nontrivial. Similarly, it seems plausible that group member insecurity

GROUPTHINK

247

and low self-esteem might amplify group think eVects. Low self-esteem certainly is one pathway to low self-eYcacy and moreover is often assumed to be related to heightened desires for social identification.8 As such, such feelings might well elevate susceptibility to normative as well as informational social influence. Certainly task diYculty could be construed as a manipulation of situational self-esteem, and as noted above, such diYculty manipulations have been found to increase conformity and concurrence seeking. Admittedly, however, research on groupthink per se has generally failed to manipulate or measure the eVects of self-esteem (or member insecurity) or to assess how such feelings relate to symptoms and decisional characteristics of groupthink. Future investigations will hopefully clarify such issues. Although ample research over the last 30 years indicates that the bulk of Janis’s antecedent conditions are clearly not necessary to trigger such phenomena as polarized judgment, outgroup stereotyping, self-censorship, and the illusion of consensus (e.g., Tetlock et al., 1992), the empirical support for the antecedent conditions specified by the ubiquity model (i.e., social identification, salient norms, and low self-eYcacy) is far from definitive. Similarly the assumptions outlined by the ‘‘moderate’’ form of the ubiquity model regarding the possible amplifying eVects of crisis, directive leadership, cohesion, and so on need to be examined with more powerful manipulations than those used in prior laboratory research if we are to gain a fuller understanding of how such factors may contribute to flawed and biased decision-making in group contexts. Our expectation is that careful research may well document meaningful relations between such variables and ‘‘groupthink phenomena’’ given that the gist of the present critique is that the pervasiveness of such reactions has been underestimated by prior theoretical accounts.

Acknowledgments I wish to thank Jerry Suls, Rene Martin, and Penny Baron for their thoughtful comments on the manuscript.

8

We oVer these predictions most cautiously given that the relation between low self-esteem and social identification is a matter of active debate (e.g., Hogg & Abrams, 1993; Long & Spears, 1998), with researchers oVering distinctions between collective self-esteem, personal selfesteem, manipulated self-esteem, trait-based self-esteem, public self-esteem, and private self-esteem.

248

ROBERT S. BARON

References Abrams, D., Marques, J., Bown, N., & Dougill, M. (2002). Anti-norm and pro-norm deviance in the bank and on campus: Two experiments on subjective group dynamics. Group Processes and Intergroup Relations, 5, 163–182. Abrams, D., Marques, J., Bown, N., & Henson, M. (2000). Pro-norm and anti-norm deviance within and between groups. Journal of Personality and Social Psychology, 78, 906–912. Aldag, R. J., & Fuller, S. R. (1993). Beyond fiasco: A reappraisal of the groupthink phenomenon and a new model of group decision processes. Psychological Bulletin, 113, 533–552. Allport, F. H. (1924). Social psychology. Boston: Houghton-Mifflin. Alvaro, E. M., & Crano, W. D. (1996). Cognitive responses to minority or majority based communications: Factors that underlie minority influence. British Journal of Social Psychology, 35, 105–121. Alvaro, E. M., & Crano, W. D. (1997). Indirect minority influence: Evidence for leniency in source evaluation and counterargumentation. Journal of Personality and Social Psychology, 72, 949–964. Aronson, E., Wilson, T. D., & Akert, R. M. (2003). Social psychology (4th ed.). Upper Saddle River, NJ: Prentice Hall. Asch, S. E. (1956). Studies of independence and submission to group pressure: I. On minority of one against a unanimous majority. Psychological Monographs, 70(9), 417–427. Asch, S. E. (1957). An experimental investigation of group influence. In Symposium on preventive and social psychiatry (pp. 15–17). Walter Reed Army Institute of Research. Washington, DC: Government Printing Office. Bargh, J. (1994). Auto-motives: Preconscious determinants of social interaction. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior (pp. 93–130). New York: Guilford Press. Baron, R. A., & Byrne, D. (2000). Social psychology (9th ed.). Boston: Allyne & Bacon. Baron, R. S. (1986). Distraction-conflict theory: Progress and problems. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 1–40). New York: Academic Press 1986. Baron, R. S. (2000). Arousal, capacity, and intense indoctrination. Review of Personality and Social Psychology, 4, 238–254. Baron, R. S., & Kerr, N. (2003). Group process, group decision, group Action (2nd ed.). Buckingham: Open University Press. Baron, R. S., Crawley, R., & Paulina, D. (2003). Aberrations of power. Leadership in totalist groups. In D. Van Knippenberg & M. Hogg (Eds.), New perspectives in leadership (pp. 169–183). Thousand Oaks, CA: Sage. Baron, R. S., Inman, M., Kao, C. F., & Logan, H. (1992). Negative emotion and superficial social processing. Motivation and Emotion, 16, 323–346. Baron, R. S., Hoppe, S., Linneweh, B., & Rogers, D. (1996). Social corroboration and opinion extremity. Journal of Experimental Social Psychology, 32, 537–560. Baron, R. S., Vandello, J., & Brunsman, B. (1996). The forgotten variable in conformity research: The impact of task importance on social influence. Journal of Personality and Social Psychology, 71, 915–927. Bernthal, P. R., & Insko, C. A. (1993). Cohesiveness without groupthink: The interactive effects of social and task cohesion. Group and Organizational Management, 18, 66–87.

GROUPTHINK

249

Bodenhausen, G. V. (1993). Emotions, arousal and stereotypic judgements: A heuristic model of affect and stereotyping. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition and stereotyping (pp. 13–37). New York: Academic Press. Bodenhausen, G. V., Sheppard, L. A., & Kramer, G. P. (1994). Negative affect and social judgment: The differential impact of anger and sadness. Affect in social judgments and cognition (Special Issue). European Journal of Social Psychology, 24, 45–62. Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitivemotivational analysis. Psychological Bulletin, 86, 307–324. Brodbeck, F. C., Kerschreiter, R., Mojzisch, A., Frey, D., & Schulz-Hardt, S. (2002). The dissemination of critical, unshared information in decisionmaking groups: The effects of prediscussion dissent. European Journal of Social Psychology, 32, 35–56. Brown, R. J. (2000). Group processes (2nd ed.). Oxford: Blackwell. Brown, R. J., Tajfel, H., & Turner, J. C. (1980). Minimal group situations and intergroup discrimination: Comments on the paper by Aschenbrenner and Shaeffer. European Journal of Social Psychology, 10, 399–414. Burnstein, G., & Vinokur, A. (1977). Persuasive arguments and social comparison as determinants of attitude polarization. Journal of Experimental Social Psychology, 13, 315–332. Campbell, D. T. (1958). Common fate, similarity and other indices of the status of aggregates of persons as social entities. Behavioral Science, 3, 14–25. Courtright, J. A. (1978). A laboratory investigation of groupthink. Communication Monographs, 45, 229–246. David, B., & Turner, J. C. (1996). Studies in self-categorization and minority conversion: Is being a member of the out-group an advantage? British Journal of Social Psychology, 35, 179–199. Darley, J. M. (1966). Fear and social comparison as determinants of conformity behavior. Journal of Personality and Social Psychology, 4, 73–78. Deutsch, M., & Gerard, H. (1955). A study of normative and informational social influence upon individual judgment. Journal of Abnormal and Social Psychology, 51, 629–636. Devine, T. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5–18. Dion, K. L., Miller, N., & Magnan, M. A. (1971). Cohesiveness and social responsibility as determinants of group risk taking. Journal of Personality and Social Psychology, 20, 400–406. Duncan, B. L. (1976). Differential social perception and attribution of inter-group violence: Testing the lower limits of stereotyping of blacks. Journal of Personality and Social Psychology, 34, 590–598. Esser, J. K. (1998). Alive and well after 25 years: A review of groupthink research. Organizational Behavior and Human Decision Processes, 73, 116–141. Esser, J. K., & Lindoerfer, J. L. (1989). Group think and the space shuttle Challenger accident: Towards a quantitative case analysis. Journal of Behavioral Decision-Making, 2, 167–177. Festinger, L., & Thibaut, J. (1951). Interpersonal communication in small groups. Journal of Abnormal and Social Psychology, 46, 92–99. Flowers, M. L. (1977). A laboratory test of some implications of Janis’s groupthink hypothesis. Journal of Personality and Social Psychology, 35, 888–896. Fodor, E. M., & Smith, T. (1982). The power motive as an influence on group decision-making. Journal of Personality and Social Psychology, 42, 178–185. Forsyth, D. (1999). Group dynamics (3rd ed.). Wadsworth, CA: Brooks Cole. Friedland, N., Keinan, G., & Tytiun, T. (1999). The effect of psychological stress and tolerance of ambiguity on stereotypic attributions. Anxiety, Stress, & Coping, 12, 397–410.

250

ROBERT S. BARON

Fuller, S. R., & Aldag, R. J. (1998). Organizational tonypandy: Lessons from a quarter-century of the groupthink phenomenon. Organizational Behavior and Human Decision Processes, 73, 163–184. Gigone, D., & Hastie, R. (1993). The common knowledge effect: Information sharing and group judgment. Journal of Personality and Social Psychology, 65, 959–974. Goethals, G. R., & Zanna, M. P. (1979). The role of social comparison in choice shifts. Journal of Personality and Social Psychology, 37, 1469–1476. Hamilton, D. L., & Sherman, S. J. (1996). Perceiving persons and groups. Psychological Review, 103, 336–355. Hensley, T. R., & Griffin, G. W. (1986). Victims of groupthink: The Kansas State University Board of Trustees and the 1977 gymnasium controversy. Journal of Conflict Resolution, 30, 497–531. Herek, G. M., Janis, I. L., & Huth, P. (1987). Decision-making during international crises: Is quality of process related to outcome? Journal of Conflict Resolution, 31, 203–226. Hodson, G., & Sorrentino, R. M. (1997). Groupthink and uncertainty orientation: Personality differences in reactivity to the group situation. Group Dynamics: Theory, Research, and Practice, 1, 144–155. Hogg, M. A. (1992). The social psychology of group cohesiveness: From attraction to social identity. New York: New York University Press. Hogg, M. A., & Abrams, D. (1993). Towards a single-process uncertainty-reduction model of social motivation in groups. In M. A. Hogg & D. Abrams (Eds.), Group motivation: Social psychological perspectives (pp. 173–190). New York: Prentice Hall. Hogg, M. A., & Abrams, D. (2001). Intergroup relations: An overview. In M. A. Hogg & D. Abrams (Eds.), Intergroup relations (pp. 1–14). Philadelphia, PA: Psychology Press. Hogg, M. A., & Hains, S. C. (1998). Friendship and group identification: A new look at the role of cohesiveness in groupthink. European Journal of Social Psychology, 28, 323–341. Hogg, M. A., Turner, J. C., & Davidson, B. (1990). Polarized norms and social frames of reference: A test of the self-categorization theory of group polarization. Basic and Applied Social Psychology, 11, 77–100. Janis, I. L. (1971). Groupthink. Psychology Today 5, November, 43–46 & 74–76. Janis, I. L. (1972). Victims of groupthink. Boston: Houghton Mifflin. Janis, I. L. (1982). Groupthink (2nd ed.). Boston: Houghton Mifflin. Janis, I. L., & Mann, L. (1977). Decision-making: A psychological analysis of conflict, choice and commitment. New York: Free Press. Kelly, J. R., Jackson, J. W., & Huston-Comeaux, S. L. (1997). The effects of time pressure and task differences on influence modes and accuracy in decision-making groups. Personality and Social Psychology Bulletin, 23, 10–22. Kerr, N. L., & Tindale, S. R. (2004). Group performance and decisionmaking. Annual Review of Psychology, 55(22), 623–655. Korte, C. (1972). Pluralistic ignorance about student radicalism. Sociometry, 35, 576–587. Kramer, R. M. (1998). Revisiting the Bay of Pigs and Vietnam decisions 25 years later: How well has the groupthink hypothesis stood the test of time? Organizational Behavior and Human Decision Processes, 73, 236–271. Larson, J. R. Jr., Foster-Fishman, P. G., & Keys, C. B. (1994). Discussion of shared and unshared information in decision-making groups. Journal of Personality and Social Psychology, 67, 446–461. Leanna, C. R. (1985). A partial test of Janis’ groupthink model: Effects of group cohesiveness and leader behavior on defective decision making. Journal of Management, 11, 5–17. Levine, J. M. (1989). Reactions to opinion deviance in small groups. In P. B. Paulus (Ed.), Psychology of group influence (2nd ed., pp. 279–326). Hillside, NJ: Erlbaum.

GROUPTHINK

251

Levine, J. M., & Thompson, L. (1996). Conflict in groups. In E. T. Higgens & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 745–776). New York: Guilford. Long, K. M., & Spears, R. (1998). Opposing effects of personal and collective self-esteem on interpersonal and intergroup comparisons. European Journal of Social Psychology, 28, 913–930. Longley, J., & Pruitt, D. G. (1980). Groupthink: A critique of Janis’s theory. In L. Wheeler (Ed.), Review of personality and social psychology (pp. 507–513). Newbury, CA: Sage. Lord, C. G. (1997). Social psychology. Orlando, FL: Holt Rinehart & Winston. Mackie, D. (1986). Social identification effects in group polarization: Pro-norm and anti-norm deviance within and between groups. Journal of Personality and Social Psychology, 50, 720–728. Mackie, D. M., & Cooper, J. (1984). Attitude polarization: Effects of group membership. Journal of Personality and Social Psychology, 46, 575–585. Mackie, D., Gastardo-Conaco, M. C., & Skelly, J. J. (1992). Knowledge of the advocated position and the processing of intergroup and out-group persuasive messages. Personality and Social Psychology Bulletin, 18, 145–151. Maier, N. R. F., & Solem, A. R. (1952). The contribution of a discussion leader to the quality of groupthink: The effective use of minority opinions. Human Relations, 5, 277–288. Marques, J., Abrams, D., & Serodio, R. G. (2001). Being better by being right: Subjective group dynamics and derogation of ingroup deviants when generic norms are undermined. Journal of Personality and Social Psychology, 81, 436–447. McCauley, C. (1989). The nature of social influence and groupthink: Compliance and internalization. Journal of Personality and Social Psychology, 57, 250–260. McCauley, C. (1998). Group dynamics in Janis’s theory of groupthink: Backward and forward. Organizational Behavior and Human Decision Processes, 73, 142–162. McGarty, C., Haslam, S. A., Hutchinson, K. J., & Turner, J. C. (1994). The effects of salient group membership on persuasion. Small Group Research, 25, 267–293. McLeod, P., Baron, R. S., Marti, M. W., & Yoon, K. (1997). The eyes have it: Minority influence in face to face and computer mediated group discussion. Journal of Applied Psychology, 82, 706–718. Michels, R. (1962). Political parties: A sociological study of the oligarchical tendencies of modern democracy. New York: Collier Books. Miller, D. T., & McFarland, C. (1991). When social comparison goes awry: The case of pluralistic ignorance. In J. Suls & T. A. Wills (Eds.), Social Comparison: Contemporary Theory and Research (pp. 287–313). Hillsdale, N.J.: L. Erlbaum Associates. Miller, D. T., & Nelson, L. D. (2002). Seeing approach motivation in the avoidance behavior of others: Implications for an understanding of pluralistic ignorance. Journal of Personality and Social Psychology, 83, 1066–1075. Moorhead, G., & Montanari, J. R. (1986). An empirical investigation of the groupthink phenomenon. Human Relations, 39, 399–410. Mullen, B., Anthony, T., Salas, E., & Driskell, J. E. (1994). Group cohesiveness and quality of decision-making: An integration of tests of the groupthink hypothesis. Small Group Research, 25, 189–204. Murphy, G. (1953). In the minds of man. New York: Basic Books. Myers, D. (2004). Psychology (9th ed.). New York: Worth. Myers, D., & Lamm, H. (1976). The group polarization phenomenon. Psychological Bulletin, 83, 602–627. Myers, D. G., & Bishop, G. D. (1970). Discussion effects on racial attitudes. Science, 169, 778–779.

252

ROBERT S. BARON

Neck, C. P., & Moorhead, G. (1992). Jury deliberations in the trial of U.S. v. John DeLorean: A case analysis of groupthink avoidance and an enhanced framework. Human Relations, 45, 1077–1091. Park, W. (1990). A review of research on groupthink. Journal of Behavioral Decision-making, 3, 229–245. Paulus, P. B. (1998). Developing consensus about groupthink after all these years. Organizational Behavior and Human Decision Processes, 73, 362–374. Peterson, R. S. (1997). A directive leadership style in group decision-making can be both virtue and vice: Evidence from elite and experimental groups. Journal of Personality and Social Psychology, 72, 1107–1121. Peterson, R. S., Owens, P., Tetlock, P. E., Fan, E. T., & Martorana, P. (1998). Group dynamics in top management teams: Groupthink, vigilance, and alternative models of organizational failure and success. Organizational Behavior and Human Decision Processes, 73, 272–305. Pettigrew, T. (1978). The ultimate attributional error: Extending Allport’s definition of prejudice. Personality and Social Psychology Bulletin, 5, 468–476. Pool, G. J., Wood, W., & Leck, K. (1998). The self esteem motive in social influence: Agreement with valued majorities and disagreement with derogated minorities. Journal of Personality and Social Psychology, 75, 967–975. Prentice, D. A., & Miller, D. T. (1993). Pluralistic ignorance and alcohol use on campus: Some consequences of misperceiving the social norm. Journal of Personality and Social Psychology, 64, 243–256. Prentice, D. A., & Miller, D. T. (1996). Pluralistic ignorance and the perpetuation of social norms by unwitting actors. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 28, pp. 161–209). San Diego, CA: Academic Press. Raven, B. (1998). Groupthink, Bay of Pigs, and Watergate reconsidered. Organizational Behavior and Human Decision Processes, 73, 352–361. Sachdev, I., & Bourhis, R. Y. (1987). Status differentials in intergroup behavior. European Journal of Social Psychology, 17, 277–293. Sanders, G. S., & Baron, R. S. (1977). Is social comparison irrelevant for producing choice shifts? Journal of Experimental Social Psychology, 13, 303–314. Schachter, S. (1951). Deviation, rejection and communication. Journal of Abnormal and Social Psychology, 46, 190–207. Sherif, M. (1935). A study of some social factors in perception. Archives of Psychology, 27, 1–60. Sherif, M., Harvey, O. J., White, G. J., Hood, W. R., & Sherif, C. (1961). Intergroup cooperation and competition: The Robbers Cave Experiment. Norman, OK: University Book Exchange. Skinner, B. F. (1956). Science and human behavior. New York: MacMillian. Stasser, G., Stewart, D. D., & Wittenbaum, G. M. (1995). Export roles and information exchange during discussion: The importance of knowing who knows what. Journal of Experimental Social Psychology, 31, 244–265. Stasser, G., & Stewart, D. D. (1992). Discovery of hidden profiles by decision-making groups: Solving a problem vs. making a judgment. Journal of Personality and Social Psychology, 63, 426–434. Stasser, G., Taylor, L. A., & Hanna, C. (1989). Information sampling of structured and unstructured discussions of three and six person groups. Journal of Personality and Social Psychology, 57, 67–78. Stasser, G., & Titus, W. (1985). Pooling of unshared information in group decision making: Biased information sampling during group discussion. Journal of Personality and Social Psychology, 48, 1467–1478.

GROUPTHINK

253

Stasser, G., Vaughan, S. I., & Stewart, D. D. (2000). Pooling unshared information: The benefits of knowing how access to information is distributed among group members. Organizational Behavior and Human Decision Processes, 82, 102–116. Stewart, D. D., & Stasser, G. (1998). The sampling of critical, unshared information in decisionmaking groups: The role of an informed minority. European Journal of Social Psychology, 28, 95–113. Stewart, D. D., Billings, R. S., & Stasser, G. (1998). Accountability and the discussion of unshared, critical information in decision-making groups. Group Dynamics: Theory, Research, and Practice, 2, 18–23. Suls, J., & Green, P. (2003). Pluralistic ignorance and college student perceptions of genderspecific alcohol norms. Health Psychology, 22, 479–486. Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks Cole. Taylor, D. M., & Jaggi, V. (1974). Ethnocentrism and causal attribution in the South Indian context. Journal of Cross-Cultural Psychology, 5, 162–171. Tetlock, P. E. (1979). Identifying victims of groupthink from public statements of decision makers. Journal of Personality and Social Psychology, 37, 1314–1324. Tetlock, P. E., Peterson, R. S., McGuire, C., Chang, S., & Feld, P. (1992). Assessing political group dynamics: The test of the groupthink model. Journal of Personality and Social Psychology, 63, 403–425. Timmons, K. (1991). Groupthink (videorecording). Carlsbad, CA: CRM films. Turner, M. E., & Pratkanis, A. R. (1998). Twenty-five years of groupthink theory and research: Lessons from the evaluation of a theory. Organizational Behavior and Human Decision Processes, 73, 105–115. Turner, M. E., Pratkanis, A. R., Probasco, P., & Leve, C. (1992). Threat, cohesion, and group effectiveness: Testing a social indemnity maintenance perspective on groupthink. Journal of Personality and Social Psychology, 63, 781–796. Wang, G., & McKillip, J. (1978). Ethnic identification and judgments of an accident. Personality and Social Psychology Bulletin, 4, 296–299. Whyte, G. (1998). Recasting Janis’s groupthink model: The key role of collective efficacy in decision fiascos. Organizational Behavior and Human Decision Processes, 73, 185–209. Williams, K. D., Cheung, C. K. T., & Choi, W. (2000). Cyber-ostracism: Effects of being ignored over the internet. Journal of Personality and Social Psychology, 79, 746–762. Williams, K. D., & Sommer, K. L. (1997). Social ostracism by co-workers: Does rejection lead to social loafing or compensation? Personality and Social Psychology Bulletin, 23, 693–706. Wilder, D. A., & Shapiro, P. (1988). Effects of anxiety on impression formation in a group context: An anxiety-assimilation hypothesis. Journal of Experimental and Social Psychology, 25, 481–499. Wittenbaum, G. M., Hubbell, A. P., & Zuckerman, C. (1999). Mutual enhancement: Toward an understanding of collective preference for shared information. Journal of Personality and Social Psychology, 77, 967–978. Wood, W., Lundgren, S., Ouellette, J. A., & Busceme, S. (1994). Minority influence: A metaanalytic review of social influence processes. Psychological Bulletin, 115, 323–345. Wood, W., Pool, G. J., Leck, K., & Purvis, D. (1996). Self-definition, defensive processing, and influence: The normative impact of majority and minority groups. Journal of Personality and Social Psychology, 71, 1181–1193.

This Page Intentionally Left Blank

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

Rupert Brown Miles Hewstone ‘‘Only if a psychological linkage is made between the image of specific individuals and the stereotype of a certain group, only when the individuals can be perceived as ‘typical representatives’ of that group, is the experience with individuals likely to affect the stereotype.’’ (Lewin & Grabbe, 1945, p. 58). ‘‘Where friendly feelings characterize the relationships between members of a group of mixed ethnic composition and individual differences from the stereotypes become apparent, the ‘isolation’ process can be appreciably lessened and the generalization of the newly developed feelings of friendship to the entire ethnic group promoted if the group leaders frequently call attention to the ethnic affiliations of the group members.’’ (Chein, Cook, & Harding, 1948, p. 49). ‘‘In order to predict the effect of contact upon attitudes we should ideally study the consequences of each of the following variables . . . (d) Is the contact perceived in terms of intergroup relations or not perceived as such? (e) Is the contact regarded as ‘typical’ or as ‘exceptional’?’’ (Allport, 1954, pp. 262–263).

*The publication of this article gives us an opportunity to reverse the order of authors’ names, following our original article (Hewstone & Brown, 1986); both authors have, however, made equal contributions to this piece of work. We gratefully acknowledge various sources of funding over the years for this program of work, including (to both authors) the Economic and Social Research Council and (to Miles Hewstone), the Central Community Relations Unit (Northern Ireland), the Templeton Foundation, and the Russell Sage Foundation. We also thank the graduate students, research staV, and other collaborators who have made major contributions to this endeavor; these are made explicit in authorship of primary studies, but we especially thank Ed Cairns, Lindsey Cameron, Anja Eller, Katy Greenland, Jake Harwood, Jared Kenworthy, Pam Maras, Stefania Paolini, Inga PfaVerott, Adam Rutland, Nicole Tausch, Tania Tam, Rhiannon Turner, Jim Vivian, Alberto Voci, Christiana Vonofakou, and Hanna Zagefka. Finally, we thank all of the following for commenting so quickly and helpfully 255 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 37

Copyright 2005, Elsevier Inc. All rights reserved. 0065-2601/05 $35.00

256

RUPERT BROWN AND MILES HEWSTONE

I. Introduction For those concerned with the quality of global intergroup relationships, the 21st century has not begun auspiciously. As we enter only its fifth year, we have already seen unprecedented incidents of international conflict and terrorism (e.g., Afghanistan, 2002; Iraq, 2003; Spain, 2004; USA, 2001), a rising tide of Islamophobia and anti-Semitism in Europe and elsewhere (e.g., Allen & Nielson, 2002), and increasingly hostile immigration receptions experienced by refugees (e.g., MORI, 2003). The task of ameliorating these problematic intergroup relationships will be arduous and complex, undoubtedly necessitating fundamental economic and political transformations, aided by the coordinated eVorts of several social science disciplines. In this article, we consider the contribution that social psychology might make to such an endeavor, focusing particularly on the role of intergroup contact in reducing prejudice and in generating more favorable intergroup attitudes and stereotypes. We begin, appropriately enough in this the half centenary of its publication, with a brief reassessment of Allport’s (1954) classic ‘Contact Hypothesis’ (see also Dovidio, Gaertner, & Kawakami, 2003). Although Allport was by no means the first to advocate bringing groups together as a means of prejudice reduction (Watson, 1947; Williams, 1947), his theorizing in The Nature of Prejudice has justly acquired landmark status because of its lucid exposition, penetrating insights, and careful analysis of the social conditions likely to facilitate the beneficial eVects of contact on intergroup attitudes and behavior. As we shall show, there is now ample evidence to support his basic contentions. However, the past 50 years have also seen important theoretical and empirical developments of the Contact Hypothesis—in particular, Brewer and Miller’s (1984) decategorization model, Gaertner and Dovidio’s (2000) common in-group identity model, Hewstone and Brown’s (1986) intergroup contact model, and Pettigrew’s (1998) theory of longitudinal contact. In this article, we aim to provide a critical assessment of these developments, focusing mainly on our own model, which provided the point of departure for our own broad research program on contact over nearly 20 years. Although our original model was conceived partly as a riposte to

on an earlier draft: Marilynn Brewer, Jack Dovidio, Anja Eller, Sam Gaertner, Katy Greenland, Jake Harwood, Jared Kenworthy, Stefania Paolini, Norman Miller, Tom Pettigrew, Tania Tam, Linda Tropp, Rhiannon Turner, Nicole Tausch, Alberto Voci, and Christiana Vonofakou, and especially Rhiannon Turner and Alberto Voci for their intellectual sustenance and practical help in preparing the final version. This article was awarded the Gordon Allport Intergroup Relations Prize (2005) by the Society for the Psychological Study of Social Issues (SPSSI).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

257

Brewer and Miller’s (1984) principal hypothesis, and it contrasts in key respects with Gaertner and Dovidio’s (2000) subsequent alternative, the evidence that has accumulated, both from our own research and that of others, now suggests to us that some rapprochement of these apparently contradictory accounts is not only possible but also highly desirable. Later in the article, we discuss how this might be achieved. We will start with a brief re-presentation of Allport’s (1954) classic hypothesis and show, with reference to recent cross-sectional and longitudinal surveys, laboratory experiments, and meta-analysis, that many of his original propositions have capably withstood the test of time. We then examine Brewer and Miller’s (1984) and Gaertner and Dovidio’s (2000) attempts to extend the Contact Hypothesis, in both of which categorization processes play a key role. This approach sets the stage for own model, first published in 1986 (Hewstone & Brown, 1986) and subsequently modified (Hewstone, 1996; Vivian, Hewstone, & Brown, 1997). In that model, we were concerned with identifying the conditions that would allow the generalization of attitudes and behavior change beyond the specific context in which the contact occurs. We hypothesized that group salience—broadly speaking, the extent to which group memberships are psychologically ‘present’ during contact— would play a key role in encouraging such generalization, essentially suggesting that it moderates the eVects of contact on prejudice reduction. We are now in a position to evaluate that hypothesis in the light of the extensive evidence that has accumulated from a wide variety of contact settings. A major portion of the article is dedicated to a review of that evidence. A further goal of our theorizing has been to understand the processes that mediate the link between contact and improved intergroup attitudes. In this review, we highlight what seems to be one of the most important of those mediating factors, intergroup anxiety. However, as we shall show, other emotions—both positive and negative—also play an important role in determining the outcomes of contact. The empirical work we present is eclectic in its variety: It includes experimental work in laboratory settings, correlational and longitudinal research in a diversity of field contexts, and studies in more applied domains such as education and acculturation. We conclude with our suggestion as to how these diVerent models might be reconciled, for which Pettigrew’s (1998) similar integration provides an instructive comparison. The article is organized into nine sections. The first briefly reviews the research inspired by the Contact Hypothesis. In Section III, we introduce the three developments of Contact Theory that occurred in the 1980s, including our own model. Sections IV–IX review the empirical research instigated by the Hewstone–Brown model. In the final Section, X, we review progress to date and attempt a theoretical integration of these models in the light of the large volume of research that they have stimulated.

258

RUPERT BROWN AND MILES HEWSTONE

For convenience of exposition, we have organized the material more or less historically. Our early work in this domain was focused on the processes of stereotype change in the face of disconfirming information, such as might be encountered in a contact situation. This is the subject of Section IV. We then move to experimental studies of contact that have attempted to manipulate intergroup salience directly to test the central hypothesis of the Hewstone–Brown model (Section V), that it is only when some group salience is maintained that one will find the generalized attitude change that is the goal of contact interventions. Necessarily, such laboratory studies are limited in the range of intergroup relationships they can investigate. In Section VI, therefore, we present the results from several correlational studies conducted in a wide range of diVerent cultural contexts. The common theme of these studies was to investigate whether group salience, as experienced and reported by participants, moderates the eVects of contact on various indicators of intergroup attitude. In the course of this work, it emerged that an occasional negative by-product of too much salience could be heightened intergroup anxiety, which is usually associated with unfavorable intergroup attitudes, as Stephan and Stephan (1985) had warned some years ago. This finding led to the realization that our original model had rather neglected the aVective processes implicated in intergroup contact, an oversight that we addressed in a revised version (Vivian et al., 1997). This instigated a series of studies that investigated the mediating role of anxiety and other emotions in determining the outcome of contact and that also extended the range of outcome measures employed, including measures of aVect and implicit attitude. These are presented in Section VII, together with some studies that have simultaneously tested moderation and mediation. In Section VIII, we present research that has attempted to apply the Hewstone–Brown model in educational settings. In Section IX, we discuss work on acculturation of ethnic groups in pluralist societies that links directly to models of intergroup contact. Section X presents our theoretical integration.

II. The Contact Hypothesis and Its Early Modifications As is well known, Allport (1954) proposed that prejudice reduction would occur when members of diVerent groups met on an equal status footing to pursue common goals through cooperative interaction, in such a way as to allow the development of close relationships with members of the out-group (i.e., it should have acquaintance potential; Cook, 1962). This contact should also have the support of institutional authorities.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

259

Allport’s (1954) conjectures proved to be enormously stimulating, both for social scientific researchers interested in the mechanisms underlying prejudice reduction and to educators and policymakers concerned with improving ethnic relations in schools, communities, and workplaces. The years following the publication of The Nature of Prejudice saw the accumulation of a vast array of research evidence on the eVects of diVerent kinds of contact on intergroup relations. This literature has now been surveyed several times, both in narrative reviews (Amir, 1969; Amir & Katz, 1976; Hewstone, 1996; Hewstone & Brown, 1986; Pettigrew, 1998; Schofield, 1995) and a meta-analysis (Pettigrew & Tropp, 2004). Although some scholars highlighted inconsistent results, the failure to generalize, diYculties of fulfilling an increasingly long list of apparently necessary conditions, and the realworld impact of the Contact Hypothesis (see Forbes, 1997; Hewstone & Brown, 1986; Stephan, 1987), there is now little doubt that its core propositions have received substantial empirical support. The topic continues to attract a great deal of research attention (see Fig. 1A), and papers focusing on intergroup contact have accounted for a substantial proportion of all papers on intergroup relations in every decade from the 1930s to the present (see Fig. 1B). Pettigrew and Tropp’s (2004) meta-analysis provides the most impressive empirical assessment. Briefly, they were able to locate more than 500 separate contact studies, conducted in a wide range of contexts and involving more than 250,000 participants of various nationalities. Across all these studies, it was apparent that contact per se had beneficial eVects in reducing prejudice: The overall relationship between contact and prejudice was significant, though modest in size (Pearson’s r of just above .20). But this eVect was substantially stronger (around .30) in those contexts in which Allport’s ‘optimal’ conditions applied. We wish to highlight one further result from this meta-analysis here: Where the contact involved out-group friendships, its eVect on prejudice was more pronounced than when it did not ( .26 vs. .22). We will return later to the issue of cross-group friendships. Some of our own research illustrates these broad trends well. Hamberger and Hewstone (1997) exploited a large database compiled from surveys using probability samples in four European countries. The primary independent variables were three diVerent kinds of out-group contact, as ‘friends,’ at ‘work,’ and in the ‘neighborhood.’ The criterion measure was blatant prejudice toward ethnic minorities. All three contact measures showed reliable negative associations with prejudice, but the first (friendship contact) was especially influential (see also Pettigrew, 1997, for a similar conclusion from an independent analysis of the same data set). Friendships with members of an out-group may be important even in quite fraught intergroup contexts. Hewstone, Cairns, Voci, Hamberger, and Niens (in press a), using

260

RUPERT BROWN AND MILES HEWSTONE

Fig. 1. (A) Published research on intergroup contact (269 studies). (B) Percentage of papers on ‘intergroup relations’ investigating intergroup contact, per decade.

representative samples (Ns ¼ 732, 797) of Catholics and Protestants in Northern Ireland, found that the number of friends, relatives, and neighbors from the ‘other side’ that people had was positively predictive of their attitude toward integrated schooling (more than 90% of schools are still segregated by religion). These two studies, like the vast majority of the contact literature, employed cross-sectional designs with their usual attendant problems of causal interpretability. In recognition of this fact, researchers are beginning to employ longitudinal designs more frequently. Maras and Brown (1996) studied young children’s attitudes toward their peers with severe disabilities in the context of a three-month exchange program between a primary school and a special school for children with profound learning and physical

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

261

disabilities. The study employed a quasi-experimental longitudinal design in which half the children were randomly selected to take part in weekly visits to the special school, and half served as controls. The visits involved cooperative play activity between the mainstream and disabled children and were accompanied by the explicit support and encouragement of teachers who also instigated much discussion among the children as to the nature of diVerent kinds of disability, thus making disability categories quite salient. On a key measure of sociometric preference (desire to play with various unknown children, both disabled and nondisabled), the children taking part in the exchange program increased their liking for disabled strangers, while the children in the control conditions did not. There can be little doubt, then, that contact under the right conditions is eYcacious in making intergroup attitudes and behavior more positive, just as Allport had surmised. What is less clear, both in Allport’s (1954) original theorizing and in much of the subsequent research, is what conditions will promote the generalization of those changes—from particular individuals to the group as a whole, from one out-group to another, from one contact situation to another, or from one dimension of evaluation to another (Pettigrew, 1998; Vivian et al., 1997). As Cook (1978) observed, much of the early research on intergroup contact failed to find much evidence of generalization and, yet, for policy purposes, successful generalization is surely crucial. Furthermore, what are the social psychological processes that underlie the changes which contact so evidently brings about? Without a good understanding of these, designing successful interventions to reduce prejudice will always be a somewhat ad hoc process. These were some of the considerations that led to three major reappraisals of contact theory in the 1980s.

III. Theoretical Developments of the Contact Hypothesis Contemporaneously, or at least within the space of five years, three new contact models were proposed in the 1980s: Brewer and Miller (1984); Gaertner, Mann, Murrell, and Dovidio (1989); and Hewstone and Brown (1986). Drawing upon a common theoretical perspective, that oVered by Tajfel and Turner’s (1979) social identity theory, they drew rather diVerent conclusions as to when and how cooperative contact would improve intergroup attitudes both within and beyond the contact situation itself. In this section, we outline these three models, briefly reviewing the evidence for each of the first two. Then, in the following sections, we provide a more detailed account of the work instigated by the third.

262

RUPERT BROWN AND MILES HEWSTONE

Two key assumptions of social identity theory are pertinent here. First, people’s memberships in various groups are incorporated into their selfconcepts to form their social identities. Second, when any particular group membership or dimension of categorization becomes salient in a situation, there will be tendencies toward intergroup diVerentiation and intragroup assimilation that will usually, though not inevitably, result in in-group favoring biases in the service of more positive identity (Tajfel & Turner, 1979; see Brewer, 1979; Hewstone, Rubin, & Willis, 2002; and Mullen, Brown, & Smith, 1992, for reviews of the pervasiveness of in-group bias).

A. THE BREWER–MILLER DECATEGORIZATION MODEL Brewer and Miller (1984, 1988; Miller, 2002) argued that contact situations should be structured so as to reduce the salience of available social categories and increase the likelihood of a more ‘interpersonal’ mode of thinking and behaving. In making this suggestion, Brewer and Miller are drawing on a distinction first articulated by Tajfel (1978) and later elaborated upon by Brown and Turner (1981). In essence, this distinction refers to a psychological continuum whose poles are defined by, on the one hand, an orientation in which the idiosyncratic characteristics of participants and their inter-individual relationships predominate, or, on the other hand, by an orientation in which people’s group memberships and the relations between those groups assume priority. For Brewer and Miller, the goal of this type of interpersonal or ‘decategorized’ contact is simultaneously to achieve greater diVerentiation among out-group members (i.e., less intragroup homogeneity) and also more personalization (i.e., greater attention to and exchange of idiosyncratic information about self, in-group, and out-group members). In this way, the original categories should lose their usefulness for organizing people’s perceptions, emotions, and behavior, and people’s (negative) mutual group stereotypes gradually become disconfirmed as more and more individuating information is acquired. As Brewer and Miller summarized it, such personalized interactions are ‘‘more likely to generalize to new situations because extended and frequent utilization of alternative information featured in interactions undermines the availability and usefulness of category identity as a basis for future interactions with the same or diVerent individuals’’ (Brewer & Miller, 1984, pp. 288–289). Moreover, Brewer and Miller proposed that Allport’s (1954) favorable conditions of contact reduced bias because these features help to decategorize and personalize interactions with out-group members. Three broad lines of evidence support the decategorization hypothesis. First, research on the eVects of crossed categorization, where two or more

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

263

categories overlap (e.g., where people can be simultaneously categorized by ethnicity and gender), suggests that, so long as the crossing categories do not result in a doubly excluded set of group members, in-group bias along any of the component category dimensions is usually reduced (for reviews see Crisp & Hewstone, 1999; Migdal, Hewstone, & Mullen, 1998; Urban & Miller, 1998). In Brewer and Miller’s terms, crossed categorization arrangements result in more diVerentiated out-groups and lessened salience of the constituent category divisions. Second, as noted earlier, there is increasing evidence that contact based on personal friendships is particularly eVective in reducing prejudice (Hamberger & Hewstone, 1997; Pettigrew, 1997, 1998; Pettigrew & Tropp, 2004; Phinney, Ferguson, & Tate, 1997). Third, direct experimental manipulations aimed at increasing personal (vs. task-oriented) contact have proved successful in reducing in-group favoritism, not just toward the members of the out-group actually encountered, but also generalized to other out-group members not yet met (Bettencourt, Brewer, Croak, & Miller, 1992). There is, however, some suggestion that this eVect may be limited to majority group members (Bettencourt, Charlton, & Kernaham, 1997; for a more detailed discussion of how the contactprejudice relationship may vary for members of majority and minority groups, see Tropp, in press). It is clear, then, that some forms of personal contact can be useful in reducing prejudice. What is less clear is, first, whether that contact is truly decategorized or ‘interpersonal’ in the sense defined by Tajfel (1978) and Brown and Turner (1981). Second, if it is, would it then oVer a practical long-term solution for the improvement in intergroup relations? Let us take these issues in turn. In the experimental paradigm that provides the strongest evidence for the Brewer–Miller model (e.g., Bettencourt et al., 1992), members of artificial groups first worked together under diVerent conditions of personal or task focus. During this phase, participants wore large badges denoting their group membership. Two features of this procedure stand out. First, the key manipulation contrasts a ‘personal focus,’ where participants are asked to form impressions of each other, with a ‘task focus,’ where participants have to concentrate more on their contributions to the task at hand. The first condition should, indeed, reduce the salience of the available (quasiminimal) categories, but it is not clear that the second increases it to make the situation more ‘intergroup’ in character. Second, in the ingenious measure of generalization devised by Bettencourt et al. (1992)—ratings of unknown in-group and out-group members shown in a video—some category salience was still maintained since the participants shown in the video retained large identification badges denoting their group membership, the same badges worn by their fellow groups members in the ‘interaction’ phase of the experiment. Similarly, in the several correlational studies testifying to the

264

RUPERT BROWN AND MILES HEWSTONE

power of out-group friendships to reduce prejudice (see Pettigrew & Tropp, 2004) it is seldom clear, nor explicitly controlled for, whether or not those relationships also contain some elements of (inter)group salience. Now turning to the second issue, some minimal level of group salience should be necessary, theoretically speaking, in order to allow generalization from individual out-group members to the out-group as a whole (Brown & Turner, 1981; Hewstone & Brown, 1986). If decategorization was completely successful, there would be no way to connect those now ‘individualized’ persons with the rest of their erstwhile group, as was recognized early on in the development of the model (Brewer, 1988, pp. 27–28; see also Scarberry, RatcliV, Lord, Lanicek, & Desforges, 1997). Furthermore, complete decategorization would presumably also mean the loss of group identities for individuals. While this is not serious in laboratory settings, where the categories are typically of little significance, where real-life groups are concerned their members may be unable or unwilling to relinquish identities so easily, especially if those groups are minorities (Simon, Aufderheide, & Kampmeier, 2001). All in all, then, there are reasons to believe that the decategorization approach, if pursued exclusively or to extremes, might not oVer an optimal strategy for the reduction of prejudice.

B. THE GAERTNER–DOVIDIO COMMON IN-GROUP IDENTITY MODEL The possibility that some in-group bias would be instigated by making a categorization salient was the starting point for Brewer and Miller’s (1984) analysis. Gaertner and Dovidio’s model also takes this as its point of departure, but from there they reach a rather diVerent conclusion (Gaertner & Dovidio, 2000; Gaertner et al., 1989). Instead of arguing for the dissolution of category boundaries, Gaertner and Dovidio propose that redrawing them is a more eVective strategy. They suggest that contact situations could be profitably engineered or transformed so that the current in-group and outgroup can be readily recategorized into a larger superordinate entity. In this way, former ‘out-groupers’ now share a new in-group membership with former ‘in-groupers,’ the so-called ‘common in-group identity,’ and the bias previously associated with the original categories should be reduced or eliminated. In contrast to the decategorization model, which implies that bias will be reduced by less positive evaluations and feelings toward the ingroup (since categories have lost their psychological significance), the common in-group identity model implies a more favorable attitude toward the out-group (since it is now part of the in-group). Like Brewer and Miller

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

265

(1984), Gaertner and Dovidio contend that Allport’s (1954) favorable conditions of contact reduce bias, in part, because they change cognitive representations from two- to one-group (for further discussion and applications to domains as diVerent as multi-ethnic high schools and banking mergers, see Anastasio, Bachman, Gaertner, & Dovidio, 1996; Gaertner, Rust, Dovidio, Bachman, & Anastasio, 1996). There is now an impressive body of evidence supporting the common ingroup identity model (for a review see Gaertner & Dovidio, 2000). Several experimental studies using artificial groups have attested to the power of a superordinate categorization to reduce the amount of in-group bias shown, especially in comparison with situations in which two group memberships remain salient, but also compared with ‘individualized’ conditions (Gaertner et al., 1989; Gaertner, Mann, Dovidio, Murrell, & Pomare, 1990; Gaertner, Dovidio, Rust, Nier, Banker, Ward et al., 1999). One drawback of this program of work is that it has seldom assessed whether the reductions in prejudice that were observed were generalized beyond the particular situations or people in them. This result was remedied in some field studies using real groups (e.g., ethnic groups in schools and colleges, bank employees), and largely convergent results have been obtained (Gaertner & Dovidio, 2000). One advantage of the common in-group identity model over the decategorization approach is that it does not advocate the wholesale abandonment of category memberships but merely their substitution by more inclusive groupings. Nevertheless, the two problems we noted concerning the Brewer– Miller model also pertain to the Gaertner–Dovidio model. The question of generalization remains unresolved because it seems unlikely that recategorization, if it were wholly successful, would facilitate the extension of the more positive intergroup attitudes to out-group members outside the recategorized situation. Also, the common in-group strategy requires groups to relinquish their existing identities (e.g., ethnic, national) in favor of some superordinate identity (e.g., national, pan-national), and this requirement may not always be so easily achieved (see Hewstone et al., 2002). For these reasons, Gaertner, Dovidio, Anastasio, Bachman, and Rust (1993) suggested that a ‘dual identity’ strategy may be preferable in some intergroup situations. In this approach, attempts are made simultaneously to retain the salience of subgroup identities within the superordinate group. We shall return to this idea later, since it bears on one of the central arguments of this article, but suYce it to say that there are already indications that it may oVer substantial advantages over the ‘pure’ common in-group identity approach, particularly in minority-majority settings (Dovidio, Gaertner, & Validzic, 1998; Gaertner & Dovidio, 2000; Hornsey & Hogg, 2000).

266

RUPERT BROWN AND MILES HEWSTONE

C. THE HEWSTONE–BROWN INTERGROUP CONTACT MODEL If the two preceding approaches imply the erosion of in-group–out-group boundaries, albeit in diVerent ways, we have argued that there can be advantages in maintaining intergroup salience, so long as the other Allport conditions apply (Hewstone & Brown, 1986; Vivian et al., 1997). Our reasoning can be deduced from the above criticisms of the other two models: Insofar as the contact takes place on an ‘interpersonal’ (Brewer–Miller) or ‘intragroup’ (Gaertner–Dovidio) basis, there is very little potential for the positive attitudes and emotions engendered by that contact to transfer to other out-group members not yet encountered. Thus, drawing on the distinction between ‘interpersonal’ and ‘intergroup’ modes of interaction (Tajfel, 1978), we proposed an alternative hypothesis: that if the contact can be arranged so that it takes place between in-group and out-group members who can be regarded as suYciently typical or representative of their groups, then the positive changes that occur should generalize to those groups as a whole. One of the necessary conditions for this to happen is that the group memberships retain some psychological salience, perhaps through symbolic representation or via structural arrangements (Brown & Turner, 1981). Although at first sight this proposal might seem rather paradoxical, it is interesting that we are neither alone nor the first to argue for ‘intergroup’ contact (see our epigrams from Allport, 1954; Chein et al., 1948; and Lewin & Grabbe, 1945). In addition to the individual-group generalization that might be conferred by maintained intergroup salience, such a strategy might also avoid the ‘assimilation’ risks carried by the decategorization and recategorization models. Since the groups concerned—be they occupational, ethnic, or national—are not obliged to give up their identities, there should be less resistance to contact interventions (Hewstone & Brown, 1986). In its earliest exposition, our model focused on two central ideas. First, under the heading ‘interpersonal vs. intergroup contact,’ we contended that: ‘‘Unless the contact can be characterized as intergroup (i.e., between individuals as group representatives or qua group members), any such positive outcomes will be primarily cosmetic, in the sense that they will leave divisive and conflictual intergroup relations unchanged’’ (Hewstone & Brown, 1986, p. 16). Second, under the heading ‘mutual intergroup diVerentiation,’ we proposed that mutual recognition of superiorities and inferiorities would characterize intergroup contact in which members acknowledged their own and the other group’s respective strengths and superiorities. Subsequently, however, we have focused almost exclusively on the first idea (i.e., that contact should be intergroup contact, with category memberships salient), and this we still see as central to our intergroup model of contact. Hence, we no longer refer to it

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

267

as the ‘mutual intergroup diVerentiation model’ but prefer to label it simply as ‘intergroup contact theory.’ The original formulation of this model was based mainly on a detailed analysis of the contact literature and especially on the frequent failure of researchers to observe much generalization from particular out-group exemplars to the wider category, combined with the insights from social identity theory. At that time, there was little direct evidence bearing on the central hypothesis concerning salience (an exception being Wilder, 1984). Over the past two decades, however, we have devoted considerable energy to the testing and refinement of the model, and in the next six sections we review the results of that work. As we hope to show, the cardinal idea of the model seems well supported. Perhaps inevitably, however, the research has also revealed shortcomings in our original thinking, particularly our neglect of the important roles that out-group friendships and intergroup emotions play in prejudice reduction.

IV. Experimental Studies of the Impact of Stereotype-Disconfirming Information: A Cognitive Analysis of Contact We and other researchers have conducted a series of studies, inspired by Rothbart and John’s (1985) cognitive analysis of intergroup contact, that focus on how perceivers process stereotype-relevant, and especially stereotype-disconfirming, information, and under what circumstances stereotypes change or are maintained. Interestingly, while most of this work focuses on stereotypes (primarily in the form of central tendency measures, of the extent to which members of a group are characterized by various attributes), most studies of intergroup contact have focused on change in attitudes (or evaluations of a group and its members). For a discussion of the relationship between stereotypes and prejudice, see Dovidio et al. (1996). Rothbart and John (1985) argue that, since objects, or exemplars, diVer in the degree to which they are viewed as prototypical examples of a category (see Barsalou, 1987), then we should accept that it is the goodness-of-fit to the stereotype, and not just a few defining features, that determines whether a person becomes associated with a given category. Rothbart and Lewis (1988) showed that as prototypicality increased, the degree of inference from member to group increased (see also Wilder, Simon, & Faith, 1996). From this view, disconfirming attributes are most likely to become associated with the stereotype if they belong to an individual who is otherwise a very good fit to the category. This view leads to the somewhat paradoxical and counterintuitive proposition that stereotype-disconfirming information should be linked

268

RUPERT BROWN AND MILES HEWSTONE

to typical out-group members (see also Wilder, 1986). This view is consistent with our idea that category salience should be maintained in contact settings (Hewstone & Brown, 1986). Unless this is the case, people tend to react to stereotype-disconfirming information not with generalization but with what Allport (1954) called ‘‘re-fencing.’’ The ‘special case’ is excluded and the category held intact (see also Williams, 1964). Rothbart and John’s (1985) prototype model has received support from experimental studies investigating three cognitive models of stereotype change: ‘bookkeeping,’ ‘conversion,’ and ‘subtyping’ (Weber & Crocker, 1983). The bookkeeping model (Rothbart, 1981) proposes a gradual modification of stereotypes by the additive influence of each piece of disconfirming information. Any single piece of disconfirming information elicits only a minor change in the stereotype; major change occurs gradually and only after the perceiver has accumulated many disconfirming instances that deviate systematically from the stereotype. The conversion model (Rothbart, 1981) envisages a radical change in response to dramatic disconfirming information but no change in response to minor disconfirming information. Finally, the ‘subtyping’ model of stereotype change views stereotypes as hierarchical structures, in which discriminations can be created in response to disconfirming information (Ashmore, 1981; Brewer, Dull, & Lui, 1981; Taylor, 1981). Subtyping occurs when a perceiver responds to members of a target group who disconfirm their stereotype by seeing them as exceptions to the rule and placing them in a separate subcategory apart from members who confirm the stereotype (Richards & Hewstone, 2001). One serious consequence of subtyping is that it may insulate the superordinate stereotype from change (Weber-Kollmann, 1985). We tested these models in a series of studies that compared stereotype change in response to diVerent patterns of disconfirming information. In these studies, information that disconfirmed the stereotype was either ‘dispersed’ across several group members (each of whom slightly disconfirms the stereotype) or ‘concentrated’ in a small number of highly disconfirming members (Weber & Crocker, 1983), and the impact of this information on the group stereotype was assessed. Our studies strongly support the subtyping model (although there is some scattered support for the other models; see Hewstone, 1994, for a review). Our studies have also specified the cognitive processes underlying stereotype change. Johnston and Hewstone (1992) showed, first, that weak disconfirming members (in the dispersed condition) were rated more typical than were strong disconfirming members (in the concentrated condition). Moreover, this perceived typicality was the only dependent measure that mediated the impact of condition (concentrated vs. dispersed) on stereotyping, exactly in line with our model of intergroup contact. This mediating role

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

269

of perceived typicality has also been demonstrated in three other independent studies, generalizing across manipulations, participant groups, and target groups (Hantzi, 1995; Hewstone, Hassebrauck, Wirth, & Waenke, 2000; Maurer, Park, & Rothbart, 1995). However, the association between perceived typicality and stereotyping is best thought of as bi-directional (Hewstone & Hamberger, 2000; Maurer et al., 1995; Park, Wolsko, & Judd, 2000). A ‘prototype-subtyping’ model seems to provide the best account of how stereotypes change in response to dispersed or concentrated patterns of disconfirming information. Stereotype change is generally eVected via the perceived typicality, or goodness-of-fit, of mild disconfirmers in the dispersed condition; it is generally impeded by the atypicality, or badness-offit, of strong disconfirmers in the concentrated condition (see also Desforges, Lord, Ramsey, Mason, Van Leeuwen, West, & Lepper, 1991; Werth & Lord, 1992). Later studies yielded data consistent with the view that subtyping is associated with active preservation of the stereotype (rather than leading either to change, or to inertia; see Hamilton & Sherman, 1994). We manipulated whether the presented information contained one or two subtypes of disconfirming members and found greater stereotyping in the two- vs. onesubtype conditions, when subtypes were based on disconfirming information (Hewstone, Macrae, GriYths, Milne, & Brown, 1994; Expt. 1). Hence, where contact allows, or worse facilitates, subtyping, generalized change in outgroup attitudes will tend to be impeded. We have also shown that the impact of the concentrated vs. dispersed pattern of disconfirming information is influenced by manipulation of the target group’s perceived variability (Hewstone & Hamberger, 2000). When the group was presented as being low in variability (i.e., most members are alike), then we found the usual eVect of more stereotyping under concentrated than dispersed information. But when the group was presented as being high in variability (i.e., group members are diVerent from one another), then we found no diVerence in the impact of the two patterns of disconfirming information. Again, perceived typicality was important. Disconfirmers were rated more typical in the dispersed than concentrated condition only when the group was presented as being low in variability. This greater perceived typicality was partly responsible for the reliably lower stereotyping found in the high variability/concentrated condition, compared with the low variability/concentrated condition. In Study 1, perceived typicality of disconfirmers partially mediated stereotyping. However, Study 2 showed that typicality and stereotyping eVects occurred simultaneously or interactively, rather than one causing the other (cf. Maurer et al., 1995). Overall, these studies provide strong support for the role of perceived typicality in experimental analogues of intergroup contact. They are

270

RUPERT BROWN AND MILES HEWSTONE

consistent with current theory about the ways in which individuals process information about out-group members (Richards & Hewstone, 2001). When group memberships are salient, individuals are perceived as members of the group, and contact has the potential to change group perceptions. However, when salience is low, it is unlikely that perceptions of the group will be changed—contact has no implications for group attitudes when the outgroup individual’s group aYliation is not noticed or processed. Group members who have better fit to the stereotype are more eVective ‘carriers’ of stereotype change, as long as information is provided in such a way that their typicality can be perceived accurately and the target group is perceived as relatively homogeneous; perceived typicality sometimes mediates stereotype change, but the relationship can also be bi-directional.

V. Experimental Studies of Group Salience The first attempt to manipulate group salience in the context of cooperative intergroup contact was by Wilder (1984; Experiment 1). He systematically varied the typicality of the out-group member in a simulated intergroup contact situation. The nature of the contact was also varied in line with traditional theorizing on contact. Thus, the contact person behaved either in a pleasant and supportive way toward the real participants or in a less pleasant and more critical fashion. The interaction took place over a cooperative task. Wilder predicted that only in the combined conditions where the interaction was pleasant and the partner could be seen as typical of her college would ratings of the out-group college become more favorable. Wilder’s results were exactly in line with his prediction. Brown, Vivian, and Hewstone (1999) extended this paradigm by contriving a contact situation in which British students played two cooperative games with a German confederate. There were two independent variables. First, we manipulated the perceived typicality of the German partner, eVected by some initial introductions which revealed him/her to have either a typically German name and some neutral to positive stereotypic attributes, or an untypical name and some counterstereotypic attributes. Second, we manipulated the perceived homogeneity of the German out-group, indicated by some fictitious survey data which showed German students to have either fairly homogeneous or heterogeneous social attitudes. According to Brown and Turner’s (1981) analysis of the distinction between ‘intergroup’ and ‘interpersonal’ situations, both increased typicality and homogeneity should cause situations to be seen in a more ‘intergroup’ fashion. These manipulations aVected generalized evaluations of Germans as a whole. On a positive

271

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

TABLE I Generalized Ratings of Out-Group as a Whole After Contact Under Different Conditions of Typicality and Out-Group Homogeneity (From Brown et al., 1999) Experimental condition Rating Dimension

Typical:Homog Typical:Hetero Atypical:Homog Atypical:Hetero

Positive stereotypic

6.43

6.04

5.43

5.84

Positive non-stereotypic

4.65

4.71

4.32

4.49

stereotype index, ratings were highest in the Typical:Homogeneous cell, as predicted (see Table I). To put it another way, the simple eVect of typicality was reliable only when the out-group was seen as reasonably homogeneous, nicely paralleling one of the findings on stereotype change reported earlier (Hewstone & Hamberger, 2000). On a stereotypically neutral but still positive index, typicality had a main eVect: Those who had worked with typically German partners rated German people more favorably than did those exposed to the atypical partner. Van Oudenhoven, Groenewoud, and Hewstone (1996) manipulated salience in a somewhat diVerent manner. Dutch school students participated in a cooperative learning group with a Turkish ‘peer’ (actually a confederate). There were two salience conditions. In one condition, participants (including the confederate) were introduced to each other by the experimenter early on in the session, and explicit references were made to their respective ethnicities so that it was obvious that the confederate was of Turkish origin (High:High salience). In the second condition, these introductions were eVected later on, about halfway through the session (Low:High salience). In the control condition, no references were made to ethnicity at any point (Low:Low). The intention of varying the timing of salience manipulation was to investigate whether there can be social advantages in not introducing group salience until some level of interpersonal intimacy has been achieved (Pettigrew, 1998), an issue we return to later. At the conclusion of the learning session, participants were asked to evaluate the particular Turkish person with whom they had worked on a number of trait ratings, and then, apparently for a diVerent study (and in a diVerent location), ‘Turkish people in general’ on a slightly larger set of traits. The results were clear. In all three conditions, the Turkish confederate was evaluated equally favorably, presumably a consequence of the pleasant cooperative interaction they had just experienced. However, the ratings of Turkish people generally showed a marked diVerence between the salience and control conditions (see Fig. 2). When the confederate’s nationality had been made explicit, whether early or

272

RUPERT BROWN AND MILES HEWSTONE

Fig. 2. Ratings of individual out-group partner and out-group as a whole under diVerent conditions of group salience (from Van Oudenhoven et al., 1996).

late in the proceedings, the favorable attitude toward him generalized to the category as a whole. The main thrust of the work presented thus far has been to contrast the eVects of individualized or ‘decategorized’ contact with encounters under conditions of intergroup salience. In that sense, they could be regarded as competitive tests of the Brewer–Miller and Hewstone–Brown models. We now turn to more recent work that has sought to examine the relationship between these two accounts and the Gaertner–Dovidio common in-group identity model. It will be recalled that Gaertner and his colleagues have suggested that a dual-identity approach might be a particularly useful strategy in real-world contexts where group identities are likely to be strongly held (Gaertner et al., 1993). This approach represents a combination of the Gaertner–Dovidio and Hewstone–Brown prescriptions since the subgroup identities should remain salient within the more inclusive category. In a series of experiments, we have compared the dual-identity strategy with the Brewer–Miller, Hewstone–Brown, and Gaertner–Dovidio approaches taken singly. The first study combined elements of Bettencourt et al.’s (1992) and Gaertner et al.’s (1989) paradigms (Gonza´lez & Brown, 2003). Ad hoc groups of equal size and status interacted first on their own (in order to generate some in-group cohesiveness). They then met to undertake a cooperative task under one of four conditions: (1) ‘separate individuals,’ where participants wore diVerent color T-shirts, had their photograph taken individually, and sat at separate tables; (2) ‘two groups,’ where the two groups wore diVerent color T-shirts, had their photographs taken as groups, and sat across a table from the other group; (3) ‘one group,’ where everyone wore the same T-shirt

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

273

(of their university), were photographed all together, and sat around a table in alternating fashion; and (4) ‘dual identity,’ where the groups wore diVerent-colored university T-shirts, were photographed both in their separate groups and together, and sat across the table from the other group. Notice that the ‘dual identity’ condition diVered from the ‘two groups’ condition in the crucial respect of simultaneously emphasizing the original subgroups and the new superordinate category. Afterward, participants rated and gave symbolic rewards both to the other participants present and to unknown participants shown in a video. Results showed no reliable eVects for conditions on the measures pertaining to the situation itself, perhaps because of the positive atmosphere engendered by the cooperative activity. However, on the generalized measures (of the videotaped participants) more bias was shown in the ‘separate individuals’ and ‘two groups’ conditions than in the ‘one group’ and ‘dual’ conditions. The latter two conditions both seemed to be eVective in reducing bias to zero and could not be distinguished from one another. In two follow-up studies using the same paradigm (but dropping the ‘two groups’ condition), group status and size were varied experimentally (Gonza´lez & Brown, in press). Once again, the ‘one group’ and ‘dual’ conditions resulted in less generalized bias than the ‘separate individuals’ condition. However, the relative benefits of the two strategies involving a superordinate categorization depended on group size. For numerical minority groups, whether of low or high status, the ‘dual’ condition resulted in the least bias. For numerical majority groups, on the other hand, none of the three categorization conditions diVered from one another, and in-group status was the determining factor—high-status groups always showed more bias than lowstatus groups, regardless of condition (see Fig. 3). The fact that dual-identity strategies seem to be diVerentially eVective for majority and minority groups has been noted by others (e.g., Gaertner & Dovidio, 2000) and will be seen again when we come to consider acculturation research.

A. LIMITS TO THE IMPORTANCE OF TYPICALITY We should be cautious, however, in arguing that perceived typicality always leads to, or is associated with, greater generalized change. To some extent, this will depend on the choice of target group, the nature of current intergroup relations, and the specific details of the contact situation. For example, experimental induction of empathy may not require additional manipulation of salience. Vescio, Paolucci, and Hewstone (2004; Study 1) manipulated the perspective that heterosexual participants adopted as they listened to a gay male individual describe his experiences with discrimination. According to

274

RUPERT BROWN AND MILES HEWSTONE

Fig. 3. Generalized in-group bias in reward allocations under diVerent conditions of group size, status, and category salience (from Gonza´lez & Brown, in press).

Batson, Polycarpou, Harmon-Jones, ImhoV, Mitchener, Bednar, et al. (1997), encouraging a person to adopt the perspective of another should arouse empathy for an out-group individual and result in an increased valuing of the stigmatized individual. Vescio et al. (2004) found that heterosexual participants who were encouraged to adopt the perspective of the gay male target (as opposed to those encouraged to remain detached and objective) reported more empathy toward the target and more favorable attitudes toward gay men in general (assessed an hour later). However, attitudes toward gay men were unrelated to the degree that stimulus targets were perceived as confirming or disconfirming stereotypes. Possible explanations for these findings include that the gay male target was suYciently typical in both conditions (although the confirming target was seen as significantly more typical than the disconfirming target, mean ratings of both targets were above the midpoint) or that the gay man schema

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

275

was chronically accessible. Batson et al. (1997) remarked specifically on the fact that their ‘emotion based’ approach ‘‘did not seem vulnerable to the subcategorization eVects that often plague cognitive approaches such as learning stereotype-inconsistent information about an individual group member’’ (p. 117) (see our earlier discussion of stereotype change and typicality). They suggested that positive empathic feelings for an individual target may generalize to the group as a whole because the group membership of the target of empathy is salient. Wolsko, Park, Judd, and Bachelor (2003) have argued that the moderating eVect of typicality further depends on the nature of the dependent variable. They propose that, in general, positive, cooperative intergroup contact results in more positive evaluations of the target group but that contact leads to changes in perceived variability only when a group member both disconfirms the stereotype and is also seen as typical of the group. They provided evidence consistent with this dual-route hypothesis for diVerent components of group views in their study of whites’ evaluations and stereotypes of Latinos. However, their results diVer from both Wilder’s (1984) and Brown et al.’s (1999) and from the bulk of evidence we report in this article showing that views of the out-group (including evaluations and attitudes) are generally moderated by category salience or typicality. One ‘‘distinct possibility’’ (Wolsko et al., 2003, p. 106) they note is that the interaction in their study was both more positive and of a much more ‘intergroup’ nature than was Wilder’s (see Brown & Turner, 1981; Hewstone & Brown, 1986; Hewstone & Lord, 1998). As a general finding in these experiments, however, heightened intergroup salience, in contact situations that otherwise satisfied Allport’s conditions (excluding institutional support), led to more favorable generalized intergroup attitudes. This supports the original Hewstone–Brown hypothesis and is inconsistent with the opposite Brewer–Miller prescription. It is also noteworthy that convergent evidence for the importance of salience has been provided by other researchers working in quite diVerent intergroup contexts (e.g., Desforges et al., 1991; Desforges, Lord, Pugh, Sia, Scarberry, & RatcliV, 1997; Scarberry et al., 1997; Werth & Lord, 1992).

VI. Correlational Studies Using Group Salience as a Moderator Variable In our experimental work, we have sought to vary salience, or its proxies, and examined the eVects on various indicators of in-group bias. Such a methodology is seldom possible in more naturalistic settings, and so there

276

RUPERT BROWN AND MILES HEWSTONE

we have adopted a diVerent approach. In a number of field studies, we have obtained measures of both the quantity and quality of contact that respondents report having with member(s) of an out-group. We also measured subjective group salience, usually by a reliable index based on perceived typicality of the out-group person(s) with respect to their group, self-reports of how frequently respective group memberships seem to feature in respondents’ interactions with members of the out-group, and how aware respondents were of group memberships during contact, and so on. Several criterion variables have been used in these studies. Often it was some measure of attitude toward the out-group as a whole; in other studies, measures of out-group trust and forgiveness were used; and sometimes more ‘cognitive’ measures such as perceived out-group homogeneity served as the dependent variable. Then, using conventional correlational techniques (e.g., multiple regression, path analysis, and structural equation modeling) we have tested whether the relationship between contact and intergroup attitude is moderated by group salience. There are two forms of what we call the ‘moderation hypothesis’ of intergroup contact. The strong form of this hypothesis is that we predict that the association between contact and outcome variables will be significant only for those participants with high levels of category salience during contact; the weak form of this hypothesis is that the association between contact and outcome variables will be stronger for those participants with high vs. low levels of category salience during contact (although both may be significant). In presenting the work in this and succeeding sections, we depend crucially on the distinction made by Baron and Kenny (1986) between moderator and mediator variables (see also Kenny, Kashy, & Bolger, 1998). We have argued that this distinction is central to understanding the mechanisms and processes of intergroup contact (Hewstone, 1997). As Baron and Kenny explain, the distinction is best understood in terms of the kinds of questions one is asking in research: Moderator variables address ‘when’ questions (e.g., when does contact between members of diVerent groups lead to an improvement in out-group attitudes?), whereas mediator variables address ‘how’ or ‘why’ questions (e.g., how or why does contact improve attitudes?). Both moderation and mediation eVects involve more than two variables; that is, they both deal with what happens when a third variable comes into play. But they do so in very diVerent ways. Moderation implies that the level of the third variable can change the relationship between the other two variables, whereas mediation implies that the relationship between the two variables can actually be created by the third variable.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

277

A. A METHODOLOGICAL CAVEAT Because almost all of our own survey research to date has been crosssectional, we will precede the next part of our review with a brief methodological digression concerning just how confidently we can draw inferences consistent with our theory from these studies. Correlational studies preclude our making strong statements about the direction of causal influence: Do varying amounts of contact bring about change of intergroup attitude, or do people with diVerent prior attitudes diVerentially seek out contact with outgroup members? Both hypotheses are plausible, and in many contexts it is likely that some circular causality will be at work. However, in a few cases, the use of longitudinal designs or sophisticated modeling analyses do permit some causal interpretations and show that contact under certain conditions does indeed lead to generalized attitude change. Moreover, this problem has been addressed in three diVerent ways, each of which gives us confidence in drawing quite strong inferences from correlational data to the eVect that contact substantially aVects attitudes and other outcomes (see Tausch, Kenworthy, & Hewstone, in press). First, researchers have tested eVects using more sophisticated statistical models. Pettigrew (1997) used a nonrecursive structural equation model and tested both directions of causality: from contact to attitudes and from attitudes to contact. He found that both paths were significant but that the path from contact to attitudes was somewhat greater than the path from attitudes to contact. Powers and Ellison (1995) used endogenous switching regression models to test the reverse path and reported similar results. We have used the same technique as Pettigrew, where possible, in our own recent research (see below). Second, researchers have assessed the eVect of contact in situations in which participants were given no choice whether to participate in intergroup contact; thus, prior attitudes could not have been driving contact. In their meta-analysis, Pettigrew and Tropp (2004) reported that no-choice studies yielded by far the largest eVect sizes between contact and attitudes. Third, the problem of causality has also been addressed in a few longitudinal studies, although longitudinal studies are still relatively rare in a database of more than 500 studies in the meta-analysis (for exceptions, see Eller & Abrams, 2003, 2004; Gerard & Miller, 1975; Greenland & Brown, 1999; Hamilton & Bishop, 1976; Levin, van Laar, & Sidanius, 2003; Maras & Brown, 1996; Stephan & Rosenfield, 1978). In a particularly impressive example, Levin et al. (2003) collected data from college students over a period of 5 years. Their results indicate that students who reported less favorable ethnic attitudes (and more intergroup anxiety) in their first year were indeed

278

RUPERT BROWN AND MILES HEWSTONE

less likely to have out-group friends during their second and third years of college, a finding that is consistent with the reverse direction, from attitudes to contact. Nevertheless, those students with more out-group friends in years 2 and 3 had more positive attitudes and were less anxious in year 5, even after their prior attitudes, friendships, and a number of relevant background variables were controlled for. Notably, both causal paths were equally strong. Similarly, Eller and Abrams (2003) found supporting evidence for both causal directions. In addition, of course, experimental studies of intergroup contact (e.g., Brown et al., 1999; Wilder, 1984; Wolsko et al., 2003) yield unambiguous evidence that manipulated contact as an independent variable can and does cause changes in attitudinal and other dependent variables. Thus, given the available empirical evidence, the most plausible answer to the question of causality seems to be the operation of a bi-directional or cumulative process, in which contact reduces prejudice, which in turn makes more contact more likely (Pettigrew, 1997).

B. EMPIRICAL EVIDENCE FOR SALIENCE MODERATION FROM SURVEYS The first of these studies was conducted in a European context (Brown et al., 1999; Study 2). Students (N ¼ 293) from six European countries were asked to nominate someone they knew in another country of the European Union. They then provided ratings of the amount and quality of the contact they had with this person and also how competitive that relationship was (recall that competition is a negative condition of contact in Allport’s hypothesis). Salience was measured in the way indicated above, and the criterion variable was the desire to live in the out-group country in question. Figure 4 shows the results of the regression analysis for the respondents who had contact with a German (similar findings were obtained collapsing across all countries). As can be seen, the salience variable proved to be a significant moderator, as predicted. Among respondents reporting that nationalities were highly salient in their relationship with their contact person, there was a reliable relationship between contact and (positive) out-group attitude; for the ‘low’ salience respondents, there was not. In a large follow-up study (N ¼ 570), Gonza´lez and Brown (2001) introduced some additional controls and more reliable criterion variables. Student respondents from four European countries rated the quantity, quality, and group salience of the contact with someone from each of four out-group countries (held constant for all respondents). In addition, they rated those contact persons on several characteristics (e.g., hard-working, intelligence,

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

279

Fig. 4. Group membership salience moderates eVects of contact on desire to live in another country (from Brown et al., 1999).

arrogance). They then rated the four out-group countries and their own country on those same descriptors and indicated how much positive aVect they felt toward each. Replicating the Brown et al. (1999) finding with a diVerent dependent measure, salience was once again a significant moderator of the contact–out-group attitude relationship: High-quality contact (defined in Allport’s terms) was more strongly associated with positive aVect toward foreigners for respondents for whom the relationship with the contact person was marked by high (rather than low) salience. In addition, Gonza´lez and Brown found some direct evidence of the power of salience to facilitate generalization from the individual to the group. Respondents were dichotomized into high- and low-salience groups, and the correlations between the ratings of individual foreigners and of their countries of provenance were computed for each group. This method provides a measure of stimulus generalization. The individual-country correlations were always higher in the high-salience group, significantly so in three of the four comparisons. For our European respondents, having a relationship with an outgroup person in which their respective nationalities played a prominent role seemed to give rise to a closer correspondence between what they thought of that person and of his/her nationality. A third study also exploited a unique aspect of European intergroup relations, namely the historical rivalry between the English and the French. Brown, Maras, Masser, Vivian, and Hewstone (2001) recruited as

280

RUPERT BROWN AND MILES HEWSTONE

respondents an opportunity sample of British adults (N ¼ 352) on a crosschannel ferry traveling between England and France. We again found a similar, albeit weaker, interaction between salience and contact; for passengers who reported a high level of group salience in their relationship with a French person they knew personally, there was a reliable correlation between amount of contact and a positive social orientation toward France or French people; among low-salience passengers, the same relationship was not significant. These moderating eVects of salience are by no means restricted to university samples or to inter-nation contexts. Gonza´lez, Saiz, Manzi, Ordones, Millar, Sirlopu, and Brown (2003a) asked Chilean school students (N ¼ 1965) for their attitudes toward several minority groups (the Mapuche [a local indigenous group], elderly people, disabled people, evangelicals, and Peruvians). Reliable measures of group salience and intergroup contact were used to predict prejudice and positive aVect toward the various target groups. By the use of structural equation modeling, the results revealed that contact significantly predicted attitude toward the out-group contact person in three of the five groups. In turn, this variable was strongly associated with attitudes (evaluative ratings and aVect) toward the out-group as a whole. However, the latter relationship was again moderated by group salience in two of the target groups (Mapuche and elderly people). Significant, positive, and stronger correlations between evaluative ratings of the person with whom they reported having had contact (e.g., an individual indigenous person) and the general evaluative rating and positive aVect of the out-group targets as a whole (e.g., Mapuche) were observed in the high compared with the low group salience respondents [see also similar findings in a survey of attitudes among Chilean political parties (Gonza´lez, Manzi, Saiz, Brewer, De Tezanos, Torres, et al., 2003b)]. The same moderation eVect was found for yet another target group, the elderly, by Harwood, Hewstone, Paolini, and Voci (in press; Study 1, N ¼ 192). We investigated whether attitudes toward elderly people were aVected by the amount and quality of contact that students had with their grandparents. We found a significant relationship between quality of contact and attitudes toward the elderly only for those who had had relatively frequent contact with their grandparents; for grandparents who were seen less often, there was no relationship. Similarly, for the ‘high frequency’ grandparents, there was a moderating eVect of salience. Those for whom the age category was salient during the contact with their grandparents (e.g., they were aware of age diVerences and saw their grandparents as typical of other older people) showed a robust relationship between contact quality and favorable attitudes toward elderly people. In contrast, the ‘low salience’ respondents showed no such relationship.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

281

We have examined slightly diVerent aspects of moderation in two recent studies conducted in Italy on Italians’ views of immigrants (Voci & Hewstone, 2004a). In the first study (N ¼ 145 Italian workers), we measured quantity and quality of contact with immigrant workers in the workplace, as well as group salience during this contact. We also measured two aspects of group representations, both perceived variability and attitude, each measured at the level of both co-workers and immigrants in general. As expected, quantity and quality of contact were both positively associated with views of co-workers; group salience was negatively associated with attitude toward co-workers; and both variability and attitudes toward coworkers were positively associated with their respective measures at the level of the group as a whole. Moreover, there were four eVects moderated by salience. Quality of contact was more strongly and positively associated with attitude toward co-workers, under high vs. low salience. Although the direct eVect of quality of contact on generalized attitudes toward immigrants was not significant, it was significantly positively associated with overall attitudes under high salience and marginally negatively associated with overall attitudes under low salience. Finally, generalization from the level of co-workers (perceived variability of, and attitudes toward) to the out-group as a whole was stronger under high than low salience (although both eVects were significant). We obtained very similar results in a second, smaller study of the quality of factory workers’ contact with both North African and Albanian coworkers, how typical they were perceived to be of their group, and attitudes toward co-workers and groups as a whole (Voci & Hewstone, 2004a; Study 2, N ¼ 76). Quality of contact was again positively associated with views of co-workers; typicality (like salience in Study 1) was negatively associated with attitude toward co-workers; and attitudes toward co-workers were positively associated with group-level attitudes. There were also two eVects moderated by salience, one for each target group. Quality of contact with North Africans did not have a significant direct eVect on generalized attitudes toward the group, but it was significantly positively associated with attitudes under high typicality and unrelated to overall attitudes under low typicality. Quality of contact with Albanians was significantly positively associated with attitude toward Albanian co-workers under high salience only. Moreover, attitudes toward the two national groups were generalized to attitudes toward the whole category of immigrants. This generalization was stronger for judgments concerning North Africans, who were considered by our respondents as more representative than were Albanians, of immigrants in general. The above studies all employed cross-sectional designs, but we did use a longitudinal design in a recent study on students at diVerent kinds of schools

282

RUPERT BROWN AND MILES HEWSTONE

in the U.K. Brown, Eller, Leeds, and Stace (2004) examined attitudes of state school students in the U.K. (N ¼ 213) toward members of a nearby private—and, hence, higher status—school over a 4-month period. There were three measures of intergroup attitude: desired closeness to out- group members, negative stereotyping of the out-group and, for the first time in contact research, out-group infrahumanization (Leyens, Paladino, Rodriguez, Vaes, Demoulin, Rodriguez, & Gaunt, 2000; the tendency to attribute more secondary, uniquely human emotions to the in- than outgroup). Frequency of contact with members of the other school at time 1 reliably predicted attitudes toward members of that school at time 2 on all three measures, even controlling for initial out-group attitude. Importantly, initial out-group attitude did not predict subsequent out-group contact (when controlling for initial contact), thus implying that the causality in this context ran from contact to intergroup attitude rather than the other way around. Brown et al. (2004) also tested whether salience moderated the eVects of contact. On the measure of negative stereotyping, we obtained a marginal longitudinal moderation eVect for salience: For high-salience respondents at time 1, quality of contact at time 1 was negatively related to stereotyping at time 2, while for low-salience respondents, the relationship was positive. Once again, reversing the causal ordering of the variables yielded nonsignificant findings.

C. SUMMARY OF MODERATIONAL EVIDENCE To summarize, there is consistent evidence from a variety of research settings that both the amount and quality of contact with individual outgroup members have stronger, more beneficial, and more generalizable eVects on intergroup attitudes when the contact person is seen as ‘typical’ of the out-group and/or the respective group memberships are psychologically salient (see Table II). What is most striking about this summary of the studies is that the modal eVect supports the strong form of our intergroup contact hypothesis, namely that the association between contact and outcome is reliable only when categories are salient during intergroup contact. The evidence for the moderation eVect of group salience in our survey research is particularly impressive, given the acknowledged diYculties of detecting interactions and moderator eVects in nonexperimental contexts (McClelland & Judd, 1993). In Section VII.D, we will review still further studies on moderation eVects. In that section, however, we will also show how group salience is related to more aVective reactions, which in turn seem to be implicated as mediators of many contact eVects.

TABLE II Category Salience Moderation Effects Moderation of contact Weak hypothesis High > Low

Study number

Target group(s)

Type of contact

Study 2

European out-group

Quantity

Live in out-group country

Gonza´lez and Brown 2001

European out-group

Quality

Out-group positive aVect

Brown et al., 2001

French

Quantity

Positive orientation to out-group

Gonza´lez et al., 2003

Indigenous group

General

Out-group attitude

yes

Out-group attitude

yes

Study Brown et al., 1999

Elderly

Outcome

to

Outcome path Strong hypothesis High sig./Low n.s. yes

yes yes

Harwood et al., in press

Study 1

Elderly

Quality

Out-group attitude

yes

Voci and Hewstone, 2004a

Study 1

Immigrants

Quality

Out-group attitude

yes

Study 2

North Africans

Quality

Out-group attitude

yes

Albanians

Quality

Attitude towards out-group co-workers

yes

Pupils at rival school

Quality

Negative traits attributed to out-group

Brown et al., 2004

yes

284

RUPERT BROWN AND MILES HEWSTONE

VII. Mediators of the Effects of Contact A. COGNITIVE AND AFFECTIVE MEDIATORS Moderation analyses, of the kind discussed in Section VI, are primarily concerned with the ‘when’ and the ‘where’ of contact eVects: For whom and under what conditions does contact bring about reductions in prejudice? Such studies may not, however, be so informative about the ‘why’ of contact: What are the processes that seem to drive those attitude changes? To answer this kind of question, mediational analyses are more useful. In this section, we review the progress we have made in pursuit of mediating variables within the framework of the original Hewstone–Brown model and its later revisions (Hewstone, 1996; Vivian et al., 1997). One of the major additions to this literature since Allport’s (1954) pioneering work has been the study of mediating variables. Indeed, with all the benefits of hindsight and a discipline that has matured theoretically, empirically, and methodologically, it is striking how little Allport seemed concerned with ‘how’ or ‘why’ contact works eVectively; yet these are questions that relatively new statistical techniques can help us to answer (notably structural equation modeling; e.g., Harwood et al., in press; Paolini et al., 2004; Voci & Hewstone, 2003b). Scholars have suggested several variables that could potentially mediate between contact and outcomes (Dovidio et al., 2003; Kenworthy, Turner, Hewstone, & Voci, 2005; Pettigrew, 1998). These can be classified as cognitive or aVective. The first cognitive mediator dates from Allport’s (1954) original formulation, which focused on contact working by improving knowledge about the out-group. However, enthusiasm for this variable had dwindled, in the light of its rather meager eVects (see Pettigrew & Tropp, 2004; Stephan & Stephan, 1984). In fact, rather than factual information per se being important, more recent research has emphasized the importance of knowing about diVerences between groups (Wolsko, Park, Judd, & Wittenbrink, 2000), which is theoretically much closer to our conception of awareness of group diVerences as a moderator of contact eVects. We do not rule out the importance of knowledge (see its role below); however, research needs a tighter focus on what kinds of knowledge contact might deliver and which ones might have an impact on views of the out-group. A second cognitive mediator we have investigated is the extent to which contact allows for individuation of the out-group member (or members) encountered during contact as an exemplar of the category. Individuation is the acquisition of knowledge about unique attributes of out-group members and was emphasized in Brewer and Miller’s model (see Miller, 2002).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

285

We argue that close relationships are likely to facilitate the acquisition of individuating information and that this, in turn, should result in a more sophisticated view of one out-group member’s experience, which in turn might translate into more sophisticated understandings of the out-group as a whole. Notwithstanding the importance of some cognitive variables, a key change in this literature over the last nearly 20 years has been the acknowledgment that intergroup contact cannot be considered only in terms of its cognitive processes (Johnston & Hewstone, 1992; Pettigrew, 1986, 1998). Indeed, a deeper understanding of intergroup relations in general requires recognition of the role of aVective processes (Fiske, Cuddy, Glick, & Xu, 2002; Mackie, Devos, & Smith, 2000). In current work, aVective factors are now considered to be particularly important (Pettigrew, 1998). AVective processes seem to be more predictive of intergroup behavior than are cognitive variables (Stangor, Sullivan, & Ford, 1991), and they also seem to play a greater role in the contact process than do cognitive factors (Pettigrew, 1998; Tropp & Pettigrew, in press). Pettigrew’s emphasis on aVective factors comes from his conviction that ‘‘the contact situation must provide the participants with the opportunity to become friends’’ (Pettigrew, 1998, p. 76). Friendship, Pettigrew argues, can both reduce negative aVect and augment positive aVect. We turn now to a review of our own cross-sectional studies on variables mediating the relationship between contact and outcomes. Because of the large number of studies, and to preserve some flavor of how this research program developed, we split this section up. We first review theory on intergroup anxiety and then summarize studies that focused exclusively on whether this one negative mediator helped explain eVects of contact. Then we review theory on additional negative as well as positive potential mediators (e.g., whether contact worked by reducing anxiety and promoting positive emotions, empathy, and perspective-taking). Within our reviews of these two classes of mediators, we also group studies according to whether they assessed direct contact (i.e., conventional self-reports of quality and quantity of face-to-face contacts) or extended contact (i.e., normally operationalized as knowing other in-group members who have out-group friends).

B. INTERGROUP ANXIETY AS A MEDIATOR OF INTERGROUP CONTACT EFFECTS In thinking about potential mediators of the eVects of contact, an early candidate was intergroup anxiety. Intergroup anxiety is a negative aVective process that is integral to the contact situation (see Greenland & Brown,

286

RUPERT BROWN AND MILES HEWSTONE

1999) and is experienced when anticipating future, or expecting actual, contact with an out-group member. The primary stimulus for this interest in anxiety was Stephan and Stephan’s (1985) hypothesis that in many intergroup settings, particularly when encountering members of an outgroup for the first time, people are liable to feel somewhat apprehensive, perhaps because of an uncertainty as to the appropriate norms of behaviour due to unfamiliarity, or because of some vestiges of culturally socialized aversion to the out-group in question. According to Stephan and Stephan, intergroup anxiety stems from the expectation of negative consequences for oneself in intergroup interactions, such as embarrassment, rejection, discrimination, or misunderstanding. Antecedents of intergroup anxiety may be minimal previous contact with members of the out-group, negative outgroup stereotypes, a history of intergroup conflict, large status diVerentials, or a high ratio of out-group to in-group members. Whatever its origins, such anxiety is not likely to be conducive to positive intergroup attitudes and behavior. In part, this is because heightened arousal is generally associated with a narrowed cognitive and perceptual focus and an increased reliance on heuristics like stereotypes, but it is also true by simple association, i.e., the negative aVect may transfer to the out-group members seen to be its cause (Bodenhausen, 1993; Stephan & Stephan, 1985; Wilder, 1993; Wilder & Simon, 1996, 2001; see Paolini, Hewstone, Voci, Harwood, & Cairns, in press). Mendes, Blascovich, Lickel, and Hunter (2002) have recently distinguished between three types of anxious or threat responses: physiological (i.e., responses of the autonomic system like sweating and increased heart rate), behavioral (i.e., depleted performance and contact avoidance), and subjective (i.e., self-reported anxiety responses). They reported higher threat responses among participants with limited vs. substantial prior contact with the out-group. Richeson and colleagues (Richeson, Baird, Gordon, Heatherton, Wyland, Trawalter, & Shelton, 2003) also reported, using a sophisticated mix of psychological and neuroscience techniques, that whites who interacted with a black individual (albeit an experimenter, rather than someone with whom they had had previous extensive contact) showed short-term resource depletion due to temporary negative eVects on executive function. Importantly, intergroup anxiety may lead to avoidance of contact and increased polarization of group judgements. Close friendships, however, are associated with reduced anxiety (La Greca & Lopez, 1998). If friendship functions as a stress-buVering mechanism (Cohen, Sherrod, & Clark, 1986), then having out-group friends may reduce negative expectations of interactions with other out-group members (Paolini et al., 2004; see also Tropp, 2003).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

287

1. Research on Intergroup Anxiety a. Direct Contact EVects. Because of the deleterious eVects associated with intergroup anxiety, it became important to establish whether the nature of that contact as ‘interpersonal’ or ‘intergroup’ aVected the level of anxiety experienced. This was the goal of a study by Islam and Hewstone (1993) that investigated inter-religious contact between Muslims and Hindus in Bangladesh, a country with a majority of Muslims (86% of the population) and a minority of Hindus (12%). Hindu and Muslim students (N ¼ 131) gave their estimates of how much contact, and of what type, they had with members of the other religion, and they also indicated whether that contact was more intergroup or interpersonal. Subsequently, they also answered scales measuring intergroup anxiety, out-group variability, and overall attitude toward the religious out-group. Both the quantity and quality of contact were directly positively associated with attitude toward the out-group and were also negatively correlated with anxiety. Importantly, perceiving the contact as ‘intergroup’ rather than ‘individual’ was also associated with greater anxiety. Finally, anxiety was predictive of decreased perceived out-group variability and a less positive attitude toward the out-group (see Fig. 5). Consistent with Stephan and Stephan’s (1985) claim, intergroup anxiety partly mediated the positive relationship between both contact quality and quantity of contact, as predictors, and out-group attitudes and perceived out-group variability, as outcomes. Moreover, it fully mediated the negative relationship between category salience and both outcomes.

Fig. 5. Intergroup anxiety as a mediator of eVects of contact on outcome variables (from Islam & Hewstone, 1993).

288

RUPERT BROWN AND MILES HEWSTONE

Confirmation of Islam and Hewstone’s findings was obtained by Greenland and Brown (1999) in a study of Anglo-Japanese attitudes. British and Japanese students (N ¼ 236), residing in their countries of origin, rated the extent that they were aware of ‘individual personalities’ (interpersonal) and ‘nationalities and culture’ (intergroup) in their contact with members of the other nationality. These two measures were both related to intergroup anxiety—negatively and positively respectively—which, in turn, was predictive of negative aVect and intergroup bias toward the other group. Interestingly, in a small longitudinal follow-up study of Japanese sojourner students in the U.K., there were indications that anxiety predicted intergroup salience rather than vice versa (Greenland & Brown, 1999, Study 2; see also Levin et al., 2003). The results of these two studies were important for two reasons: first, because they were the first intimation of a possible drawback of the Hewstone–Brown proposal that contact should occur at an intergroup level for maximum eVect. They suggested that some balance needed to be struck between reaping the generalization benefits of ‘intergroup’ interactions and the aVective advantages of a more ‘interpersonal’ kind of contact. Actually, as we shall see shortly, heightened group salience does not inevitably have negative side eVects. But a second, and more fundamental, implication of Islam and Hewstone’s (1993) results was that they gave significant pointers as to what might be the key processes that mediate the eVects of contact. We have also obtained evidence for the mediational role of anxiety in several studies conducted in Northern Ireland, investigating Catholic-Protestant relations (see Hewstone et al., in press a; Hewstone, Cairns, Voci, McLernon, Niens, & Noor, 2004; Hewstone, Cairns, Voci, Paolini, McLernon, Crisp, & Niens, 2005, for reviews). This intergroup relationship provides quite a challenge for the Contact Hypothesis given the long and often acrimonious history of religious conflict in this setting. If contact variables can be shown to be eVective here, then there are good grounds for optimism for their applicability in other, less fractious, contexts. In fact, we have found strong and consistent evidence that contact is associated with reduced intergroup anxiety, even in Northern Ireland. We have found this eVect in surveys using Catholics and Protestants as students at the mixed University of Ulster (most schools in the province are segregated) and in representative samples of adults. One of our first studies (which surveyed 440 16- to 18-year olds in 1997–8), sought to predict outgroup evaluation from evaluation of out-group friends, anxiety, general out-group contact (quality and quantity), and importance attached to own and other’s religion (Craig, Cairns, Hewstone, & Voci, 2002; reported in Hewstone et al., 2005; see Fig. 6). As Fig. 6 shows, importance of religion was negatively associated with contact and positively associated with anxiety. Next, contact was positively associated with evaluation of out-group

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

289

Fig. 6. Estimated model of the mediating role of anxiety in the relation between contact and out-group evaluations. Note. Numbers are standardized parameters. Goodness-of-fit indexes: 2 (25) ¼ 40.94, p ¼ .023; RMSEA ¼ 0.038; SRMR ¼ 0.024; CFI ¼ 0.99; N ¼ 440. ***p < .001 (Hewstone et al., 2005).

friends and negatively associated with anxiety. Anxiety itself was negatively associated with the out-group in general, both directly and indirectly via evaluation of out-group friends. Finally, we have conducted two studies of heterosexual students’ contact with and attitudes toward gays (Vonofakou, Hewstone, & Voci, 2003). Our first study (N ¼ 73) measured friendship contact, intergroup anxiety, and several detailed measures of attitude, as well as action tendencies. We tried to predict not only overall attitude but also attitude accessibility (Fazio, 1990) and attitude strength (Krosnick, Boninger, Chuang, Berent, & Carnot, 1993). We predicted that direct contact would be more highly associated with more accessible and stronger out-group attitudes. This is exactly what we found. Cross-group friends had direct, positive eVects on accessibility, attitude strength, and action tendencies. In addition, contact had an indirect eVect, mediated by intergroup anxiety, on all four outcome measures. b. Extended Contact EVects. In its classic formulation, the Contact Hypothesis proposed that direct contact with members of disliked outgroups could, under the right conditions, promote improved intergroup attitudes. All the work reviewed thus far has concerned itself with assessing the eVects of such direct contact. But, as Wright, Aron, Mclaughlin-Volpe,

290

RUPERT BROWN AND MILES HEWSTONE

and Ropp (1997) have pointed out, this may not be the only way that contact has its eVects. It is possible that the mere knowledge that fellow in-group members enjoy friendships with out-group members may be suYcient also to bring about positive attitude change, even in the absence of any direct contact. This could occur because observing such relationships might instigate a change in the perceived in-group norms as to what kinds of intergroup relationships are permissible or desirable. Furthermore, observing out-group members holding friendships with in-group members could begin to modify stereotypes of that out-group since they may be seen as counterstereotypical exemplars. And, finally, through spontaneous empathy with those in-group members—what Wright et al. (1997) call ‘inclusion of the other in the self’ (Aron & Aron, 1996)—it is possible that the ‘observer’ will experience some of the positive aVect associated with friendships, and this aVect is then directed toward the observed, but not actually encountered, out-group members. Wright et al. also noted, that since one is observing the contact indirectly rather than experiencing it directly, there is less risk of the heightened intergroup anxiety that Stephan and Stephan (1985) warned about. Crucially for the present argument, Wright et al. (1997) speculated that these socalled ‘extended contact’ (or indirect contact) eVects may be most potent when and if the group memberships of the protagonists are salient in the mind of the observer (i.e., a moderation eVect) and because the other is included in the self and less anxiety is experienced when observing extended contact than participating in direct contact (i.e., a mediation eVect). Evidence supportive of this ‘extended contact’ hypothesis is beginning to emerge (Liebkind & McAlister, 1999; Wright et al., 1997). Some of our own work in this direction is now outlined. In the first study, Paolini et al. (2004; Study 1) surveyed Catholic and Protestant students in Northern Ireland (N ¼ 341). Respondents were asked first how many close friends they had belonging to the other religious community and also how many friends of their own community had friends from the other community. These were the measures of direct and extended friendship, respectively. The criterion variables were prejudice toward the religious out-group and a measure of perceived out-group variability. Intergroup anxiety served as the potential mediator. As predicted, both direct and extended contact were associated with lower levels of prejudice and with increased perceived out-group variability. Moreover, both contact variables were also substantially and negatively correlated with intergroup anxiety, which itself was negatively associated with prejudice and positively associated with out-group variability. Further analyses revealed that anxiety completely mediated the relationship between direct contact and perceived variability and between extended contact and prejudice, and partially mediated the direct contact—prejudice and extended contact—variability links (see Fig. 7).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

291

Fig. 7. Estimated model of direct and indirect cross-group friendship for university student sample. Values are standardized beta weights. 2 (27, N ¼ 341) ¼ 49.48, p > .005, CFI ¼ .99, RMSEA ¼ .049; SRMR ¼ .023. *p < .05. **p < .01. ***p < .001. (Paolini et al., 2004; Study 1).

A follow-up study with a larger (N ¼ 735) and representative sample was then conducted (Paolini et al., 2004; Study 2) using similar, if slightly simplified, measures. As before, both direct and extended contact were associated with lessened prejudice, with increased perceived out-group variability, and reduced intergroup anxiety. Consistent with the hypothesized role of anxiety in the contact-attitude relationship, anxiety was found to mediate the eVects of direct contact on prejudice (partially) and perceived variability (totally) and to partially mediate the relationship between extended contact and those same two criterion variables. An important point to note about both of these studies is that the measures of contact they employed were derived from estimates of the number of friendships with members of the out-group that the respondents enjoyed, whether directly or indirectly. This underlines the significance of the ‘acquaintanceship potential’ factor in Allport’s (1954) original hypothesis, suggesting that one of its key functions may be to lessen intergroup anxiety. Superficially, the powerful role played by friendships in this study might also be seen to support the Brewer–Miller decategorization–personalization hypothesis. However, the absence of any measures of category salience in these studies means that we cannot be definitive about how those friendships were seen by the participants. In any case, as we shall argue later, close interpersonal relationships may sometimes coexist with moderate levels of intergroup salience.

292

RUPERT BROWN AND MILES HEWSTONE

C. ADDITIONAL NEGATIVE AND POSITIVE MEDIATORS OF CONTACT EFFECTS Whereas intergroup anxiety is certainly an important factor in explaining the eVect of contact on intergroup attitudes, recent work has acknowledged two additional negative mediators. First, Stephan and colleagues emphasized the importance of perceived threats to the in-group as predictors of prejudice (Stephan & Renfro, 2003; Stephan & Stephan, 2000; Stephan, Boniecki, Ybarra, Bettencourt, Ervin, Jackson, McNatt, & Renfro, 2002). They distinguish realistic threats (e.g., threats to the in-group’s political and economic power) from symbolic threats (e.g., threats to the in-group’s value system, belief system, or worldview) as proximal predictors of prejudice. Available studies underscore the potential role of contact in ameliorating perceived threats and their mediating role in the relationship between contact and attitudes (see Stephan & Stephan, 2000). Second, recent research has gone beyond the focus on one negative emotion (intergroup anxiety) to the recognition that there are multiple (negative and positive) potentially relevant intergroup emotions. Thus, researchers have shifted their interest from general evaluations to specific emotions felt toward an out-group (Mackie & Smith, 2002; Smith, 1993). According to Mackie and Smith, emotions like fear, anger, and disgust are related to specific action tendencies such as flight, fight, and avoidance; their distinction thus allows for a better prediction of actual behavior. However, this exclusive focus on negative aVect is still limited. Pettigrew (1997) has usefully suggested the value of promoting positive intergroup aVect, especially via cross-group friendships. A key positive aVect in this context is empathy, as Allport (1954, p. 436) originally noted. Batson et al. (1997) have shown that empathy is closely associated with perspective-taking. Taking the perspective of a stigmatized person results in a greater understanding of the eVects of prejudice (Coke, Batson, & McDavis, 1978), and empathy with a member of a stigmatized group reduces bias against the group as a whole (Batson et al., 1997; Finlay & Stephan, 2000; Galinsky & Moskowitz, 2000). Empathy is generally defined as a vicarious emotional state triggered by witnessing the emotional state of another. Empathy involves imagining how another person perceives their situation and how they might feel as a result. Perspective-taking and empathy have a number of positive consequences for intergroup relations (see Galinsky & Moskowitz, 2000, for a review). By inducing a merging of, or a perception of increased overlap between, the self and the other (see Aron, McLaughlin-Volpe, Mashek, Lewandowski, Wright, & Aron, in press), perspective-taking makes thoughts about the target individual become more self-like (Galinsky & Moskowitz, 2000).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

293

This increased self-target overlap means that a greater percentage of traits used to describe the self will be attributed to the target individual. Because one of the main causes of in-group bias is thought to be the association of the in-group with the self (Cadinu & Rothbart, 1996), attributing traits associated with the self to the out-group target is likely to lead to a more positive evaluation of her or him, which may then generalize to the outgroup as a whole. Perspective-taking and empathy can operate through other routes as well. They can increase the perception that a common humanity and destiny is shared with the other group, which (as proposed by the Common In-group Identity model, reviewed earlier) can reduce prejudice and discrimination. Empathy is also likely to evoke altruistic motivation to relieve any diYculties experienced by the person for whom empathy is felt (Batson, Early, & Salvarani, 1997), as well as motivations to restore justice in response to discrimination (Dovidio, ten Vergert, Stewart, Gaertner, Johnson, Esses, Riek, & Pearson, 2004). Pettigrew (1997, 1998) also identified self-disclosure as a central process in cross-group friendship (see also Dovidio, Gaertner, Validzic, Matoka, Johnson, & Frazier, 1997a, who showed that a common in-group identity was associated with less biased self-disclosure to in- vs. out-group members). Self-disclosure is the presentation of significant aspects of oneself to another person and is important in the development of interpersonal relationships; it may also contribute toward more positive attitudes in an intergroup situation. In addition to reducing anxiety for the recipients of disclosures, selfdisclosure serves to give the disclosers control of how others see them (Berger & Bradac, 1982). By self-disclosing, disclosers tell others how to understand the way they see themselves, or how to empathize with them. Self-disclosure also promotes relational intimacy and depth (Brewer & Gaertner, 2001; Laurenceau, Barrett, & Pietromonaco, 1998; Reis & Shaver, 1988), which may result in more positive aVect toward the out-group. By personalizing an interaction, self-disclosure also focuses attention on the individuating features of those involved, which may reduce the use of stereotypes in a contact situation (Fiske & Neuberg, 1990). Central to the notion of self-disclosure as a mediator is the idea that it is a mode of communication that establishes mutual trust and detailed knowledge about the other party that may disconfirm negative attitudes (leading to personalization, Miller, 2002; see Alicke, Klotz, Breitenbecher, Yurak, & Vredenburg, 1995). Logically, more positive intergroup relations can result from both the reduction of negative aVective processes assumed to be operative in intergroup relations and encounters, as well as the induction of positive aVect that leads to greater liking of the out-group. As we shall see, however, research on mediating factors in contact began with an emphasis on

294

RUPERT BROWN AND MILES HEWSTONE

reducing the negative and has only recently turned to accentuating the positive as well. It is important to consider positive and negative aVective processes separately because they are not necessarily negatively related (Cacioppo & Berntson, 2001). Moreover, although cognitive and aVective processes are conceptually distinct, they frequently operate in concert. 1. Mediation of Contact EVects by Additional Negative and Positive Variables a. Direct Contact EVects. We carried out our first studies including both negative and positive potential mediators in Northern Ireland. Our first study (Voci, Hewstone, Cairns, & McLernon, 2001), using a representative sample of adults (N ¼ 936), measured opportunities for contact, contact and two mediators (negative: intergroup anxiety; positive: perspective-taking) to predict three outcomes (prejudice, out-group trust, and forgiveness). Opportunities positively predicted contact with friends, which itself had positive, direct eVects on all three outcomes. In addition, contact aVected all three outcomes indirectly, via both reduced anxiety and increased perspectivetaking. Thus the positive mediator, perspective-taking, made a unique contribution to the prediction of prejudice, trust, and forgiveness (being only weakly correlated with anxiety). In this study, we were also able to test for both potential causal eVects, from contact to outcomes, and vice versa (see Pettigrew, 1997). The results could not have been clearer; all causal paths from contact to outcomes were significant, but none of the reverse paths were. Our second study (Tam, Hewstone, Kenworthy, Voci, Cairns, & Geddes, 2003; Study 1) moved away from intergroup anxiety to measure intergroup emotions more generally, including both negative and positive emotions (using items based on Mackie et al., 2000), and empathy (both aVective empathy and cognitive empathy, or perspective-taking; after Davis, 1994; Batson et al., 1997). In Study 1 (N ¼ 83 college students in Northern Ireland), we followed Smith’s (1993) intergroup appraisal theory and sought to predict positive and negative action tendencies from prior emotions and contact. We found that cross-group friendships were a negative predictor of negative emotions and a positive predictor of both empathy and positive emotions. In turn, negative emotions predicted negative action tendencies, positive emotions predicted positive action tendencies, and empathy predicted both types of action tendencies (negatively and positively, respectively). Thus, again, the addition of positive mediators (positive emotions, empathy) improved prediction based only on negative mediators (which included, but went beyond, intergroup anxiety). We have also included both negative and positive mediators in a program of research on intergroup contact between Hindus and Muslims in India (Tausch, Hewstone, Singh, Ghosh, & Biswas, 2004). We collected data from

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

295

both religious groups in two cities (Ns ¼ 438 and 304), one which had been relatively peaceful, and one with a history of communal riots. We tested a model with three predictors (quality and quantity of contact, and conflict experience), four mediators (intergroup anxiety, realistic and symbolic threat, and knowledge), and out-group attitude as outcome. To summarize this complex set of results: Contact generally had the hypothesized eVects, being negatively associated with threat and anxiety, although quantity of contact was positively related to realistic threat for Muslims in the riot-prone city. Anxiety and symbolic threat were mediators only for Hindus (in both cities); the mediators for Muslims in the peaceful city were knowledge and realistic threat (no mediators were found for Muslims in the riot-prone city). These findings again indicate the value of adding to a single mediator (e.g., intergroup anxiety) further negative and positive potential mediators. Our final two studies of mediators both considered self-disclosure, alongside other negative and/or positive mediators. Both studies are also unique in adding an implicit measure of attitudes to the self-report or explicit measures used in all other studies to date. The first study (N ¼ 60 white school children, aged 8–11 years) measured cross-group friendships with Asian children as predictor, out-group self-disclosure and intergroup anxiety as mediators, and explicit and implicit attitudes as outcomes (Turner, Hewstone, & Voci, 2004; Study 1). The implicit measure was a word-free form of the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998) specially developed for use with children. Cross-group friendship had a direct, positive eVect on IAT; it also aVected both mediators—self-disclosure positively, and intergroup anxiety negatively. Finally, both mediators aVected the explicit measure of attitude. Self-disclosure had a positive eVect, and anxiety had the usual negative eVect. Thus, adding a positive mediator (self-disclosure) to anxiety increased our ability to predict the explicit measure. To our knowledge, this is the first correlational study showing that self-disclosure mediates the eVect of contact, as predicted by Pettigrew (1997, 1998). [Ensari & Miller (2002) had earlier manipulated self-disclosure experimentally and shown that subjectively experienced self-disclosure aVected attitudes toward new out-group members.] The next study (Tam, Hewstone, Harwood, & Voci, 2004) considered selfdisclosure, anxiety, and empathy as mediators of intergroup contact with the elderly and young people’s explicit and implicit (IAT) attitudes toward the elderly (N ¼ 77 students). In this study, we measured all three mediators at the level of one out-group exemplar, namely the grandparent with whom participants interacted most regularly. Given the importance of selfdisclosure in increasing closeness and intimacy, we predicted that in this context self-disclosure to the grandparent might be associated not only with reduced anxiety but also with increased empathy toward out-group members. Our model showed that quantity of contact with older people

296

RUPERT BROWN AND MILES HEWSTONE

had a direct, positive eVect on implicit attitudes and also positively predicted self-disclosure. In turn, self-disclosure negatively predicted anxiety and positively predicted empathy; and anxiety negatively predicted explicit attitudes, while empathy positively predicted them. Thus, once again, this subset of negative and positive mediators helped us to predict explicit out-group attitudes. To our knowledge, these are the first two studies to demonstrate eVects of contact on implicit measures which, because they do not require participants to report their attitudes directly, are less likely to be influenced by social desirability than are explicit measures. Implicit measures are also important, because they may better predict spontaneous behaviour than do explicit measures (Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997b). It is interesting that both eVects were direct, unmediated eVects. In other words, it seems as though white children who were more familiar with Asians and young people who were more familiar with elderly people held more positive/less negative implicit associations with them. In line with the mere exposure eVect, familiarity (or environmental associations) breeds liking on an implicit level (Bornstein, 1989; Karpinski & Hilton, 2001). b. Extended Contact EVects. We have also investigated extended contact in two studies of young people’s ethnic prejudice in England (Turner et al., 2004; Studies 2 and 3). The first study questioned Asian and white youth (N ¼ 96) about their opportunities for contact, and both direct and extended out-group friendships, as predictors. Opportunity for contact (Wagner, Hewstone, & Machleit, 1989) is a subjective measure of segregation, which actually correlates well with objective measures (see Hewstone et al., 2004). We also measured out-group self-disclosure and intergroup anxiety as mediators. Finally, we measured implicit and explicit out-group attitudes and perceived out-group variability. For both ethnic groups, opportunity for contact was a direct predictor of implicit attitudes and also of direct out-group friendships. Interestingly, opportunity for contact did not predict extended friendships, supporting the idea that this form of contact is both diVerent from direct contact and less reliant on experience of desegregated environments. Direct friendships predicted self-disclosure; extended friendship was a positive predictor of self-disclosure, a negative predictor of intergroup anxiety, and had a direct, positive eVect on out-group attitudes. Finally, self-disclosure was a positive predictor of out-group variability and out-group attitudes, and intergroup anxiety was a negative predictor of out-group attitudes. We then conducted a follow-up study, using white respondents only (N ¼ 164). Given the importance of self-disclosure in the previous and other studies (see Harwood et al., in press; Tam et al., 2004), this study used a more detailed measure of reciprocal self-disclosure (i.e., self to out-group and out-group members to self). The sample was suYciently large to confirm almost exactly the same model using a latent-variables analysis (the one

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

297

diVerence was that extended friendship no longer had a direct, positive eVect on out-group attitudes). These studies improve our knowledge of extended contact by showing that it aVects outcomes not only via reduced intergroup anxiety (as in Paolini et al., 2004) but also via increased out-group selfdisclosure. In addition, consistent with the earlier studies of the eVect of contact on implicit measures (Tam et al., 2004; Turner et al., 2004, Study 1), eVects of contact on implicit measures were unmediated and appeared due to environmental associations (once opportunity for contact was included in the model this, not cross-group friends, was the predictor). We again tested mediational eVects of both direct and extended contact in an extension of Paolini et al. (2004), conducted in Northern Ireland. Tam et al. (2003, Study 2, N ¼ 180) measured opportunities for contact and both direct and extended contact as predictors. In this study, the direct-contact measure was based on cross-group friendships, but the extendedcontact measure was new. We included multiple items tapping various ingroup members (neighbors, friends, and close friends at both home and college) known to have cross-group friends. We also added a measure of the ‘importance’ of contact (van Dick, Wagner, Pettigrew, Christ, Wolf, & Petzel, 2004) and included as mediators empathic emotions, perspectivetaking, positive emotions, and negative emotions. Outcome measures were positive and negative action tendencies and out-group attitude. Opportunities for contact predicted cross-group friends which, together with extended contact, predicted importance of contact. As van Dick et al. (2004) had argued, importance (which may be a proxy for ‘intimacy of contact’) mediated between cross-group friendships and all four aVective mediators (negatively for negative emotions). In the final step of the model each outcome had multiple significant predictors. Empathic emotions, perspective-taking, and positive emotions predicted positive action tendencies; empathic emotions, perspective-taking, and negative emotions predicted out-group attitude; and perspective-taking (negatively) and negative emotions predicted negative action tendencies. In this study too we were able to test both potential causal eVects (see Pettigrew, 1997). Once again, paths from contact to variables were significant or marginal (negative emotions, attitude), but reverse paths were not. 2. Summary Table III summarizes the findings across a range of studies using aVective mediators to test how contact (quantity, quality, direct and extended friends) has its eVects on a range of outcomes. The most impressive evidence has been accrued for the mediating role of intergroup anxiety, contact aVecting a range of outcomes by reducing anxiety. There is now also a solid base of studies showing that contact works too via positive aVective mediators,

TABLE III Mediators of Contact Effects Mediators

Study

Study number

Target group(s)

Type of contact

Outcome

Intergroup anxiety

Threat

Negative emotions

Positive emotions

Empathy

Perspective taking

Direct contact Islam and Hewstone, 1993

Hindu and Muslim

Quality

Quantity

Greenland and Brown, 1999

Craig et al., 2002

Voci et al., 2001

British/ Japanese

Catholics and Protestants

Catholics and Protestants

General

General

Friends

Out-group attitude

Yes

Perceived variability

Yes

Out-group attitude

Yes

Perceived variability

Yes

Negative out-group aVect

Yes

Intergroup bias

Yes

Out-group evaluation

Yes

Evaluation of out-group friends

Yes

Out-group evaluation

Yes

Forgiveness

Yes

Trust

Yes

Yes

Yes Yes

Selfdiscosure

Tam et al., 2003

Study 1

Study 2

Catholics and Protestants

Catholics and Protestants

Friends

Friends (direct and extended)

Yes

Positive action tendencies Negative action tendencies

Yes

Yes

Out-group attitude

Yes

Yes

Yes

Yes

Yes

Yes

Positive action tendencies Negative action tendencies Paolini et al., 2004

Study 1

Study 2

Tam et al., 2004

Catholics and Protestants

Catholics and Protestants

Elderly

Friends

Yes

Out-group prejudice

Yes

Perceived variability

Yes

Out-group prejudice

Yes

Perceived variability

Yes

Quality  quantity

Out-group attitudes

Yes

Quality and quantity

Out-group attitude

Yes

Quality and quantity

Out-group attitude

Quality and quantity

Out-group attitude

Friends

Yes

Yes

Yes

Yes

Tausch et al., 2004 Peaceful city Hindus

Hindus and Muslims Yes (symbolic)

Peaceful city Muslims

Yes (realistic)

Riot-prone city Hindus

Yes

Yes (symbolic)

(continued )

TABLE III (continued) Mediators

Study

Study number

Target group(s)

Type of contact

Outcome

Intergroup anxiety

Threat

Negative emotions

Positive emotions

Empathy

Perspective taking

Selfdiscosure

Riot-prone city Muslims Turner et al., 2004

Out-group attitude

Study 1

Asian

Friends

Out-group attitude

Study 2

Asian and White

Friends

Out-group attitude

Yes

Perceived variability

Yes

Out-group attitude

Yes

Perceived variability

Yes

Study 3

Vonofakou et al., 2003

Quality and quantity

Study 1

Asian

Gay men

Friends

Friends

Yes

Attitude

Yes

Accessibility

Yes

Attitude strength

Yes

Action tendencies

Yes

Prejudice

Yes

Extended contact Paolini et al., 2004

Study 1

Catholics and Protestants

Extended friends

Yes

Study 2

Turner et al., 2004

Study 2

Catholics and Protestants

Asian and White

Extended friends

Extended friends

Perceived variability

Yes

Prejudice

Yes

Perceived variability

Yes

Out-group attitude

Yes

Yes

Perceived variability Study 3

Asian

Extended friends

Out-group attitude Perceived variability

Yes

Yes

Yes Yes

302

RUPERT BROWN AND MILES HEWSTONE

notably empathy/perspective-taking and positive out-group self-disclosure. These findings provide strong support for Pettigrew’s (1998) emphasis on aVective mediators.

D. SIMULTANEOUS MODERATION AND MEDIATION Thus far, we have discussed separately evidence for moderation and mediation eVects. Thanks to recent developments in statistics and structural equation modeling (Jo¨reskog & So¨rbom, 1996; Wegener & Fabrigar, 2000), it is now possible to analyze whether group salience during contact moderates any of the eVects involving contact, mediators or outcomes—that is, do variations in the moderator aVect the relation between a predictor and a mediator or between a mediator and an outcome? Thus far, very few studies have examined this, but their results are rather consistent (for methodological details, see Voci and Hewstone, 2003b). 1. Correlational Evidence of Moderated Mediation EVects Voci and Hewstone (2003b) investigated the mediating role of anxiety in two studies of the eVect of contact on Italians’ attitudes towards immigrants in Italy. We created an index of contact by multiplying quantity by quality, consistent with the moderational idea underlying Allport’s (1954) thesis that more quantity is valuable only if it has positive quality. The first study, involving Italian students (N ¼ 310), revealed positive, direct eVects of contact on both out-group variability and out-group attitude, and a direct, negative eVect on subtle prejudice (Pettigrew & Meertens, 1995; see Fig. 8). There was also an eVect of contact, mediated by anxiety; contact negatively predicted anxiety, which negatively predicted out-group attitude and positively predicted subtle prejudice. Echoing results presented in Section VI, this study also found two instances in which salience moderated the eVects of contact. First, the eVect of contact on favorable attitudes toward immigrants was significantly higher for those reporting high than low intergroup salience. Second, salience moderated the negative eVect of contact on intergroup anxiety. The relationship was stronger for those reporting high than low salience. In a second study, this time with hospital workers as participants (N ¼ 94), we measured contact with immigrant workers in the hospital, intergroup anxiety at work, and group salience during contact. We found that contact had direct, positive eVects on both attitudes to immigrants as a group and a more policy-oriented measure of endorsement of rights for immigrants (see Fig. 9). Contact was also positively associated with attitudes toward

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

303

Fig. 8. Path model of the eVects of contact on judgments concerning immigrants. Note. *p < .05; **p < .01; ***p < .001; N ¼ 310. Only significant paths are reported. (Voci & Hewstone, 2003b; Study 1).

Fig. 9. Path model of the eVects of contact on judgments concerning immigrants. Note. *p < .05; **p < .01; ***p < .001; N ¼ 94. Only significant paths are reported. (Voci & Hewstone, 2003b; Study 2).

304

RUPERT BROWN AND MILES HEWSTONE

co-workers and negatively associated with intergroup anxiety at work. In turn, anxiety negatively predicted attitude toward immigrant co-workers, which positively predicted both outcome measures. Finally, three eVects were moderated by salience. Contact was reliably negatively associated with anxiety only under high salience. In addition, attitudes toward co-workers were more strongly related to attitudes toward the out-group as a whole under high than low salience, and co-worker attitudes generalized only to policy endorsements under high salience. Harwood et al. (in press; Study 2) included the most detailed set of potential mediators of the eVect of contact with the most regularly seen grandparent on both attitudes toward and perceived variability of the elderly (N ¼ 100 students). We included five possible mediators: intergroup anxiety, perspective-taking, individuation, self-disclosure, and accommodation. We have introduced all these mediators theoretically in an earlier section, with the exception of accommodation, a measure of the degree to which young respondents adapted communicatively to their grandparents. Accommodation is a crucial signal of interpersonal solidarity, and the absence of accommodation is often a signal of intergroup diVerentiation (Shepard, Giles, & LePoire, 2001). This study measured mediators at the ‘individual’ level (as in Tam et al., 2004) rather than at the ‘group’ level (as in most studies). This means that, rather than looking at mediators that are relevant to people’s reactions to out-group members in general or in hypothetical scenarios (e.g., the use of intergroup anxiety in Islam & Hewstone, 1993), mediators in this study were tapped at the level of the specific relationship with an out-group member and concerned interaction with the same grandparent with whom we assessed quality of contact. We believe that both types of mediators are necessary for a complete understanding of the mechanisms behind contact eVects (Paolini et al., in press). If quality of contact with a specific out-group individual influences more general out-group attitudes, then there may be very specific interactional experiences that aVect group-level mediators and serve as mechanisms for such influence. Experiencing a specific aVect, cognition, or behavior in interaction with a particular out-group member (i.e., individual level mediator) makes it more likely that such a phenomenon might occur or be seen as possible with other out-group members (i.e., group-level mediator). This perception may generalize to expectations for intergroup contact more generally, and hence the nature of the out-group as a whole. When examined separately, three of these variables proved to be reliable and full mediators of the eVects of contact quality on attitudes: anxiety, perspective-taking, and accommodation (see Table IV). However, it was the two other variables (self-disclosure and individuation) that, when entered separately, partially mediated perceived variability. When all significant

305

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

TABLE IV Mediation in the Relation Between Contact Quality and Dependent Measures (Top Panel, Attitudes; Bottom Panel, Perceived Variability) (From Harwood et al., In press; Study 2)

Mediator

Contactmediator

Mediatorattitude

Contactattitude via mediator

R2

Goodman test

Perspective taking

.48 (.07)***

.58 (.14)***

.08 (.12)

.24

3.55***

Anxiety

.64 (.09)***

.41 (.11)***

.10 (.12)

.22

3.28**

Individuation

.32 (.08)***

.09 (.14)

.33 (.11)**

.11

0.62

Disclosure

.64 (.08)***

.11 (.10)

.29 (.14)*

.11

1.08

Accommodation

.21 (.04)***

.74 (.23)

.20 (.11)

.19

2.71**

Contactmediator

Mediatorvariability

Perspective taking

.48 (.07)***

2.89 (2.59)

9.55 (2.19)***

.28

1.10

Anxiety

.64 (.09)***

1.83 (1.99)

9.76 (2.21)***

.28

0.90

Individuation

.32 (.08)***

8.26 (2.16)***

8.29 (1.82)***

.37

2.72**

Disclosure

.64 (.08)***

5.22 (2.27)*

7.59 (2.28)***

.31

2.21*

Accommodation

.21 (.04)***

1.08 (4.13)

10.71 (2.01)***

.27

0.26

Mediator

Contactvariability via mediator

R2

Goodman test

Note. Scores are unstandardized regression coeYcients, standard error in parentheses. For the relation between contact quality and attitude without mediation: b ¼ 0.36 (.11)***, R2 ¼ .11. For the relation between contact and variability without mediation: b ¼ 10.94 (1.80)***, R2 ¼ .27. *p < .05. **p < .01. ***p < .001.

mediators were entered together, perspective-taking was the only significant mediator of contact on attitudes, and individuation was the only significant mediator of contact on variability; both eVects were positive. The mediation eVects for out-group attitudes (but not perceived variability) were also moderated by group salience, holding only for respondents for whom the ‘young-elderly’ relationship was above the average in salience. These moderated mediation eVects may occur because the moderator aVects the relation between the predictor and the mediator, or between the mediator and the criterion variable (Wegener & Fabrigar, 2000). To clarify this point, we examined the presence of moderational eVects in the links between contact and each mediator and between each mediator and the criterion variables. Results showed that the paths aVected by group salience were the ones between contact quality and mediators. Contact quality aVected

306

RUPERT BROWN AND MILES HEWSTONE

perspective-taking more when salience was high than when salience was low; anxiety was reduced by contact more when group salience was high than when salience was low; and the link between contact and accommodation was significant only when salience was high. Empathy was also a central mediator variable in three recent studies on Italians’ attitudes towards immigrants (Voci & Hewstone, 2003a). In the first (N ¼ 169 Italian adults), we created an index of contact by multiplying quantity by quality. In addition to contact as a predictor, we measured anxiety and empathy as mediators, and out-group attitudes and subtle prejudice (Pettigrew & Meertens, 1995) as outcomes. We found direct eVects of contact on both out-group attitude (positive eVect) and subtle prejudice (negative eVect). Contact also aVected both outcomes via the mediators. Contact negatively predicted anxiety, which negatively predicted out-group attitude and positively predicted subtle prejudice. Contact also positively predicted empathy, which positively predicted out-group attitude and negatively predicted subtle prejudice. We then tested for moderations of the mediational eVects and found three such eVects. First, salience moderated the direct eVect of contact on subtle prejudice; the negative eVect of contact on prejudice was significant only for those reporting high versus low salience. Then, two moderated mediations were detected: Contact increased empathy and reduced anxiety only when group salience was high. The second study (N ¼ 172 Italian adults) distinguished ‘intimate’ from more ‘generic, social contact’ and tested whether their eVects on attitudes and subtle prejudice were mediated by empathy. Social contact had a direct negative eVect on subtle prejudice and a direct positive eVect on out-group attitudes. Both types of contact also aVected both outcomes via positive eVects on empathy. Empathy negatively predicted prejudice and positively predicted out-group attitude. In this study, salience moderated the eVects of intimate contact on empathy (the positive association was significant only under high salience) and general contact on subtle prejudice (the negative association was significant only under high salience). The third study (N ¼ 143 Italian factory workers) distinguished ‘workrelated’ from more ‘generic, social contact’ and tested whether empathy mediated their eVects on out-group attitudes and a measure of perceived crime by immigrants. Generic contact had a direct negative eVect on perceived crime and a direct positive eVect on attitudes. Both types of contact also aVected both outcomes via positive eVects on empathy. Empathy negatively predicted perceived crime and positively predicted out-group attitude. Salience moderated the eVect of work contact on empathy (the positive association was significant only under high salience) and general contact on perceived crime (the negative association was significant only under high salience).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

307

Two of our most recent studies in Northern Ireland have also included both moderators and mediators (Hewstone, Voci, Cairns, & McLernon, 2003). The first study used a representative sample of adults in Northern Ireland (N ¼ 970) and measured five variables. First, we assessed self-reports of opportunity for contact (Wagner et al., 1989). Next, we assessed contact (including quantity and quality, e.g., out-group friends) intergroup anxiety, salience, and out-group evaluation. As expected, opportunity for contact predicted actual contact, which then aVected out-group evaluation directly, and indirectly, via reduced anxiety. The path from contact to out-group evaluation was moderated by salience; the association between contact and evaluation was significantly positive under high salience but only marginally so under low salience. The second study followed up the idea that contact that is ‘high’ on both interpersonal and intergroup dimensions should be most eVective. We tested a sample of students at the University of Ulster (N ¼ 343) and measured six variables: group identification (how strongly respondents identified with their religious group membership), contact with out-group friends, intergroup anxiety, salience, out-group evaluation, and forgiveness. The last measure was specially developed for this research and consisted of a multiitem measure indicating willingness to forgive the other group for its part in the conflict (for more details see Hewstone et al., 2004). There was a direct, negative path from group identification to forgiveness, and identification was also negatively associated with contact with out-group friends. In turn, contact with out-group friends was positively associated with both out-group evaluation and forgiveness and negatively associated with anxiety. In the last step, intergroup anxiety negatively mediated the eVects of contact with out-group friends on both outcomes. Finally, this data set yielded two moderation eVects involving salience. First, the association between contact with out-group friends and out-group evaluation was significantly positive only under high salience. Second, the association between contact with out-group friends and forgiveness was also significantly positive only under high salience. Contact with out-group friends was associated with more positive out-group evaluation and greater forgiveness only for respondents who reported high (vs. low) awareness of group memberships during contact. Our final study testing mediation and moderation assessed heterosexual students’ views of gay men (N ¼ 160; Vonofakou et al., 2003; Study 2). As in the first study with this target group, which assessed mediation only, we found that cross-group friends directly predicted both attitude accessibility and attitude strength. But it also had indirect eVects on all four outcomes, mediated by closeness of the out-group friendship and intergroup anxiety. There was also one moderation eVect. Closeness of the out-group friendship

308

RUPERT BROWN AND MILES HEWSTONE

predicted lower intergroup anxiety only when that friend was perceived as highly typical. 2. Summary Table V summarizes the findings across the nine studies that have tested for moderated mediation of contact eVects. There is considerable evidence for moderation of the paths from both contact-to-mediators and contact-tooutcomes. Although few of these studies have included it, there is also evidence of moderation for the generalization eVect, from out-group targets to the out-group as a whole. Once again, the findings also lend support to the strong version of our hypothesis, with most moderated mediation eVects indicating that the association holds only when categories are salient during contact. E. SUMMARY OF MEDIATIONAL EVIDENCE We have shown in this section that the cumulative evidence in support of mediating processes, especially aVective processes, is just as impressive and plentiful as the evidence on moderation. Moreover, by simultaneously analyzing for moderation and mediation, we have shown that key mediational eVects are moderated by category salience, in line with predictions from our theory. This research shows quite convincingly that both positive and negative aVect play a key role in mediating the eVects of contact on intergroup attitudes. Moreover, these aVective variables mediate the eVects of both direct and extended contact. Certainly, all of the mediating mechanisms that we have discussed above are central to our understanding of the role of contact in intergroup relations. The relative importance of each depends on the given situation, the groups involved (see, for example, Stephan & Stephan, 2000, on the diVerential role of symbolic and realistic threats for majority and minority group members), and on the outcomes (see Harwood et al., in press; Study 2). From the empirical evidence presented above, it seems that ‘hot’ processes are very important in intergroup conflict. The diVerent mediating processes, however, have not been tested against each other in diVerent settings of conflict. There is some evidence that diVerent mediating mechanisms make unique contributions in the reduction of intergroup bias (e.g., anxiety reduction and increased perspective-taking; Craig et al., 2002; see also Harwood et al., in press), but more systematic research is needed to predict which mediating mechanisms work under which conditions. Knowing which psychological processes are driving the eVect of contact on attitudes and intergroup relations in a given context could help

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

309

in the design and implementation of optimal interventions that specifically target the most important processes.

VIII. Evidence from Educational Settings We now discuss some research germane to our model conducted in the naturalistic setting of the school classroom. Three features of this work distinguish it from the other research we have discussed thus far and warrant its presentation in a section of its own. First, it was concerned with evaluating the consequences of ongoing educational policy interventions that also just happened to be relevant to the relative merits of the Hewstone–Brown and Brewer–Miller contact models. This provided an opportunity to see how contact models work in practice, relatively ‘uncontaminated’ by the researchers’ presuppositions. Second, it involved much younger participants (6–11 years) than most of the other research we have reviewed and thus provides a glimpse as to the developmental applicability of contact hypotheses. Third, it was concerned with attitudes toward stigmatized groups— children with various kinds of disabilities, refugees—that have been little studied by contact researchers (for exceptions see, Armstrong, Johnson, & Balour, 1981; Desforges et al., 1991). We have briefly presented the first of these studies (Maras & Brown, 1996) in our discussion of general contact eVects (see Section II). The study involved young children (aged around 9 years), some of whom participated in a weekly exchange with a school for children with learning disabilities and some of whom did not. The exchange program lasted 3 months. In addition to the improvement over time in the sociometric preferences of the ‘experimental group’ already discussed, several other results from the study are noteworthy. One of the measures involved a categorization task in which children were asked to sort a number of photographs of unknown children, with and without diVerent kinds of disabilities. These photographs showed children with physical disabilities, learning disabilities, and hearing diYculties (two of each type), and six children without disabilities. At the beginning of the study there was little diVerence between the sorting strategies used in the ‘experimental’ and ‘control’ groups. Both tended to classify simply by gender and disability, lumping all three types of disability stimuli into the same category. But, by the end of the study, although the ‘control’ children’s classifications had changed little, those in the ‘experimental’ group tended to have a more diVerentiated representation of disability, distinguishing between the three subgroups present in the photographic

TABLE V Moderated Mediation of Contact Effects Relationships moderated by category salience

Study Hewstone et al., 2003

Study number

Target group(s)

Type of contact

Study 1

Catholics and Protestants

Quantity and quality

Mediator

Catholics and Protestants

Out-group friends

Contact-tomediator

Out-group evaluation

Yes (b)

Evaluation

Yes (b)

Forgiveness Study 1

Immigrants

Quantity  quality

Generalization: target-to-category

Anxiety Unmediated paths

Voci and Hewstone, 2003b

Contact-tooutcome

Anxiety Unmediated path

Study 2

Outcome

Anxiety

Yes (b) Yes (a)

Out-group attitude

Yes (a)

Perceived variability Subtle prejudice Study 2

Voci and Hewstone, 2003a

Study 1

Immigrants

Immigrants

Quantity  quality

Quality  quantity

Anxiety

Yes (b) Out-group attitude

Yes (a)

Immigrant rights

Yes (b)

Anxiety

Yes (b) Out-group attitude Subtle prejudice

Empathy

Yes (b) Out-group attitude Subtle prejudice

Unmediated path

Out-group attitude Subtle prejudice

Study 2

Immigrants

Intimate

Empathy

Yes (b) Yes (b)

Out-group attitude Subtle prejudice Social

Empathy Out-group attitude Subtle prejudice

Study 3

Immigrants

Work

Empathy

Yes (b) Yes (b)

Out-group attitude Perceived crime Unmediated path Social

Empathy Out-group attitude Perceived crime

Harwood et al., In press

Vonofakou et al., 2003

Study 2

Study 2

Elderly

Gay men

Quality

Friends

Anxiety

Out-group attitude

Yes (b) Yes (a)

Perspective taking

Yes (a)

Accommodation

Yes (b)

Anxiety

Closeness of friendship to anxiety

Yes (b)

Note. (a) Denotes significantly higher association between contact and other variable under High vs. Low salience. (b) Denotes association between contact and other variable only significant under High salience.

312

RUPERT BROWN AND MILES HEWSTONE

stimuli. These cognitive changes were also visible in the children’s evaluations of the capabilities of the disability out-groups along diVerent dimensions. The ‘experimental’ children showed a sharp increase over time in their ratings of the auditory capacity of learning and physically disabled children, but not hearing-impaired children, thus showing an improved and more realistic understanding of diVerent forms of disability. Furthermore, they also increased in their ratings of ‘thinking ability’ for all disabled groups. In contrast, the ‘control’ children, without the benefit of regular cooperative contact with the special school, showed virtually no changes in these evaluations (see Fig. 10). These findings speak to the potential of contact experiences to promote not just changes toward a particular out-group and its associated stereotypic attributes (in this case, children with learning disabilities and ‘thinking ability’) but also to other out-groups (children with physical disabilities or hearing impairments) and along other dimensions (e.g., hearing ability). As Pettigrew (1997) and Vivian et al. (1997) have noted, very little contact research has investigated these kinds of generalization. The more usual type of generalization studied is from individual to group. The importance of these generalization eVects is that they occurred in a contact situation that could plausibly be described as resembling the high-salience conditions advocated by the Hewstone–Brown model.

Fig. 10. Changes in evaluations of children with diVerent kinds of disability over 3 months of an exchange program (from Maras & Brown, 1996).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

313

A second cross-sectional study investigated attitudes toward those same disability groups but by comparing attitudes in four diVerent schools (Maras & Brown, 2000): two where hearing-impaired children were integrated into mainstream classes and two control schools that had no hearing-impaired children. The method of integration in the former two schools diVered, however. In one, labeled ‘categorized contact,’ the children with hearing impairments attended a special unit in its own building within the school grounds and participated in some mainstream classes in a structured way. They were encouraged to use more visible (i.e., body-worn) hearing aids. The nature of hearing impairment was discussed formally with the mainstream children, and they were given information as to how to communicate with their hearing-impaired peers. All these conditions should have served to heighten the salience of the categories in the minds of the mainstream children and hence correspond to the Hewstone–Brown perspective. In the other (‘decategorized’) school, on the other hand, only individual diVerences/similarities were acknowledged; the special unit was not separate, and there was no particular attempt to integrate the disabled children in any structured way; the latter were discouraged from wearing body-worn aids, and there was little discussion of hearing impairment as a disability nor of ways to communicate with hearing-impaired children. These conditions more closely resemble the decategorized contact advocated by Brewer and Miller. How did the children’s attitudes compare in these three types of school, the ‘categorized,’ ‘decategorized,’ and ‘control’? In general, the attitudes in the ‘decategorized’ school were more favorable than those in the ‘categorized’ school, with respect to both the particular group they had contact with (hearing impaired) and other disability groups (learning and physical disability). In fact, they more closely resembled the attitudes of the ‘control’ children who had had little such contact, while those in the ‘categorized’ school showed much sharper (and more negative) diVerences between ‘normal’ and ‘disabled’ children. Figure 11 provides a good illustration of this general trend. Notice that, although all children rate ‘normal’ children more favorably on ‘thinking ability’ than the others, diVerentiation is much sharper in the ‘categorized’ school. In this case, and in stark contrast to the results of Maras and Brown (1996), ‘salient’ intergroup contact seemed to be having negative eVects on the children’s attitudes. Actually, this apparently contrary finding (for the Hewstone–Brown model) is nevertheless instructive. A plausible explanation for those more negative attitudes in the ‘categorized’ school is that the everyday practical constraints of the classroom meant that not all of Allport’s necessary conditions were present (there was little equalstatus cooperation, for example). If so, it is quite likely that contact in such circumstances could have produced unfavorable outcomes that, because of

314

RUPERT BROWN AND MILES HEWSTONE

Fig. 11. Evaluations of children with diVerent kinds of disability in schools where disability category is or is not salient (from Maras & Brown, 2000).

the salience of the disability category, then generalized to the out-group as a whole. Consistent with this explanation, there was greater individual-group generalization in sociometric preferences in the ‘categorized’ contact school than in the ‘decategorized’ school. Still, the finding in this study that group salience can sometimes make matters worse rather than better (if suboptimal contact conditions prevail), as we had originally warned (Hewstone & Brown, 1986), is a useful caution against an overly optimistic view of its applicability. It should also be noted here that Pettigrew and Tropp’s (2004) meta-analysis shows that the eVects of contact tend to be weakest in the context of disability, compared with other intergroup contexts. A more positive outcome of an antiprejudice school intervention based on a combination of Wright et al.’s (1997) extended contact idea and the Hewstone–Brown model has been found in two studies by Cameron, Rutland, and Brown (2004). Both were conducted in primary schools and involved children aged between 5 and 11 years (Ns ¼ 72, 159). The focus of the interventions was to change children’s attitudes toward disabled children (Study 1) or refugees (Study 2). One type of intervention presented the children with fictional stories involving members of the in-group (nondisabled or host-society children) having friendships with an out-group

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

315

member (child with a disability or refugee). In accordance with our thinking outlined in Section VII, the stories emphasized both individual and category attributes and also made the main (out-group) protagonist seem rather typical of his or her group. The other type of intervention was more cognitive in kind. It involved repeated exercises in ‘socio-cognitive skills training,’ modeled on an idea promulgated by Bigler (1995). In these, the children had to practice classifying photographs of in-group and out-group children on multiple dimensions, the aim being to get them to see out-group members in a more diVerentiated and ‘cross-categorized’ fashion. Both types of intervention took place once a week for a 6-week period. A control group of children received no special treatment. In both studies, the extended contact plus group salience treatment elicited more favorable attitudes toward the out-group than either of the control or socio-cognitive conditions (which did not diVer from each other).

IX. Acculturation Orientations and Intergroup Attitudes Running parallel to the research program inspired by the Contact Hypothesis and its derivatives has been a raft of research investigating the social and psychological impact of diVerent acculturation strategies adopted by host society and immigrant groups. A major influence on this work has been Berry’s typology of four diVerent acculturation strategies: integration, assimilation, separation, and marginalization (e.g., Berry, 1984). These orientations are considered to result from the combination of two orthogonal orientations: a desire to maintain (or relinquish) one’s ethnic identity and a desire to engage with and have contact with other groups (or not). As has been pointed out by Dovidio, Kawakami, and Gaertner (2000), with minor modifications these four strategies bear a striking resemblance to the diVerent contact models discussed earlier in the article. Integration corresponds to the dual-identity option (simultaneous salience of subgroup and superordinate categories), assimilation to Gaertner and Dovidio’s common in-group identity model (low subgroup but high superordinate group salience), separation to the original Hewstone–Brown mutual intergroup diVerentiation formulation (high subgroup but low superordinate group salience), and marginalization to the decategorization idea (low group salience altogether). Much research testifies to the fact that minority groups tend to favor integration (Berry, 1997; Van de Vijver, Helms-Lorenz, & Feltzer, 1999; Van Oudenhoven, Prins, & Buunk, 1998), and that this strategy seems to be associated with better acculturation outcomes (Berry, 1997; Liebkind,

316

RUPERT BROWN AND MILES HEWSTONE

2001). Among majority or host-society groups, the picture is less clear. Some have found that they also prefer integration (Piontkowski, Florack, Hoelker, & Obdrzalek, 2000), while others have found a preference for assimilation (Van Oudenhoven et al., 1998). Little of this research has examined the implications of these strategy preferences for mutual intergroup attitudes, although some theoretical work has been undertaken (Bourhis, Moise, Perreault, & Senecal, 1997). In an attempt to fill this gap, we conducted three studies of acculturation attitudes in Europe (PfaVerott & Brown, 2004; Zagefka & Brown, 2002; Zagefka, Brown, Broquard, & Leventoglu, 2002). In the first (Zagefka & Brown, 2002), minority- and majority-school students in Germany (N ¼ 321) were classified into Berry’s four acculturation strategies according to their responses to questions addressing issues of cultural maintenance and desire for contact with the other group. A preponderance of both groups favored the integration strategy (minority, 75%; majority, 61%), with assimilation being endorsed by only 18% in both groups. Marginalization also featured among some (19%) majority-group members but hardly at all in the minority group (2%). The in-group bias displayed by the acculturationpreference groups diVered both among themselves and between minority and majority groups (see Fig. 12). As can be seen, for minority groups those favoring integration showed less bias than any of the other groups; for members of the host-society majority, the pattern was rather diVerent, with respondents who preferred assimilation showing marginally the least bias. Comparable acculturation preferences were observed in Zagefka et al. (2002), but this time focusing solely on majority-group members in Belgium and Turkey. For these respondents, though integration was still the modal

Fig. 12. In-group bias as a function of acculturation orientations of minority and majority group members in Germany (Zagefka & Brown, 2002).

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

317

preference (around 48%), assimilation was also much in evidence as an orientation (Belgium: 45%; Turkey: 23%). In PfaVerott and Brown’s (2004) study, the divergence between strategies preferred by the minority and majority groups was even more marked. The study was conducted in an urban setting in Germany among host-society and minority (mainly Turkish) adolescents. Once again, there was a clear preference (85%) among the latter for the integration option, a choice made by only 45% of the majority group. The other main preference of the host majority was again assimilation (also 45%). The fact that the dual-identity and assimilationist strategies seem to sometimes be diVerentially attractive and eYcacious for minority and majority groups has been observed in other studies also (Dovidio et al., 2000; Gonza´lez & Brown, in press) and presents something of a problem for the formulation of a comprehensive model of intergroup contact. That task is the focus of the next section. We note here, however, that the role of cultural/ historical forces is crucial to an understanding of acculturation attitudes and that the United States, for example, has seen a pendulum swing for specific groups from assimilation to integration to separatism.

X. Toward an Integration In the previous six sections, we have summarized a large body of research—more than 40 independent studies—that has been dedicated to the testing and refinement of our intergroup model of contact, a model we described as ‘‘as yet untested’’ in 1986 (Hewstone & Brown, 1986, p. 36). As we hope has become clear, the central thesis of that model has received substantial support in a wide range of intergroup contexts: Maintaining some group salience in settings that otherwise satisfy Allport’s criteria for favorable contact leads to positive and generalized change in people’s orientations toward out-groups. Moreover, in the course of that research eVort, substantial progress has been made toward identifying the underlying causal mechanisms involved in successful intergroup contact: Several cognitive and aVective factors have been found to mediate the eVects of contact on intergroup outcomes. Finally, a start has been made on exploring the applied implications of the model. In this concluding section, we seek to do two main things. First, we evaluate progress in terms of each of the basic components of the model. We summarize our achievements and the lessons learned to date, and we consider a number of priorities for future research. Second, we compare our intergroup model of contact with the three main alternative perspectives that

318

RUPERT BROWN AND MILES HEWSTONE

have had the most impact on this research area. We attempt an integration, with a view toward making helpful prescriptions for policy, focused on applying our conclusions in educational settings.

A. INTERGROUP CONTACT THEORY: PROGRESS AND PRIORITIES Stripped down to its psychological essentials, our model consists of four components (Hewstone, 1996; Hewstone & Brown, 1986; Vivian et al., 1997): dimensions of contact, group salience, mediators, and generalized outcomes for intergroup relations. We now elaborate on each of those four components in the light of the research we have reviewed earlier, beginning with contact. Early on in the development of our thinking, we took as a ‘given’ Allport’s (1954) four criteria for contact to result in improved intergroup relations—that is, equal status, cooperative contact that had high ‘acquaintance potential’ (in Cook’s [1962] later phrase) and which enjoyed institutional support. It is probably fair to say that our model emphasized the first two and the last of these conditions more than the third, mainly because our thinking diverged from Allport’s rationale for that condition— that it would permit the discovery of interpersonal similarities between members of diVerent groups. Our model instead stressed the utility of the mutual recognition of group diVerences. Over the past decade, we have come to revise our position somewhat. As is apparent from several of the studies reviewed earlier, one of the most potent forms of contact is contact through cross-group friendship, a conception that is obviously close to Allport’s ‘acquaintance’ condition. We now believe that this is important because of its potential to generate positive orientations and emotions like perspective-taking and empathy and to reduce the negative emotion of anxiety, all important mediators of the eVects of contact. The emergence of friendship as a critical source of contact also provides a link to the Brewer–Miller decategorization model, as we indicate below. However, and crucially for both their model and ours, we still argue that it is possible for cross-group friendships to develop in conjunction with awareness and acknowledgment of respective group memberships. 1. Dimensions of Contact In terms of measurement, we advise that most studies should first measure opportunities for contact (such as demographic indices and group proportions) that set constraints for actual intergroup contact (see also Pettigrew & Tropp, 2004; Wagner et al., 1989). Next, researchers should include

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

319

multi-item measures of quality and quantity of direct intergroup contact, including especially cross-group friendships. As Pettigrew and Tropp (2004) reported, having good measures of contact is more closely related to the contact-outcome eVect size than is having good measures of prejudice. There is growing evidence that it is also worth assessing extended contact. In Wright et al.’s (1997) original research, however, they measured only how many in-group members were known who had an out-group friend. We propose that future research should investigate the impact of two relationships: (1) that between the respondent and the other in-group member and (2) that between the other in-group member and the member of the outgroup. Both relationships could vary from relatively superficial to relatively intimate. But both in-group and out-group members could also vary in terms of how typical or representative they are of their group. We would predict that extended contact would be highly eVective when a typical ingroup member had a close relationship with a typical out-group member (see Cameron et al., 2004). A further priority for future research is to move away from a limited, dyadic view of cross-group contact to one that assesses diversity of social networks. As work in sociology and political science has emphasized, ‘weak ties’ (Granovetter, 1973) or ‘bridging capital’ (Putnam, 2000) that link in-group to out-group social networks are particularly important sources of social capital. Relatedly, a broader notion of extended contact would include analysis of the impact on, e.g., racial attitudes of out-group members in positions of authority, such as mayors and (local) politicians (see Hajnal, 2001). A final suggestion regarding extended contact is that, in part because it is associated with lower intergroup anxiety and higher-category salience, it could prove a potent means of preparing participants for intergroup contact. Thus, exposing people to extended contact prior to direct contact might be a particularly powerful prescription to challenge pernicious prejudice. There are two other reasons to be so sanguine about the potential impact of extended contact eVects. First, it appears less dependent than direct contact on opportunities for contact (see Turner et al., 2004) and is therefore especially useful in highly segregated intergroup contexts (e.g., Northern Ireland). Second, as Wright et al. (1997) noted, one person’s out-group friends have the potential to aVect the attitudes of many others who are not direct friends of the out-group member. 2. The Moderational Role of ‘Intergroup’ Contact We have operationalized category salience as awareness of group memberships, awareness of group diVerences, perceived typicality of the outgroup member(s) with whom there is contact, perceived homogeneity of

320

RUPERT BROWN AND MILES HEWSTONE

the out-group, or diVerent combinations of these. The most reliable moderators have proved to be awareness of membership and perceived typicality (e.g., Brown et al., 1999, 2001, 2004a; Gonza´lez & Brown, 2001; Gonza´lez et al. 2003a,b; Voci & Hewstone, 2004a). Theoretically, this makes good sense. People will not easily make a connection between an individual outgroup exemplar and the out-group as a whole unless they are somewhat conscious of their respective group memberships, and that connection should be stronger to the extent that the exemplar is seen as a typical representative of the group. As we have seen, however measured, category salience has consistently proved to moderate the contact-outcome relationship: Participants for whom the groups are salient reliably manifest a stronger association between contact and outcome measures of intergroup relations than do those for whom they are less salient; they also show stronger individual-to-group generalization (Gonza´lez & Brown, 2001; Gonza´lez et al., 2003a,b; Voci & Hewstone, 2003b, Study 2; Voci & Hewstone, 2004). We cannot, however, conclude that category salience is always necessary. Consider, first, the central finding of Pettigrew and Tropp’s (2004) metaanalysis, which does not include category salience as a moderator, that contact systematically aVects prejudice. Either there must have been suYcient salience in the contact situation in most of these studies (which Pettigrew, personal communication, believes to be the case), or salience is not necessary. Results of our own moderational analyses show clearly that category salience does moderate contact eVects in most cases (i.e., at least some eVects in most studies). Moreover, they tend to support the strong form of our hypothesis, whereby the association between contact and mediator or outcome holds only when there is category salience during contact. We have shown significant moderation eVects whether respondents’ selfreports on contact refer to interactions with a single out-group member (e.g., Brown et al., 1999, 2001; Harwood et al., in press; Hewstone et al., 2003) or are averaged across interactions with multiple out-group members in various settings (e.g., Hewstone et al., 2003). Our normal practice has been to construct reliable indices and use these as moderators. This approach is methodologically defensible but cannot inform us of whether moderation is stronger for one (type of) measure than others. Originally, we argued that, to ensure generalization, contact should make ‘‘group aYliations more salient’’ (Hewstone & Brown, 1986, p. 18, emphasis in original). We have subsequently proposed, simply, that group membership must be suYciently salient to ensure generalization but not so salient that it leads to intergroup anxiety or otherwise exacerbates tensions (Hewstone, 1996, p. 333). As our moderated mediation approach emphasizes, participants who are relatively more aware of group memberships

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

321

during contact are, in fact, those most likely to benefit from the cumulative, anxiety-reducing eVect of repeated exposure to the out-group (e.g., Harwood et al., in press, Study 2; Voci & Hewstone, 2003b, Study 1). Intuitively, our sense is that the optimal measure of moderation refers to awareness of group aYliations or perceived typicality of out-group partner(s). Where tensions are high, however, an item such as ‘during contact we discuss intergroup diVerences’ may trigger negative intergroup diVerentiation. We note three future priorities for research on moderation. First, we need systematic studies of which measures of salience are best moderators of contact eVects and which, if any, actually have negative eVects. This could be done with both survey research and, ideally, experimental studies. We know of only one relevant experiment, in which Ensari and Miller (2002; Study 2) manipulated both salience and typicality along with self-disclosure. They found that typicality plus self-disclosure led to more favorable ratings of out-group targets, regardless of salience; however, when the out-group member was atypical, ratings of the out-group were significantly more positive in the salience than nonsalience conditions. Thus, consistent with our view, some level of category salience must be maintained if positive attitudes are to generalize. Second, when should salience be introduced into the contact setting? The one study on this question showed only that salience must be introduced but found no eVect of when this was done (van Oudenhoven et al., 1996). However, we still believe it is reasonable to predict that, when it is feared or expected that category salience is likely to have negative eVects (e.g., under high anxiety or threat, competition, intractable bargaining), category salience should be introduced gradually, or perhaps after initial interpersonal interaction (see also Wilder, 1993). Third, there is a need for further research on generalization. We have accumulated considerable evidence that individual-to-group generalization is greatly enhanced by maintaining group salience in contact situations. There is also evidence, as reported in Sections VI and VIII, that outgroup-to-out-group generalization can be easier under conditions of high group salience, although little is yet known about the processes that govern this. We assume that there must be a degree of perceived similarity between out-groups to permit the transfer of positive feeling from one to another. Thus, in Voci and Hewstone (2004a, Study 2, reported in Section VI), the generalization from individual national out-groups (North Africans and Albanians) to ‘immigrants’ as a whole must have been facilitated by the fact that the former categories are often associated with the latter in the minds of host-society members. Similarly, in Maras and Brown (1996; see Section VIII), the extension of positive evaluations of one group (learning disability) to others (hearing impairment, physical disability) was

322

RUPERT BROWN AND MILES HEWSTONE

presumably possible because of the common ‘disability’ category that they all shared. The limits of this perceived similarity are not known. For example, will an intervention designed to reduce prejudice toward ethnic groups within a country also lessen prejudice towards ‘foreigners’? Also underresearched is how to maximize cross-situational generalizability, an issue relatively neglected since early research on the contact hypothesis (e.g., Minard, 1952). It remains an important topic for practical applications of contact to discover how prejudice-reduction programs can avoid becoming so parochially limited in their eVects. 3. Mediation of Intergroup Contact EVects The third component of the model is the set of mediating variables. In its first version, our model emphasized various cognitive processes that we believed would lead to positive (or negative) outcomes of contact (social categorization, stereotyping, expectancies, and attribution processes). This led to our work on the cognitive processes implicated in stereotype change reviewed in Section IV. However, the subsequent decade (prompted by Pettigrew, 1986) saw a gradual shift in emphasis to more aVective mediators, starting with intergroup anxiety, which has proved to be a potent variable in several contexts. Since Islam and Hewstone’s (1993) study, research has consistently found that good-quality contact reliably leads to less anxiety, which in turn leads to more favorable intergroup attitudes. Other mediators have also been identified, principally empathy, perspective-taking, and selfdisclosure. The pattern is consistent. Again, favorable contact generally leads to increases in these emotional and cognitive variables, which then leads to a more positive intergroup orientation. Once more, this is highly suggestive of how successful prejudice-reduction interventions could be implemented. In terms of measures, we would recommend that a range of potential mediators are assessed (preferably simultaneously), including both cognitive (e.g., knowledge) and, especially, aVective variables. The chosen set of aVective mediators should, however, allow for the possibility that contact may improve intergroup relations by both reducing negative and increasing positive aVect. We have also made progress in identifying mediators for both direct and extended contact eVects, notably intergroup anxiety and self-disclosure (e.g., Paolini et al., in press; Turner et al., 2004). An important new development for the model was the realization that salience might not only moderate the contact-outcome relationship but could act on the hypothesized mediators as well. As we saw in Section VII.D, there are now several documented examples where salience qualifies the relationship between contact and the mediator. Recognition of this additional

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

323

complexity of the model again has implications for designing contact interventions to reduce prejudice. A priority for future research, given the relative recency of studies that have tested for moderated mediation, is to ascertain which, if any, mediators are (or are not) moderated by salience. Thus far, the few studies have shown that salience moderates intergroup anxiety, as predicted, in most studies (e.g., Harwood et al., in press, Study 2; Voci & Hewstone, 2003a, Studies 1 and 2; Vonofakou et al., 2003). There remains, however, some doubt in the case of empathy (but see Voci & Hewstone, 2003a, Studies 1–3), because experimental studies have shown generalized eVects of empathy without the need for salience (Vescio et al., 2004; Studies 1 and 2), consistent with Batson et al.’s (1997) claim, but Voci and Hewstone (2003a) found that salience did moderate empathy (see Paolini et al, in press, for further discussion). Future work on empathy—commonly defined, as we said earlier, as the capacity to imagine and feel how others are feeling—should also try to unpack it further. For example, Iyer, Leach, and Crosby (2003) have suggested that empathic feelings can be diVerentiated according to the main focus of attention. Putting oneself in the shoes of a disadvantaged group can cause a perceiver to feel bad personally (‘Thinking about that group’s plight causes me distress’), or it could focus the perceiver’s attention on the emotions of the out-group (‘It must be distressing for that group to experience such a plight’). Iyer and her colleagues speculate that the latter emotional reaction might be more closely linked to positive feelings toward the outgroup than the former. The literature on empathy also distinguishes between cognitive (e.g., Galinsky & Moskowitz, 2000) and emotional (e.g., Batson et al., 1997) forms of empathy. Future research on contact could usefully investigate further whether these variants of empathy can be empirically distinguished and, if so, whether they have diVerent mediational eVects (see Voci & Hewstone, 2004b). Finally, there are doubtless other intergroup emotions that are more relevant in specific intergroup contexts, such as collective guilt. Researchers are beginning to discover that collective guilt has important implications for intergroup relations (e.g., Branscombe & Doosje, 2004). There are indications that feelings of collective guilt on the part of the majority are associated with attempts to make reparation to minorities who have experienced long histories of discrimination (Brown, Gonza´lez, Zagefka, Manzi, & Saiz, 2004b; Doosje, Branscombe, Spears, & Manstead, 1998; Iyer et al., 2003). The beneficial eVects of collective guilt may be mediated by empathy (Brown et al., 2004), an emotion which, as we have seen, is a strong mediator of contact eVects. In our work in Northern Ireland, which has suVered conflict for more than 30 years, we found that forgiveness was positively associated with both collective guilt and more contact with out-group friends (Hewstone et al., 2005).

324

RUPERT BROWN AND MILES HEWSTONE

4. Outcome Measures Initially, our research program concentrated mainly on stereotype change (see Section IV) and measures of generalized attitude change toward the outgroup as a whole (e.g., Brown et al., 1999; Van Oudenhoven et al., 1996). However, an important extension to the model has been the discovery that it seems to apply well to an increasingly wide range of social indicators of the state of intergroup relations. We have found that measures of intergroup aVect (liking), trust, and forgiveness are all predicted by various kinds of contact, either directly or indirectly. Although there are other important outcome variables still to be investigated (see below), finding similar eVects on such a broad range of intergroup measures gives us considerable confidence in the model’s practical utility. We have also recently found that detailed measures of out-group contact do predict a weakened own-race eVect in discriminating faces of own from other ethnic groups (Walker & Hewstone, 2004a,b). We have also begun to investigate the eVects of contact on indirect and implicit measures of intergroup attitude. As reported earlier (Section VII), there are indications that implicit measures are sensitive to variations in contact; what we do not yet know is what, if any, variables mediate those eVects. In addition, other indirect measures of intergroup attitude may well prove profitable to explore (e.g., linguistic intergroup bias; Maass, 1999). Recently, we have shown that contact also aVects relatively subtle measures of ‘infra-humanization’ (Leyens et al., 2000), such as the tendency to attribute more secondary, uniquely human emotions to the in- than out-group (Brown et al., 2004). In a study of intergroup contact in Northern Ireland, we found that the more contact respondents had with members of the other group, the less they favored the in-group with the ascription of secondary emotions. Moreover, this measure of infra-humanization mediated the eVect of contact on forgiveness (Cairns, Niens, Tam, Hewstone, & Voci, in press). Of course, the ‘holy grail’ will be to find whether, and if so how, contact interventions have a lasting impact on actual behavior. When will people act more warmly, or at least less negatively or indiVerently, toward the outgroup as a result of positively structured cooperative encounters with some individual members of it? We wish, however, to focus on two main priorities for future research in this section, namely the relatively weak eVect of contact on stereotypes and the link between contact and attitude functions. Allport (1954) explicitly argued that contact would aVect both attitudes and stereotypes, and Pettigrew and Tropp’s (2004) meta-analysis reported that contact was generally associated with more positive attitudes toward and less negative stereotypes of the out-group. The latter eVect was, however, smaller, and

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

325

stereotypes have been used less often than attitudes as dependent measures in contact research (Tropp & Pettigrew, in press; note that the use of stereotypes as indicators peaked in the 1970s). Three interesting lines of inquiry concern types of contact situations, aVective processes in stereotype change, and whether moderation is necessary for stereotype change. First, Rothbart and John’s (1985) perceptive analysis of stereotype change noted that characteristics of the contact situation are correlated with the observation of particular types of behaviors (e.g., one is more likely to observe extroverted behaviors in informal than formal settings). Thus, future research should study whether stereotype change may be relatively more constrained than attitude change by the restricted range and nature of settings in which intergroup contact actually occurs. Second, research on stereotype change (see Section IV) has been almost exclusively cognitive and has neglected aVective processes. If aVect is stored with schemas, then contact might need to target these ‘aVective tags’ (Fiske, 1982) and not simply cognitive, information-based components of stereotypes. Moreover, the relative weight of the aVective component of people’s reactions to groups may vary not only across individuals (Stangor et al., 1991) but across intergroup contexts (Haddock, Zanna, & Esses, 1993). Thus, future research should undertake a more systematic analysis of individual diVerences in aVect and stereotyping across a range of settings (Paolini et al., in press). Third, Wolsko et al. (2003) have argued that stereotypes, especially, are moderated by group salience. If this is the case, then perhaps meta-analytic results would be diVerent once there are suYcient studies of the eVect of contact on stereotypes that have also measured category salience during contact. In our more recent research, we have systematically studied multiple aspects, and measures, of attitudes as outcome variables. We have introduced new measures of, and shown eVects of contact on, attitude strength and attitude accessibility (Vonofakou et al., 2003). A research priority is to link intergroup contact more closely to the functions fulfilled by relevant intergroup attitudes (Maio & Olson, 2000). For example, Herek’s (1987) ‘Attitude Function Inventory’ assessed four functions, two of which seem closely linked to diVerent types of contact. Thus, the ‘experiential-schematic’ function (e.g., ‘‘my opinions about [the target group] are based mainly on my personal experiences with [specific members of the target group]’’) appears to be based on direct contact, whereas the ‘social-expressive’ function (e.g., ‘‘my opinions about [the target group] are based mainly on learning how [the target group] are viewed by the people whose opinions I respect’’) appears to be based on extended contact. Again, this potential link between contact and

326

RUPERT BROWN AND MILES HEWSTONE

attitude functions would have great importance for planned interventions, which will be most eVective if they target the motivation expressed by an attitude (Petty, Wheeler, & Bizer, 2000) and ensure that they target the appropriate function for the chosen level of abstraction (e.g., attitudes may have diVerent functions for superordinate [‘immigrants’] and subordinate groups [‘Poles’, ‘Albanians’]; see Watt, Maio, Rees, & Hewstone, 2004, under review).

B. THEORETICAL INTEGRATION: INTERGROUP CONTACT, PERSONALIZATION, AND COMMON IN-GROUP IDENTITY Thus far, we have reviewed how our model has expanded to accommodate the array of empirical research findings it has stimulated. Next we consider how easily it can be reconciled with the two models it was first juxtaposed against (Brewer & Miller, 1984; Gaertner & Dovidio, 2000), and Pettigrew’s (1998) subsequent attempted integration. These models can be viewed, to some extent at least, as complementary (see Hewstone, 1996; Miller, 2002), and their major strengths can and should be integrated. If we start with the Brewer–Miller decategorization model, it is now possible to discern clear convergences between the two approaches. The frequent finding that contact based on friendship with out-group members is a powerful predictor of a positive orientation toward the out-group as a whole is entirely consistent with Brewer and Miller’s contention that contact should be organized to maximize personalized interactions. Moreover, the fact that self-disclosure has proved to be a reliable mediator also sits well with that same perspective. However, what is also clear is that encouraging such intimate relationships between in-group and out-group members is not incompatible with maintaining some category salience in interactions between them which, as we have seen, is still an important ingredient to facilitate generalized attitude change. As we saw in Section VII.D, there are now several studies showing that ‘interpersonal’ variables like friendship and self-disclosure are linked more clearly to each other and to positive intergroup outcomes under high salience. Ensari and Miller (2002) reached exactly this same conclusion based on their experimental investigations into self-disclosure, typicality, and group salience. We have argued (Hewstone, 1996), theoretically, that the bipolar distinction between interpersonal and intergroup behavior (Brown & Turner, 1981; Tajfel, 1978), should now be abandoned in favor of a twodimensional conception, much as proposed by Stephenson (1981); see Fig. 13. Implicitly, our earlier chapter acknowledged the importance of both interpersonal and intergroup contact, because we proposed that Allport’s

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

327

Fig. 13. Interpersonal and intergroup contact as orthogonal dimensions (after Stephenson, 1981, p. 196).

(interpersonal) conditions should be paired with our own (intergroup) condition (see Hewstone & Brown, 1986, p. 18). Miller (personal communication) argues that social category salience was also implicit in the Brewer and Miller (1984) decategorization model. Its relevant experimental research ensured category salience (see Section III.A), and Miller (2002) assumes that some level of category salience is involved in personalization. Now we are explicit in recognizing the importance of both interpersonal and intergroup dimensions. In such a model, we would argue, the optimal conditions occur in contact situations that are ‘high’ on both the interpersonal and the intergroup dimensions. In contrast, the other combinations—low on either or both of the dimensions—are likely to be less successful. A situation that promotes high interpersonal intimacy but where group salience is absent would be unlikely to lead to generalization and will result in only piecemeal attitude change. High group salience with little possibility for developing interpersonal acquaintance might create conditions leading to intergroup anxiety (e.g., Islam & Hewstone, 1993). Finally, situations devoid of both interpersonal and intergroup significance would be unlikely to generate (or inhibit) any of the aVect that we have shown to be so important in mediating attitude change. Parenthetically, we note that this point of view is at variance with the claim by Turner and his colleagues (1987) that the intergroup and interpersonal levels are functionally antagonistic and cannot vary independently

328

RUPERT BROWN AND MILES HEWSTONE

as we propose here. Still, we believe that the position we articulate here is more consistent with the empirical reality, now documented by many studies reviewed here and elsewhere (e.g., Deschamps & Devos, 1998), that an interpersonal mode of interaction is not necessarily at odds with the maintenance of high levels of intergroup salience. The links between our model and the Gaertner–Dovidio common ingroup identity model should also be apparent. Indeed, such a connection was made by Gaertner and his colleagues themselves in an early revision of their model (Gaertner et al., 1993). Their proposal of a ‘dual identity’ strategy—simultaneous salience of subordinate and superordinate categories—was an explicit recognition that a wholesale adoption of a single common in-group identity would be likely to be resisted by minorities, to which we would add only that the ‘dual identity’ approach oVers the additional advantage of helping to promote generalization. Despite this easy dove-tailing of the Gaertner–Dovidio and Hewstone–Brown models, there remains the problem that the dual-identity strategy appears to be more attractive to minorities than to the majority, as we noted in Sections V and IX. One reason for this is suggested by Mummendey and Wenzel’s (1999) ingroup projection theory. According to this account, the image of the superordinate category held by the majority coincides perfectly with their image of themselves. Hence, they will have little reason or need to conceive of a more diVerentiated common in-group. Minorities, on the other hand, by virtue of their lesser numerical and social status, find it less easy to ‘project’ their ingroup onto the superordinate group and hence prefer the dual or integration option. Mummendey and Wenzel’s suggested solution to this problem is to advocate the construction of less monolithic and more multifaceted superordinate category representations that will be more resistant to in-group projection processes by the majority (Waldzus, Mummendey, Wenzel, & Weber, 2003). While such a cognitive strategy—presumably as part of a multicultural approach in educational curricula and citizenship programs— may well be helpful, we suspect that it may need to be supplemented by more aVectively laden interventions. Pettigrew (1998) oVered an integration of the three models that was based on a particular temporal ordering of interventions. He proposed that the most successful strategy would be one in which contact was structured so as to minimize category salience initially (Brewer–Miller), followed by a phase in which subgroup categories were made salient (Hewstone–Brown), and concluding with the development of a common in-group identity (Gaertner– Dovidio). Such a sequence, he argued, should reduce intergroup anxiety and promote cross-group friendships, permit individual-to-group generalization of positive attitude change, and capitalize on the social benefits of having a more inclusive category with which to identify. There has been little

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

329

empirical research to test this ingenious idea. Thus, the ‘jury is still out’ on how necessary Pettigrew’s proposed sequence is for optimal prejudice reduction. Practically speaking, it may not always be convenient or feasible to arrange intergroup contact in such a manner, simply because the exigencies of some educational, occupational, or community contexts may not permit such a careful staging and sequencing of intergroup encounters. Also, in light of our remarks above about the possibility and desirability of encouraging simultaneous interpersonal and intergroup interactions, the first two steps of the proposed sequence may be collapsible into one. Of course, it is theoretically important to show where models converge and diverge, but practically it is also a research priority to use the insights gained from the past two decades of testing and refining these models to design eVective prejudice-reduction interventions. Unsurprisingly, we believe that the starting point for any such intervention should be to provide ample opportunities for participants to become aware of and appreciate their diverse group backgrounds, be these ethnic, religious, or national (or whatever identity is contextually most significant). But, whereas once we would have stressed such group salience augmentation more or less exclusively, we now advocate its implementation alongside activities that promote interpersonal closeness. Consistent with our argumentation earlier, we believe that encouraging such potentially individuating behaviors as self-disclosure and perspective-taking is not incompatible with group salience-inducing discussions of category memberships and diVerences. Indeed, since such encounters would be likely to increase empathy and possibly decrease anxiety, two of the most powerful mediators of contact eVects that we have identified, they should probably be regarded as desirable if not essential features of any such intervention. Finally, and coherent with our espousal of the dualidentity approach to prejudice reduction, such activities could be usefully supplemented by restructuring the environment so as to promote the feeling of being ‘diVerent groups on the same team’ (in the words of one of the measures used by Gaertner & Dovidio, 2000), each having something valuable to contribute to the common cause. Practitioners who can successfully combine interpersonal, intergroup, and common-group elements in this way will, we believe, have a good chance of achieving genuine and enduring reductions in intergroup tensions.

C. CONCLUSIONS In this article, we have revisited and revised our 1986 model of intergroup contact and reviewed the now-extensive body of relevant research. The major changes in the new model are: (1) the shift from an exclusive emphasis on

330

RUPERT BROWN AND MILES HEWSTONE

intergroup contact to the view that contact should be ‘high’ on both intergroup and interpersonal dimensions; (2) the inclusion of a range of mediators, especially aVective ones, whose eVects are frequently also moderated by category salience; and (3) an openness to theoretical integration of our own and competing models. Given the breadth of intergroup settings considered (e.g., from contact between young children with and without disabilities to Catholic-Protestant contact in Northern Ireland), the variety of predictors, moderators, mediators and outcomes measured, and the range of methodologies used, we are confident that our model can be widely applied. We were, in hindsight, too extreme in our 1986 focus on intergroup contact alone; but so too were Brewer and Miller (1984) in their focus on personalization. Yet this is not something we regret. We contend that our respective arguments for extreme, albeit contradictory, positions represent good examples of what Popper (1959) called ‘bold’ theorizing; i.e., they are more likely to be falsified (see Pettigrew, 1986). In light of the nearly 20 years of subsequent research, both initial theories have, we argue, been falsified, but this has led to a more integrative theory that unites the interpersonal and intergroup perspectives (a view now shared by proponents of all three models; e.g., Gaertner & Dovidio, 2000; Hewstone, 1996; Miller, 2002).1 Notwithstanding the positive conclusions we draw from this research program, focused here especially on our own work, we also wish to acknowledge some limitations in the ultimate impact of intergroup contact. We are quite sure that contact impacts on prejudice (Pettigrew & Tropp, 2004), and we have reported its impact on multiple other social-psychological outcomes. But does contact have the power to have an impact on social conflict? Commentators outside of social psychology have concluded that contact often lowers prejudice at the individual level of analysis but is ineVective at the group level (Forbes, 1997). Moreover, Green and Seher (2003) have argued that prejudice plays only a limited role in ethnic conflict, from which we can conclude that a prejudice-based intervention such as contact can, at best, have only a limited role in reducing such conflict. Without wishing to challenge these latter views, we point out that ‘subjective’ elements of conflict often persist long after its ‘objective’ elements disappear. They can become independent of the initiating, more ‘objective,’ causes of the conflict and contribute to an escalation and continuation of violence even after the initial causes have become irrelevant (Deutsch, 1973; Tajfel & Turner, 1979). Thus, the formal resolution of a conflict is 1 Readers might be amused to know that, in almost two decades of theoretical polarization, all the main protagonists (Marilynn Brewer, Rupert Brown, Jack Dovidio, Sam Gaertner, Miles Hewstone, and Norman Miller) have shared friendships, bottles of wine, good dinners, and the occasional Mozart sonata, thus testifying to our own interpersonal and intergroup relationships.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

331

often merely the first step toward peaceful co-existence. To promote peace and reconciliation, a psychological process is required to change people’s often deeply entrenched beliefs and feelings about the out-group, their in-group, and the relationship between the two (Bar-Tal, 2000). It is here that contact can help even in the case of conflict. One example is given by our own extensive work in Northern Ireland (see Hewstone et al., in press, 2004, 2005). Another is provided by Varshney’s (2002) sociological analysis of the positive impact of inter-communal networks of civic life on resisting pressures that polarize the Hindu and Muslim communities in India. We do not wish to end on the Panglossian note that contact theory and research in general, and our contributions in particular, are perfect. They are not. But they are constantly developing, and their falsifiable hypotheses are constantly subjected to empirical testing. Nonetheless, we believe that our research program testifies to the central importance of intergroup contact in reducing prejudice and resolving intergroup conflict. As an intervention, intergroup contact cannot possibly deal with all the dimensions of a fullblown, typically ethnic, intergroup conflict, but it is diYcult to imagine successful programs aimed at improving intergroup relations without sustained, positive contact between members of the two previously competing groups.

References Alicke, M. D., Klotz, M. L., Breitenbecher, D. L., Yurak, T. J., & Vredenburg, D. S. (1995). Personal contact, individuation, and the better-than-average effect. Journal of Personality and Social Psychology, 68, 804–825. Allen, C., & Nielson, J. S. (2002). Summary report on Islamophobia in the EU after 11 September 2001. Vienna: European Monitoring Centre on Racism and Xenophobia. Allport, G. W. (1954). The nature of prejudice. Addison-Wesley. Amir, Y. (1969). Contact hypothesis in ethnic relations. Psychological Bulletin, 71, 319–342. Amir, Y., & Katz, P. (1976). The role of intergroup contact in change of prejudice and ethnic relations. In P. A. Katz (Ed.), Towards the elimination of racism (pp. 245–308). New York: Pergamon. Anastasio, P. A., Bachman, B. A., Gaertner, S. L., & Dovidio, J. F. (1996). In R. Spears, P. J. Oakes, N. Ellemers, & S. A. Haslam (Eds.), The social psychology of stereotyping and group life (pp. 236–256). Oxford: Blackwell. Armstrong, B., Johnson, D. W., & Balour, B. (1981). Effects of co-operative versus individualistic learning experiences on interpersonal attraction between learning disabled and normal progress elementary school students. Contemporary Educational Psychology, 15, 604–616. Aron, A., & Aron, E. N. (1996). Self and self-expansion in relationships. In G. Fletcher & J. Fitness (Eds.), Knowledge structures in close relationships: a social psychological approach (pp. 325–344). Hillsdale, NJ: Erlbaum.

332

RUPERT BROWN AND MILES HEWSTONE

Aron, A., McLaughlin-Volpe, T., Mashek, D., Lewandowski, G., Wright, S. C., & Aron, E. N. (in press). Including others in the self. European Review of Social Psychology, 16. Ashmore, R. D. (1981). Sex stereotypes and implicit personality theory. In D. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 37–81). Hillsdale, NJ: Erlbaum. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Barsalou, L. W. (1987). The instability of graded structure: Implications for the nature of concepts. In U. Neisser (Ed.), Concepts and conceptual development (pp. 101–140). New York: Cambridge University Press. Bar-Tal, D. (2000). From intractable conflict through conflict resolution to reconciliation: Psychological analysis. Political Psychology, 21, 351–365. Batson, C. D., Early, S., & Salvarani, D. (1997). Perspective talking: Imagining how another feels versus imagining how you would feel. Personality and Social Psychology Bulletin, 23, 751–758. Batson, C. D., Polycarpou, M. P., Harmon-Jones, E., Imhoff, H. J., Mitchener, E. C., Bednar, L. L., Klein, T. R., & Highberger, L. (1997). Empathy and attitudes: Can feeling for a member of a stigmatized group improve feelings toward the group? Journal of Personality and Social Psychology, 72, 105–118. Berger, C. R., & Bradac, J. J. (1982). Language and social knowledge: Uncertainty in interpersonal relations. London: Edward Arnold. Berry, J. W. (1984). Cultural relations in plural societies: Alternatives to segregation and their sociopsychological implications. In N. Miller & M. B. Brewer (Eds.), Groups in contact: The psychology of desegregation (pp. 11–28). New York: Academic Press. Berry, J. W. (1997). Immigration, acculturation, and adaptation. Applied Psychology: An International Review, 46, 5–68. Bettencourt, A., Charlton, K., & Kernaham, C. (1997). Numerical representation of groups in co-operative settings: Social orientation effects on ingroup bias. Journal of Experimental Social Psychology, 33, 630–659. Bettencourt, B. A., Brewer, M. B., Croak, M. R., & Miller, N. (1992). Cooperation and the reduction of intergroup bias: The role of reward structure and social orientation. Journal of Experimental Social Psychology, 28, 301–309. Bigler, R. (1995). The role of classification skill in moderating environmental influences on children’s gender stereotyping: A field study of the functional use of gender in the classroom. Child Development, 66, 1072–1087. Bodenhausen, G. V. (1993). Emotions, arousal, and stereotypic judgments: A heuristic model of affect and stereotyping. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 13–37). San Diego: Academic Press. Bornstein, R. F. (1989). Exposure and affect: Overview and meta-analysis of research. Psychological Bulletin, 106, 265–289. Bourhis, R. Y., Moise, L. C., Perreault, S., & Senecal, S. (1997). Towards an interactive acculturation model: A social psychological approach. International Journal of Psychology, 32, 369–389. Branscombe, N. R., & Doosje, B. (Eds.) (2004). Collective guilt: International perspectives. New York: Cambridge University Press. Brewer, M. B. (1979). Ingroup bias in the minimal intergroup situation: A cognitivemotivational analysis. Psychological Bulletin, 86, 307–324.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

333

Brewer, M. B. (1988). A dual process model of impression formation. In R. S. Srull & R. S. Wyer (Eds.), Advances in Social Cogntion (Vol. 1, pp. 1–36). Hillsdale, NJ: Erlbaum. Brewer, M. B., Dull, V., & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes. Journal of Personality and Social Psychology, 41, 656–670. Brewer, M. B., & Gaertner, S. L. (2001). Toward reduction of prejudice: Intergroup contact and social categorization. In R. Brown & S. L. Gaertner (Eds.), Blackwell handbook of social psychology: Intergroup processes (pp. 451–474). Oxford: Blackwell. Brewer, M. B., & Miller, N. (1984). Beyond the contact hypothesis: Theoretical perspectives on desegregation. In N. Miller & M. B. Brewer (Eds.), Groups in contact: The psychology of desegregation (pp. 281–302). Orlando, FL: Academic Press. Brewer, M. B., & Miller, N. (1988). Contact and cooperation: When do they work? In P. Katz & D. Taylor (Eds.), Eliminating racism: Profiles in controversy (pp. 315–326). New York: Plenum. Brown, R., Eller, A., Leeds, S., & Stace, K. (2004a). Intergroup contact, perceived typicality and intergroup attitudes: A longitudinal study. University of Sussex. Unpublished manuscript. Brown, R., Gonza´lez, R., Zagefka, H., Manzi, J., & Saiz, J. (2004b). Collective guilt, shame and sympathy, their implications for reparation policies: A longitudinal study in Chile. Paper presented at the Society for the Psychological Study of Social Issues, Washington, DC. Brown, R., Maras, P., Masser, B., Vivian, J., & Hewstone, M. (2001). Life on the ocean wave: Testing some intergroup hypotheses in a naturalistic setting. Group Processes and Intergroup Relations, 4, 81–97. Brown, R., & Turner, J. C. (1981). Interpersonal and intergroup behaviour. In J. C. Turner & H. Giles (Eds.), Intergroup Behaviour (pp. 33–65). Oxford: Blackwell. Brown, R., Vivian, J., & Hewstone, M. (1999). Changing attitudes through intergroup contact: The effects of group membership salience. European Journal of Social Psychology, 29, 741–764. Cacioppo, J. T., & Berntson, G. T. (2001). The affect system and racial prejudice. In J. A. Bargh & D. K. Apsley (Eds.), Unravelling the complexities of social life: A festschrift in honor of Robert B. Zajonc (pp. 95–110). Washington, DC, US: American Psychological Association. Cadinu, M. R., & Rothbart, M. (1996). Self-anchoring and differentiation processes in the minimal group setting. Journal of Personality and Social Psychology, 70, 661–677. Cairns, E., Niens, U., Tam, T., Hewstone, M., & Voci, A. (in press). Intergroup forgiveness and intergroup conflict. In E. Worthington (Ed.), Handbook of forgiveness. NY: Brunner/ Routledge. Cameron, L., Rutland, A., & Brown, R. (2004). Changing children’s attitudes towards minority groups. Paper presented at the Jena workshop on intergroup relations, Oppburg, Germany. Chein, I., Cook, S. W., & Harding, J. (1948). The field of action research. American Psychologist, 3, 43–50. Cohen, S., Sherrod, D. R., & Clark, M. S. (1986). Social skills and the stress-protective role of social support. Journal of Personality and Social Psychology, 50, 963–973. Coke, J. S., Batson, C. D., & McDavis, K. (1978). Empathic mediation of helping: A two-stage model. Journal of Personality and Social Psychology, 36, 752–766. Cook, S. W. (1962). The systematic analysis of socially significant events: A strategy for social research. Journal of Social Issues, 18, 66–84. Cook, S. W. (1978). Interpersonal and attitudinal outcomes in cooperating interracial groups. Journal of Research and Development in Education, 12, 97–113. Craig, J., Cairns, A., Hewstone, M., & Voci, A. (2002). Young people’s attitudes to and contact with members of the religious out-group. Unpublished manuscript, University of Ulster.

334

RUPERT BROWN AND MILES HEWSTONE

Crisp, R. J., & Hewstone, M. (1999). Differential evaluation of crossed category groups: Patterns, processes, and reducing intergroup bias. Group Processes and Intergroup Relations, 2, 307–333. Davis, M. H. (1994). Empathy: A social psychological approach. Madison, WI: Brown & Benchmark. Deschamps, J.-C., & Devos, T. (1998). Regarding the relationship between social identity and personal identity. In S. Worchel, J. F. Morales, D. Paez, & J.-C. Deschamps (Eds.), Social identity: International perspectives (pp. 1–12). London: Sage. Desforges, D. M., Lord, C. G., Pugh, M. A., Sia, T. L., Scarberry, N. C., & Ratcliff, C. D. (1997). Role of group representativeness in the generalization part of the contact hypothesis. Basic and Applied Social Psychology, 19, 183–204. Desforges, D. M., Lord, C. G., Ramsey, S. L., Mason, J. A., Van Leeuwen, M. D., West, S. C., & Lepper, M. R. (1991). Effects of structured cooperative contact on changing negative attitudes towards stigmatized social groups. Journal of Personality and Social Psychology, 60, 531–544. Deutsch, M. (1973). The resolution of conflict: Constructive and destructive processes. New Haven, CT: Yale University Press. Doosje, B., Branscombe, N. R., Spears, R., & Manstead, A. S. R. (1998). Guilty by association: When one’s group has a negative history. Journal of Personality and Social Psychology, 75, 872–886. Dovidio, J. F., Brigham, J. C., Johnson, B. T., & Gaertner, S. L. (1996). Stereotyping, prejudice, and discrimination: Another look. In N. Macrae, C. Stangor, & M. Hewstone (Eds.), Foundations of stereotypes and stereotyping (pp. 276–319). New York: Guilford. Dovidio, J. F., Kawakami, K., Johnson, C., Johnson, B., & Howard, A. (1997b). On the nature of prejudice: Automatic and controlled processes. Journal of Experimental Social Psychology, 33, 510–540. Dovidio, J., Kawakami, K., & Gaertner, S. L. (2000). Reducing contemporary prejudice: Combating explicit and implicit bias at the individual and intergroup level. In S. Oskamp (Ed.), Reducing prejudice and discrimination (pp. 137–163). Mahwah, NJ: Erlbaum. Dovidio, J. F., Gaertner, S. L., & Kawakami, K. (2003). Intergroup contact: The past, present and future. Group Processes and Intergroup Relations, 5–21. Dovidio, J. F., Gaertner, S. L., & Validzic, A. (1998). Intergroup bias, status, differentiation, and common ingroup identity. Journal of Personality and Social Psychology, 75, 109–120. Dovidio, J. F., Gaertner, S. L., Validzic, A., Matoka, K., Johnson, B., & Frazier, S. (1997a). Extending the benefits of re-categorization: Evaluations, self-disclosure and helping. Journal of Experimental Social Psychology, 33, 401–420. Dovidio, J. F., ten Vergert, M., Stewart, T. L., Gaertner, S. L., Johnson, J. D., Esses, V. M., Riek, B. M., & Pearson, A. R. (2004). Perspective and prejudice: Antecedents and mediating mechanisms. Personality and Social Psychology Bulletin, (in press). Eller, A., & Abrams, D. (2003). ‘Gringos’ in Mexico: Cross-sectional and longitudinal effects of language school-promoted contact on intergroup bias. Group Processes and Intergroup Relations, 6, 55–75. Eller, A., & Abrams, D. (2004). Come together: longitudinal comparisons of Pettigrew’s reformulated intergroup contact model and the common ingroup identity model in Anglo-French and Mexican-American contexts. European Journal of Social Psychology, 34, 1–28. Ensari, N., & Miller, N. (2002). The out-group must not be so bad after all: The effects of disclosure, typicality, and salience on intergroup bias. Journal of Personality and Social Psychology, 83, 313–329.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

335

Fazio, R. H. (1990). Multiple processes by which attitudes guide behavior: The MODE model as an integrative framework. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 75–109). San Diego, CA: Academic Press. Finlay, K. A., & Stephan, W. G. (2000). Improving intergroup relations: The effects of empathy on racial attitudes. Journal of Applied Social Psychology, 30, 1720–1737. Fiske, S. T. (1982). Schema-triggered affect: Applications to social perception. In M. S. Clark & S. T. Fiske (Eds.), Affect and cognition: The 17th Annual Carnegie Symposium (pp. 55–77). Hillsdale, N.J: Erlbaum. Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878–902. Fiske, S. T., & Neuberg, S. L. (1990). A continuum model of impression formation from category-based to individuating processes: Influences of information and motivation on attention and interpretation. In M. P. Zanna (Ed.), Advances in experimental social psychology (pp. 1–74). San Diego, CA: Academic Press. Forbes, H. D. (1997). Ethnic conflict: Commerce, culture, and the contact hypothesis. New Haven, CT: Yale University Press. Gaertner, S., & Dovidio, J. (2000). Reducing intergroup bias: The common ingroup identity model. Hove: Psychology Press. Gaertner, S., Mann, J., Murrell, A., & Dovidio, J. (1989). Reducing intergroup bias: The benefits of recategorization. Journal of Personality and Social Psychology, 57, 239–249. Gaertner, S. L., Dovidio, J., Anastasio, P. A., Bachman, B. A., & Rust, M. C. (1993). The common ingroup identity model: Recategorization and the reduction of intergroup bias. European Review of Social Psychology, 4, 1–26. Gaertner, S. L., Dovidio, J., Rust, M., Nier, J., Banker, B., Ward, C., Mottola, G., & Houlette, M. (1999). Reducing intergroup bias: Elements of intergroup co-operation. Journal of Personality and Social Psychology, 76, 388–402. Gaertner, S. L., Mann, J. A., Dovidio, J., Murrell, A. J., & Pomare, M. (1990). How does cooperation reduce intergroup bias? Journal of Personality and Social Psychology, 59, 692–704. Gaertner, S. L., Rust, M. C., Dovidio, J. F., Bachman, B. A., & Anastasio, P. A. (1996). The contact hypothesis: the role of a common ingroup identity on reducing intergroup bias among majority and minority group members. In J. L. Nye & A. M. Brower (Eds.), What’s social about social cognition? (pp. 230–360). Newbury Park, CA: Sage Publications. Galinsky, A. D., & Moskowitz, G. B. (2000). Perspective-taking: Decreasing stereotype expression, stereotype accessibility, and in-group favoritism. Journal of Personality and Social Psychology, 78, 708–724. Gerard, H. B., & Miller, N. (1975). School desegregation. New York: Plenum. Gonza´lez, R., & Brown, R. (2001). The contact hypothesis and levels of categorization: Predicting inter-nation attitudes in the European Union. Paper presented at the EAESP Small Group Meeting on Social Identity: Motivational, Affective and Cross-Cultural Aspects, Calgiari, Sardegna. Gonza´lez, R., & Brown, R. (2003). Generalization of positive attitude as a function of subgroup and superordinate identifications in intergroup contact. European Journal of Social Psychology, 33, 195–214. Gonza´lez, R., & Brown, R. (in press). Dual identities in intergroup contact: Group status and size moderate the generalization of positive attitude change. Journal of Experimental and Social Psychology. Gonza´lez, R., Saiz, J., Manzi, J., Ordones, G., Millar, A., Sirlopu, D., & Brown, R. (2003a). Intergroup contact, social norms, and group salience: predicting prejudice and positive affect

336

RUPERT BROWN AND MILES HEWSTONE

towards minority groups in secondary and high school students in Chile. Paper presented at the conference on Anti-bias education: practice, research and theory, Evanston, Illinois. Gonza´lez, R., Manzi, J., Saiz, J., Brewer, M., De Tezanos, P., Torres, D., Aldunate, N., & Aravena, T. (2003b). Political parties and coalitions in Chile: intergroup political attitudes and positive affects as a function of political identity and intergroup contact. Paper presented at the International Society of Polı´tical Psychology 26th Annual Scientific Meeting: Political leadership in divided societies and a dangerous world, Boston, Massachusetts. Granovetter, M. (1973). Strength of weak ties. American Journal of Sociology, 78, 1360–1380. Green, D. P., & Seher, R. L. (2003). What role does prejudice play in ethnic conflict? Annual Review of Political Science, 6, 509–531. Greenland, K., & Brown, R. (1999). Categorization and intergroup anxiety in contact between British and Japanese nationals. European Journal of Social Psychology, 29, 503–522. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74, 1464–1480. Haddock, G., Zanna, M. P., & Esses, V. M. (1993). Assessing the structure of prejudicial attitudes: The case of attitudes toward homosexuals. Journal of Personality and Social Psychology, 65, 1105–1118. Hajnal, Z. (2001). White residents, black incumbents, and a declining racial divide. American Political Science Review, 95, 603–670. Hamberger, J., & Hewstone, M. (1997). Inter-ethnic contact as a predictor of blatant and subtle prejudice: Tests of a model in four West European nations. British Journal of Social Psychology, 36, 173–190. Hamilton, D. L., & Bishop, G. D. (1976). Attitudinal and behavioral effects of initial integration of white suburban neighbourhoods. Journal of Social Issues, 32, 47–67. Hamilton, D. L., & Sherman, J. W. (1994). Stereotypes. In R. S. Wyer & T. K. Srull (Eds.), Handbook of social cognition (2nd ed., pp. 1–68). Hillsdale, NJ: Lawrence Erlbaum Associates. Hantzi, A. (1995). Change in stereotypic perceptions of familiar and unfamiliar groups: The pervasiveness of the subtyping model. British Journal of Social Psychology, 34, 463–477. Harwood, J., Hewstone, M., Paolini, S., & Voci, A. (in press). Grandparent-grandchild contact and attitudes towards older adults: Moderator and mediator effects. Personality and Social Psychology Bulletin. Herek, G. M. (1987). Can functions be measured? A new perspective on the functional approach to attitudes. Social Psychology Quarterly, 50, 285–303. Hewstone, M. (1994). Revision and change of stereotypic beliefs: In search of the elusive subtyping model. European Review of Social Psychology, 5, 69–109. Hewstone, M. (1996). Contact and categorization: Social psychological interventions to change intergroup relations. In N. Macrae, C. Stangor, & M. Hewstone (Eds.), Stereotypes and stereotyping (pp. 323–368). New York: Guildford. Hewstone, M. (1997). Three lessons from social psychology: Multiple levels of analysis, methodological pluralism, and statistical sophistication. In C. McGarty & A. Haslam (Eds.), The message of social psychology (pp. 166–181). Oxford: Blackwell. Hewstone, M., & Brown, R. (1986). Contact is not enough: An intergroup perspective on the ‘Contact Hypothesis’. In M. Hewstone & R. Brown (Eds.), Contact and conflict in intergroup encounters (pp. 1–44). Oxford: Blackwell. Hewstone, M., E., Voci, A., Hamberger, J., & Niens, U. (2004). Intergroup contact, forgiveness, and experience of ‘The Troubles’ in Northern Ireland. Journal of Social Issues.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

337

Hewstone, M., Cairns, E., Voci, A., McLernon, F., Niens, U., & Noor, M. (2004). Intergroup forgiveness and guilt in Northern Ireland: Social psychological dimensions of ‘The Troubles.’ In N. Branscombe & B. Doosje (Eds.), Collective guilt: International perspectives (pp. 193–215). New York: Cambridge University Press. Hewstone, M., Cairns, Voci, A., Paolini, S., McLernon, F., Crisp, R., & Niens, U. (2005). Intergroup contact in a divided society: Challenging segregation in Northern Ireland. In D. Abrams, J. M. Marques, & M. A. Hogg (Eds.), The social psychology of inclusion and exclusion (pp. 265–292). Philadelphia, PA: Psychology Press. Hewstone, M., & Hamberger, J. (2000). Perceived variability and stereotype change. Journal of Experimental Social Psychology, 36, 103–124. Hewstone, M., Hassebrauck, M., Wirth, A., & Waenke, M. (2000). Pattern of disconfirming information and processing instructions as determinants of stereotype change. British Journal of Social Psychology, 39, 399–411. Hewstone, M., & Lord, C. G. (1998). Changing intergroup cognitions and intergroup behavior: The role of typicality. In C. Sedikides, J. Schopler, & C. A. Insko (Eds.), Intergroup cognition and intergroup behavior (pp. 367–392). Mahwah, NJ: Lawrence Erlbaum. Hewstone, M., Macrae, C. N., Griffiths, R., Milne, A., & Brown, R. (1994). Cognitive models of stereotype change: (5). Measurement, development, and consequences of subtyping. Journal of Experimental Social Psychology, 30, 505–526. Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias. Annual Review of Psychology, 53, 575–604. Hewstone, M., Voci, A., Cairns, E., & McLernon, A. (2003). Intergroup contact and intergroup attitudes in Northern Ireland: Mediational and moderational evidence. Paper presented at Society for Experimental Social Psychology, Boston. Hornsey, M. J., & Hogg, M. A. (2000). Subgroup relations: A comparison of mutual intergroup differentiation and common ingroup identity models of prejudice reduction. Personality and Social Psychology Bulletin, 26, 242–256. Islam, M. R., & Hewstone, M. (1993). Dimensions of contact as predictors of intergroup anxiety, perceived out-group variability, and out-group attitude: An integrative model. Personality and Social Psychology Bulletin, 19, 700–710. Iyer, A., Leach, C. W., & Crosby, F. (2003). White guilt and racial compensation: The benefits and limits of self-focus. Personality and Social Psychology Bulletin, 29, 117–129. Johnston, L., & Hewstone, M. (1992). Cognitive models of stereotype change (3): Subtyping and the perceived typicality of disconfirming group members. Journal of Experimental Social Psychology, 28, 360–386. Jo¨reskog, K. G., & So¨rbom, D. (1996). LISREL 8 User’s Reference Guide. Chicago: Scientific Software International. Karpinski, A., & Hilton, J. L. (2001). Attitudes and the Implicit Association Test. Journal of Personality and Social Psychology, 81, 774–788. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychology, 4th ed., (Vol. 1, pp. 233–265). New York: McGraw-Hill. Kenworthy, J., Turner, R., Hewstone, M., & Voci, A. (2005). Intergroup contact: When does it work, and why? In J. Dovidio, P. Glick, & L. Rudman (Eds.), Reflecting on Allport (pp. 278–292). Oxford: Blackwell. Krosnick, J. A., Boninger, D. S., Chuang, Y. C., Berent, M. K., & Carnot, C. G. (1993). Attitude strength: One construct or many related constructs? Journal of Personality and Social Psychology, 65, 1132–1151. La Greca, A., & Lopez, N. (1998). Social anxiety among adolescents: Linkages with peer relations and friendships. Journal of Abnormal Child Psychology, 26, 83–94.

338

RUPERT BROWN AND MILES HEWSTONE

Laurenceau, J.-P., Barrett, L. F., & Pietromonaco, P. R. (1998). Intimacy as an interpersonal process: The importance of self-disclosure, partner disclosure, and perceived partner responsiveness in interpersonal exchanges. Journal of Personality and Social Psychology, 74, 1238–1251. Levin, S., van Laar, C., & Sidanius, J. (2003). The effects of ingroup and outgroup friendships on ethnic attitudes in college: A longitudinal study. Group Processes and Intergroup Relations, 6, 76–92. Lewin, K., & Grabbe, P. (1945). Conduct, knowledge, and acceptance of new values. Journal of Social Issues, 2, 53–64. Leyens, J. Ph., Paladino, P. M., Rodriguez, R. T., Vaes, J., Demoulin, S., Rodriguez, A. P., & Gaunt, R. (2000). The emotional side of prejudice: The attribution of secondary emotions to ingroups and outgroups. Personality and Social Psychology Review, 4, 186–197. Liebkind, K. (2001). Acculturation. In R. Brown & S. Gaertner (Eds.), Blackwell handbook of social psychology (Vol. 4, pp. 387–405). Oxford: Blackwell. Liebkind, K., & McAlister, A. (1999). Extended contact through peer modelling to promote tolerance in Finland. European Journal of Social Psychology, 29, 765–780. Maass, A. (1999). Linguistic intergroup bias: Stereotype-perpetuation through language. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 31, pp. 79–121). San Diego, CA: Academic Press. Mackie, D. M., Devos, T., & Smith, E. (2000). Intergroup emotions: explaining offensive action tendencies in an intergroup context. Journal of Personality and Social Psychology, 79, 602–616. Mackie, D. M., & Smith, E. R. (Eds.) (2002). From prejudice to intergroup emotions: Differentiated reactions to social groups. New York & Hove: Psychology Press. Maio, G. R., & Olson, J. M. (Eds.) (2000). Why we evaluate: Functions of attitudes. Mahwah, New Jersey: Lawrence Erlbaum Associates. Maras, P., & Brown, R. (1996). Effects of contact on children’s attitudes toward disability: A longitudinal study. Journal of Applied Social Psychology, 26, 2113–2134. Maras, P., & Brown, R. (2000). Effects of different forms of school contact on children’s attitudes towards disabled and non-disabled peers. British Journal of Educational Psychology, 70, 337–351. Maurer, K. L., Park, B., & Rothbart, M. (1995). Subtyping versus subgrouping processes in stereotype representation. Journal of Personality and Social Psychology, 69, 812–824. McClelland, G. H., & Judd, C. M. (1993). The statistical difficulties of detecting interactions and moderator effects. Psychological Bulletin, 114, 376–390. Mendes, W. B., Blascovich, J., Lickel, B., & Hunter, S. (2002). Challenge and threat during social interactions with white and black men. Personality and Social Psychology Bulletin, 28, 939–952. Migdal, M., Hewstone, M., & Mullen, B. (1998). The effects of crossed categorization on intergroup evaluations: A meta-analysis. British Journal of Social Psychology, 37, 303–324. Miller, N. (2002). Personalization and the promise of contact theory. Journal of Social Issues, 58, 387–410. Minard, R. D. (1952). Race relationships in the Pocahontas coal field. Journal of Social Issues, 8, 29–44. MORI (2003). British views on immigration. London. Mullen, B., Brown, R., & Smith, C. (1992). Ingroup bias as a function of salience, relevance, and status: An integration. European Journal of Social Psychology, 22, 103–122. Mummendey, A., & Wenzel, M. (1999). Social discrimination and tolerance in intergroup relations: Reactions to intergroup difference. Personality and Social Psychology Review, 3, 158–174.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

339

Paolini, S., Hewstone, M., Cairns, E., & Voci, A. (2004). Effects of direct and indirect crossgroup friendships on judgments of Catholics and Protestants in Northern Ireland: The mediating role of an anxiety-reduction mechanism. Personality and Social Psychology Bulletin, 30, 770–786. Paolini, S., Hewstone, M., Voci, A., Harwood, J., & Cairns, E. (in press). Intergroup contact and the promotion of intergroup harmony: The influence of intergroup emotions. In R. Brown & D. Capozza (Eds.), Social identities: Motivational, emotional, cultural Influences. Hove, UK: Psychology Press. Park, B., Wolsko, C., & Judd, C. M. (2000). Measurement of subtyping in stereotype change. Unpublished manuscript, University of Colorado. Pettigrew, T. F. (1986). The contact hypothesis revisited. In M. Hewstone & R. Brown (Eds.), Contact and conflict in intergroup encounters (pp. 169–195). Oxford: Blackwell. Pettigrew, T. F. (1997). Generalized intergroup contact effects on prejudice. Personality and Social Psychology Bulletin, 23, 173–185. Pettigrew, T. F. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65–85. Pettigrew, T. F., & Meertens, R. W. (1995). Subtle and blatant prejudice in western Europe. European Journal of Social Psychology, 25, 57–75. Pettigrew, T. F., & Tropp, L. (2004). A meta-analytic test of intergroup contact theory. Manuscript under review. Petty, R. E., Wheeler, C., & Bizer, G. Y. (2000). Attitude functions and persuasion: An elaboration likelihood approach to matched versus mismatched messages. In G. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 133–162). Mahwah, New Jersey: Lawrence Erlbaum Associates. Pfafferott, I., & Brown, R. (2004). Acculturation attitudes of immigrant and German adolescents in Germany. Unpublished MS, University of Sussex. Phinney, J. S., Ferguson, D. L., & Tate, J. D. (1997). Intergroup attitudes among ethnic minority adolescents: A causal model. Child Development, 68, 955–969. Piontkowski, U., Florack, A., Hoelker, P., & Obdrzalek, P. (2000). Predicting acculturation attitudes of dominant and non-dominant groups. International Journal of Intercultural Relations, 24, 1–26. Popper, K. (1959). The logic of scientific discovery. London: Hutchinson. Powers, D. A., & Ellison, C. G. (1995). Interracial contact and black racial attitudes: The contact hypothesis and selectivity bias. Social Forces, 74, 205–226. Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster. Reis, H. T., & Shaver, P. (1988). Intimacy as an interpersonal process. In S. Duck (Ed.), Handbook of personal relationships (pp. 367–389). Chichester, England: Wiley. Richards, Z., & Hewstone, M. (2001). Subtyping and subgrouping: Processes for the prevention and promotion of stereotype change. Personality and Social Psychology Review, 5, 52–73. Richeson, J. A., Baird, A. A., Gordon, H. L., Heatherton, T. F., Wyland, C. L., Trawalter, S., & Shelton, J. N. (2003). An fMRI investigation of the impact of interracial contact on executive function. Nature Neuroscience, 6, 1323–1328. Rothbart, M. (1981). Memory processes and social beliefs. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 145–181). Hillsdale, NJ: Erlbaum. Rothbart, M., & John, O. P. (1985). Social categorization and behavioural episodes: a cognitive analysis of the effects of intergroup contact. Journal of Social Issues, 41, 81–104.

340

RUPERT BROWN AND MILES HEWSTONE

Rothbart, M., & Lewis, S. (1988). Inferring category attributes from exemplar attributes: Geometric shapes and social categories. Journal of Personality and Social Psychology, 55, 861–872. Scarberry, N. C., Ratcliff, C. D., Lord, C. G., Lanicek, D. L., & Desforges, D. M. (1997). Effects of individuating information on the generalization part of Allport’s contact hypothesis. Personality and Social Psychology Bulletin, 23, 1291–1299. Schofield, J. W. (1995). Review of research on school desegregation’s impact on elementary and secondary school students. In J. A. Banks & C. A. McGee (Eds.), Handbook of research on multicultural education (pp. 597–616). New York: MacMillan. Shepard, C. A., Giles, H., & LePoire, B. A. (2001). Communication accommodation theory. In W. P. Robinson & H. Giles (Eds.), The new handbook of language and social psychology (pp. 33–56). Chichester, UK: John Wiley & Sons. Simon, B., Aufderheide, B., & Kampmeier, C. (2001). The social psychology of minoritymajority relations. In R. Brown & S. Gaertner (Eds.), Blackwell handbook of social psychology: intergroup processes (pp. 303–323). Oxford: Blackwell. Smith, E. R. (1993). Social identity and social emotions: Toward new conceptualizations of prejudice. In D. L. Hamilton (Ed.), Affect, cognition and stereotyping (pp. 297–315). San Diego: Academic Press. Stangor, C., Sullivan, L. A., & Ford, T. E. (1991). Affective and cognitive determinants of prejudice. Social Cognition, 9, 359–380. Stephan, W. G. (1987). The contact hypothesis in intergroup relations. In C. Hendrick (Ed.), Group processes and intergroup relations: Review of personality and social psychology (Vol. 9, pp. 13–40). Newbury Park, California: Sage. Stephan, W. G., Boniecki, K. A., Ybarra, O., Bettencourt, A., Ervin, K. S., Jackson, L. A., McNatt, P. S., & Renfro, C. L. (2002). The role of threats in the racial attitudes of blacks and whites. Personality and Social Psychology Bulletin, 28/9, 1242–1254. Stephan, W. G., & Renfro, C. L. (2003). The role of threat in intergroup relations. In D. M. Mackie & E. R. Smith (Eds.), From prejudice to intergroup emotions: Differentiated reactions to social groups (pp. 191–207). New York: Psychology Press. Stephan, W. G., & Rosenfield, D. (1978). Effects of desegregation on racial attitudes. Journal of Personality and Social Psychology, 36, 759–804. Stephan, W. G., & Stephan, C. W. (1984). The role of ignorance in intergroup relations. In N. Miller & M. B. Brewer (Eds.), Groups in contact: The psychology of desegregation (pp. 229–255). Orlando, FL: Academic Press. Stephan, W. G., & Stephan, C. (1985). Intergroup anxiety. Journal of Social Issues, 41, 157–175. Stephan, W. G., & Stephan, C. W. (2000). An integrated threat theory of prejudice. In S. Oskamp (Ed.), Reducing prejudice and discrimination (pp. 23–46). Mahwah, N.J: Erlbaum. Stephenson, G. M. (1981). Intergroup bargaining and negotiation. In J. Turner & H. Giles (Eds.), Intergroup behaviour (pp. 168–198). Oxford: Blackwell. Tajfel, H. (1978). Interindividual and intergroup behaviour. In H. Tajfel (Ed.), Differentiation between social groups: Studies in the social psychology of intergroup relations (pp. 27–60). London: Academic Press. Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). California: Brooks & Cole. Tam, T., Hewstone, M., Kenworthy, J., Voci, A., Cairns, E., & Geddes, L. (2003). The mediational role of empathy and intergroup emotions in contact between Catholics and Protestants in Northern Ireland. Paper presented at Conference on Social Exclusion, University of Kent at Canterbury.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

341

Tam, T., Hewstone, M., Harwood, J., & Voci, A. (2004). Intergroup contact and grandparentgrandchild communication: Do self-disclosure, empathy, and reduced anxiety improve implicit and explicit attitudes toward older people? Manuscript under review. Tausch, N., Hewstone, M., Singh, P., Ghosh, E. S. K., & Biswas, U. N. (2004). Crosscommunity contact between Hindus and Muslims in India. Paper presented at the Meeting of the Society for the Psychological Study of Social Issues, Washington, DC. Tausch, N., Kenworthy, J. B., & Hewstone, M. (in press). The role of intergroup contact in the reduction of prejudice and the improvement of intergroup relations. In M. Fitzduff & C. M. Stout (Eds.), Psychological approaches to conflict and war. New York, NY: Praeger. Taylor, S. E. (1981). A categorization approach to stereotyping. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 83–114). Hillsdale, NJ: Erlbaum. Tropp, L. R. (2003). The psychological impact of prejudice: Implications for intergroup contact. Group Processes and Intergroup Relations, 6, 131–149. Tropp, L. R., (in press). Stigma and intergroup contact among members of minority and majority status groups. In S. Levin & C. van Laar (Eds.), Stigma and group inequality: Social psychological perspectives. Mahwah, NJ: Erlbaum. Tropp, L. R., & Pettigrew, T. F. (in press). Intergroup contact and the central role of affect in intergroup prejudice. In C. W. Leach & L. Tiedens (Eds.) The social life of emotion. Cambridge, UK: Cambridge University Press. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. (1987). Rediscovering the social group: a self-categorization theory. Oxford: Blackwell. Turner, R. N., Hewstone, M., & Voci, A. (2004). The impact of cross-group friendship on explicit and implicit prejudice in children. Manuscript submitted for publication. Urban, L. M., & Miller, N. M. (1998). A meta-analysis of crossed categorization effects. Journal of Personality and Social Psychology, 74, 894–908. Van de Vijver, F. J. R., Helms-Lorenz, M., & Feltzer, M. J. A. (1999). Acculturation and cognitive performance of migrant children in the Netherlands. International Journal of Psychology, 34, 149–162. van Dick, R., Wagner, U., Pettigrew, T. F., Christ, O., Wolf, C., & Petzel, T. (2004). The role of perceived importance in intergroup contact. Journal of Personality and Social Psychology, 87, 211–227. Van Oudenhoven, J. P., Groenewoud, J. T., & Hewstone, M. (1996). Cooperation, ethnic salience and generalisation of interethnic attitudes. European Journal of Social Psychology, 26, 649–661. Van Oudenhoven, J. P., Prins, K. S., & Buunk, B. P. (1998). Attitudes of minority and majority members towards adaptation of immigrants. European Journal of Social Psychology, 28, 995–1013. Varshney, A. (2002). Ethnic conflict and civic life: Hindus and Muslims in India. New Haven: Yale University Press. Vescio, T. K., Paolucci, M. P., & Hewstone, M. (2004). Tea and empathy: Empathy arousal and improved attitudes toward gay men. Manuscript under review. Vivian, J., Hewstone, M. & Brown, R. J. (1997). Intergroup contact: Theoretical and empirical developments. In R. Ben-Ari & Y. Rich (Eds.), Understanding and enhancing education for diverse students: An international perspective (pp. 13–46). Ramat-Gan: Bar-Ilan University Press. Voci, A., & Hewstone, M. (2003a). Contact and prejudice reduction in the Italian context: The impact of empathy, perspective taking, and group salience. Paper presented at the Meeting of the Society for Experimental Social Psychology, Boston, MA.

342

RUPERT BROWN AND MILES HEWSTONE

Voci, A., & Hewstone, M. (2003b). Intergroup contact and prejudice towards immigrants in Italy: The mediational role of anxiety and the moderational role of group salience. Group Processes and Intergroup Relations, 6, 37–54. Voci, A., & Hewstone, M. (2004a). Contact with immigrants among Italian workers: Effects of generalization and diversification. Paper presented at the Meeting of the Society for the Psychological Study of Social Issues, Washington, DC. Voci, A., & Hewstone, M. (2004b). La relazione tra contatto e riduzione del pregiudizio: Variabili di mediazione e moderazione. [The relationship between contact and the reduction of prejudice: Mediational and moderational variables.] Convegno Immigrazione, acculturazione e ipotesi del contatto [Conference on immigration, acculturation and the contact hypothesis]. Enna, Italy. Voci, A., Hewstone, M., Cairns, E., & McLernon, A. (2001). Intergroup contact and intergroup attitudes in Northern Ireland. EAESP Small-Group Meeting, Sardinia. Vonofakou, C., Hewstone, M., & Voci, A. (2003). Contact with out-group friends as a predictor of strength and accessibility of attitudes towards gay men. Postgraduate Psychology Conference, University of Cambridge, U.K. 22nd May. Wagner, U., Hewstone, M., & Machleit, U. (1989). Contact and prejudice between Germans and Turks: A correlational study. Human Relations, 42, 561–574. Waldzus, S., Mummendey, A., Wenzel, M., & Weber, U. (2003). Towards tolerance: Representations of superordinate categories and perceived ingroup prototypicality. Journal of Experimental Social Psychology, 39, 31–47. Walker, P. M., & Hewstone, M. (2004a). Intergroup experience and the own-race face effect. Manuscript submitted for publication. Walker, P. M., & Hewstone, M. (2004b). A developmental investigation of other-race contact and the own-race face effect. Manuscript submitted for publication. Watson, G. (1947). Action for unity. New York: Harper. Watt, S. E., Maio, G. M., Rees, K., & Hewstone, M. (2004). Functions of attitudes towards ethnic groups: Effects of level of abstraction. Manuscript submitted for publication. Weber, R., & Crocker, J. (1983). Cognitive processes in the revision of stereotypic beliefs. Journal of Personality and Social Psychology, 45, 961–977. Weber-Kollmann, R. (1985). Subtyping: The development and consequences of differentiated categories for stereotyped groups. Unpublished PhD, Northwestern University. Wegener, D. T., & Fabrigar, L. R. (2000). Analysis and design for nonexperimental data. Addressing causal and noncausal hypotheses. In H. T. Reis & C. M. Judd (Eds.), Handbook of research methods in social and personality psychology (pp. 412–450). Cambridge: Cambridge University Press. Werth, J. L., & Lord, C. G. (1992). Previous conceptions of the typical group member and the contact hypothesis. Basic and Applied Social Psychology, 13, 351–369. Wilder, D. A. (1984). Intergroup contact: The typical member and the exception to the rule. Journal of Experimental Social Psychology, 20, 177–194. Wilder, D. A. (1986). Social categorization: Implications for creation and reduction of intergroup bias. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 293–355). New York: Academic Press. Wilder, D. A. (1993). The role of anxiety in facilitating stereotypic judgements of outgroup behavior. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 87–109). San Diego: Academic Press. Wilder, D., & Simon, A. F. (1996). Incidental and integral affect as triggers of stereotyping. In R. M. Sorrentino & E. T. Higgins (Eds.), Handbook of motivation and cognition: The interpersonal context (Vol. 3, pp. 397–419). New York: Guilford Press.

AN INTEGRATIVE THEORY OF INTERGROUP CONTACT

343

Wilder, D., & Simon, A. F. (2001). Affect as a cause of intergroup bias. In R. Brown & S. L. Gaertner (Eds.), Blackwell handbook of social psychology: Intergroup processes (pp. 153–172). Malden, MA & Oxford, UK: Blackwell. Wilder, D. A., Simon, A. F., & Faith, M. (1996). Enhancing the impact of counterstereotypic information: Dispositional attributions for deviance. Journal of Personality and Social Psychology, 71, 276–287. Williams, R. M. (1947). The reduction of intergroup tensions. New York: Social Science Research Council. Williams, R. M. (1964). Strangers next door: Ethnic relations in American communities. Englewood Cliffs, NJ.: Prentice-Hall. Wolsko, C., Park, B., Judd, C., & Wittenbrink, B. (2000). Framing interethnic ideology: Effects of multicultural and color-blind perspectives on judgments of groups and individuals. Journal of Personality and Social Psychology, 78, 635–654. Wolsko, C., Park, B., Judd, C. M., & Bachelor, J. (2003). Intergroup contact: Effects on group evaluations and perceived variability. Group Processes and Intergroup Relations, 6, 93–110. Wright, S. C., Aron, A., McLaughlin-Volpe, T., & Ropp, S. A. (1997). The extended contact effect: Knowledge of cross-group friendships and prejudice. Journal of Personality and Social Psychology, 73, 73–90. Zagefka, H., & Brown, R. (2002). The relationship between acculturation strategies, relative fit and intergroup relations: Immigrant-majority relations in Germany. European Journal of Social Psychology, 32, 171–188. Zagefka, H., Brown, R., Broquard, M., & Leventoglu, S. (2002). Acculturation preferences, fit, intergroup attitudes and economic competition. Paper presented at the BPS Social Psychology Section Conference, Huddersfield.

This Page Intentionally Left Blank

SAYS WHO?: EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

Arie W. Kruglanski Amiram Raviv Daniel Bar-Tal Alona Raviv Keren Sharvit Shmuel Ellis Ruth Bar Antonio Pierro Lucia Mannetti

This chapter features the concept of ascribed epistemic authority (Kruglanski, 1989) oVered as a unique perspective on source eVects in social judgment. In contrast to prior approaches that viewed the source of communication as external to the self, we assume that both the self and external sources may be assigned diVerent degrees of epistemic authority in diVerent domains and that this determines how individuals process information, make decisions, and undertake actions. The present framework traces the socio-developmental aspects of epistemic authority assignments and considers individual diVerences in the distribution of authority assignments across sources. From this perspective, we claim a central role in human judgment to the information’s source, and the assessment of its epistemic authority is seen to constitute an essential preliminary phase in individuals’ approach to information.

345 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 37

Copyright 2005, Elsevier Inc. All rights reserved. 0065-2601/05 $35.00

346

ARIE W. KRUGLANSKI et al.

I. Introduction A critical aspect of human social functioning concerns people’s informational dependence on others (Kelley & Thibaut, 1969). As we negotiate our way through the labyrinths of interpersonal relations and task exigencies, we encounter a continuous flow of information in the form of communications, advice, exhortations, and pleas from a variety of sources. These pose the ubiquitous question of whom to (informationally) trust and whose statements to discount or regard with suspicion. The issue has profound implications not only for one’s personal dealings with one’s social and physical environments but also for the workings of society itself. In this vein, the political scientist Robert Putnam (2000) commented on the declining trust in government characteristic of contemporary American society and on the dangers this poses to its communal functioning. In other words, for many Americans the government has ceased to represent an informational source to be relied on, encouraging political disengagement and civic apathy. The significance of source variables for individuals’ reactions to the information given has not been lost on social psychologists. To the contrary, such variables have figured prominently in major theories of communication and persuasion beginning with the classic works of the Yale group (Hovland, Janis, & Kelley, 1953; Hovland & Weiss, 1951). This continues with the currently influential dual-mode theories of persuasion (Chaiken, Lieberman, & Eagly, 1989; Petty & Cacioppo, 1986) and the parametric unimodel (Erb, Kruglanski, Chun, Pierro, Mannetti, & Spiegel, 2003; Kruglanski & Thompson, 1999a,b; Pierro, Mannetti, Kruglanski, & Sleeth-Keppler, 2004). Yet these various treatments have been limited in scope and have not addressed the full array of issues that source characteristics may raise in the domain of social judgment. The present article introduces a broader perspective on source eVects framed from the subjective standpoint of the information’s recipient. We first review the treatment of source eVects in several major models of persuasion. We then introduce the present framework focused on the concept of ‘‘epistemic authority’’ and review empirical research conducted under its aegis. A final discussion highlights the unique properties of the present approach and considers its implications for the place of source eVects in notions of information processing and human judgment.

II. The Early Work: Hovland et al.’s (1953) Learning Theory Paradigm of Source EVects Hovland et al.’s (1953) conception of persuasion and attitude change rested on the learning theory paradigm dominant at the time. From that

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

347

perspective, communicators’ influence derives from the incentives they provide. These can be tangible or intangible and based on such source characteristics as expertness, trustworthiness, similarity, or likability. In turn, these (and other possible characteristics) are learned through experiencing the consequences of accepting or rejecting a given source’s influence and are generalized to other similar sources. In general, Hovland et al. (1953) suggest that the learned characteristics of the source determine individuals’ motivation to attend to and comprehend the message, as well as accept its implications (McGuire, 1969). Within this conceptual framework, Hovland et al. (1953) focused in particular on the variable of communicator credibility, which they parsed into its two constituents, expertness and trustworthiness, the latter referring to the communicator’s intent to convey valid information. Early empirical studies (e.g., Hovland and Weiss, 1951) indeed showed that messages ascribed to high-credibility sources tend to be accepted to a greater extent than ones ascribed to sources of low credibility. An issue that occupied researchers of source eVects concerned a separation and comparison of the two components of credibility, namely expertise and trustworthiness (Cohen, 1964; Hovland et al., 1953). In an empirical investigation of this problem, Kelman and Hovland (1953) exposed high school students to a communication advocating a more lenient treatment of juvenile delinquents. This communication was attributed to one of three diVerent sources: (1) a well-informed and intentioned judge, i.e., a source possessing a high degree of expertise as well as of trustworthiness, (2) a member of the studio audience, i.e., a source of intermediate presumed expertise and trustworthiness, and (3) a juvenile delinquent on bail, charged with drug dealings and other shady business, i.e. a source of high presumed familiarity (and in this sense ‘‘expertise’’) in the domain, yet of low trustworthiness. As might be expected, the message delivered by the judge had significantly greater persuasive impact than one delivered by the juvenile delinquent with the studio audience member’s message exerting an intermediate impact, closer in magnitude to that of the judge than to the delinquent. Hovland et al. (1953) interpreted these findings as suggesting that trustworthiness may play a greater role in attitude change than expertness. Subsequent research confirmed the significant role that the communicator’s intent may play in attitude change (e.g., Allyn & Festinger, 1961; Weiss & Fine, 1955, 1956). Though the distinction between expertise and ‘‘trustworthiness’’ (understood as the intent to speak the truth) is of interest, the attempt to separate the two and pit them against each other could be misleading. For the question always turns on whom to (informationally) trust (the ubiquitous ‘‘says who?’’ question in the title of this chapter). In this sense expertise, intent, as well as a possible host of other learned source characteristics

348

ARIE W. KRUGLANSKI et al.

(e.g., membership in one’s in-group, age, gender) drive the acceptance of communications through the overall variable of credibility.1 In our subsequent analysis, we refrain from analytically disentangling the diVerent source characteristics that could contribute to a source’s epistemic authority, and we instead address the attribution of such authority as a Gestalt.

III. Source EVects from the Dual-Mode Perspective: The Elaboration Likelihood Model An important shift in the way persuasion eVects in general and source eVects in particular were approached by persuasion researchers was occasioned by the work of Petty and Cacioppo (1986) on the Elaboration Likelihood Model (ELM) and Chaiken’s (1979) work on the Heuristic-Systematic Model (HSM). Conceptually, this shift signaled a break from the neobehavioristic paradigm adopted by Hovland et al. (1953) and a move toward the cognitive Zeitgeist as part of the ‘‘cognitive revolution’’ that has been transforming the psychological science as a whole. Specifically, both of these dual-mode theories regarded persuasion as essentially a cognitive activity wherein conclusions are reached on the basis of more or less extensive processing (or ‘‘elaboration’’) of various types of information available in the persuasive context. Besides its novel conceptual perspective on source eVects, the dual-mode paradigm was based on the recognition that prior research on this topic yielded a disappointing crop of empirical findings. As Petty and Cacioppo (1986, p. 125) put it: ‘‘. . . although it might seem reasonable to propose that by associating a message with an expert source agreement could be increased (e.g., see Aristotle’s Rhetoric), the accumulated contemporary research literature suggested that expertise eVects were considerably more complicated . . . Sometimes expert sources had the expected eVects . . . sometimes no eVects were obtained . . ., and sometimes reverse eVects were noted. Unfortunately, the conditions under which each of these eVects could be obtained and the processes involved in producing these eVects were not at all apparent . . .’’ In the ELM, source characteristics are often considered as ‘‘cues,’’ that is, as ‘‘peripheral’’ signals likely to exert an on impact persuasion under conditions of a low ‘‘elaboration’’ likelihood, that is, in circumstances where an 1

For instance, one might evolve a lay theory that even well-intentioned experts cannot be trusted because of their arrogance and tendency to operate through the blinders of their misconceptions.

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

349

individual’s motivation and/or capacity to process information are low. For instance, in the classic study by Petty, Cacioppo, and Goldman (1981), source expertise was manipulated orthogonally to message argument quality. Cross-cutting both, the researchers manipulated participants’ issue involvement, considered as a determinant of elaboration likelihood. It was found that source expertise had persuasive eVects only under conditions of low issue involvement but not under conditions of high issue involvement. By contrast, message argument quality had persuasive eVects only under high, but not under low, issue involvement. These findings were interpreted to demonstrate the cue function that source characteristics may often fulfill, and according to which they may exert impact under low-elaboration (but not under high-elaboration) likelihood conditions. It is noteworthy that in the ELM, source characteristics may fulfill functions other than that of a ‘‘cue.’’ In particular, they may constitute ‘‘message arguments,’’ such as when a physically attractive person advertises a beauty product (her or his appearance serving as proof of the product’s eVectiveness) (Petty & Cacioppo, 1986). Finally, according to the ELM, source characteristics may serve a motivating function, prompting extensive or restricted elaboration of the information. Knowing that a source is expert or that he/she represents a majority position (Mackie, 1987) may motivate recipients to pay close attention to its message and to take it seriously. Similarly, knowing that the source is inexpert might reduce the recipients’ consideration of the message.

IV. Source EVects in the Heuristic Systematic Model In the HSM, source characteristics are regarded as ‘‘heuristic’’ information, related to simple and general rules of thumb or ‘‘heuristics,’’ such as ‘‘expertise implies correctness,’’ ‘‘friends can be trusted,’’ or ‘‘majority opinions are valid.’’ Knowing that a source is a friend, an expert, or member of a majority may then prompt an acceptance of her or his recommendations through an application of the corresponding heuristic. Much like the ELM, the HSM too distinguishes between two separate and qualitatively distinct modes of persuasion: the heuristic mode depicted above (likely to be adopted when the recipients’ cognitive and motivational resources are limited) and a systematic mode in which the message information is carefully considered (likely to be adopted when the recipients’ cognitive and motivational resources are ample) (Chaiken et al., 1989). Furthermore, the heuristic mode is assumed to aVord lesser judgmental confidence than the systematic mode; hence it is assumed to be opted for when lower confidence threshold is

350

ARIE W. KRUGLANSKI et al.

deemed suYcient for the recipient’s purposes. When it is insuYcient, that is, when the issue is important enough to require considerable confidence, the systematic mode is assumed to ‘‘kick in’’ instead.

V. Source EVects in the Unimodel Both the ELM and the HSM stress the qualitative distinctions between the two modes of persuasion (i.e., the peripheral and central modes in the ELM and the heuristic and systematic modes in the HSM). By contrast, a recently proposed ‘‘unimodel’’ (Erb et al., 2003; Kruglanski & Thompson, 1999a,b; Kruglanski, Fishback, Erb, Pierro, & Mannetti, 2004; Pierro et al., 2004; Chun, Spiegel, & Kruglanski, 2002) stress the commonalities between those modes. In these terms, both ‘‘peripheral cues’’ and ‘‘message arguments’’ function as evidence for various conclusions that recipients may draw from information in the persuasive context. Some such conclusions may refer to actual recommendations espoused by the communication source or flowing from the message arguments. Other conclusions may relate to the desirability of paying close attention to the source’s pronouncements, e.g., because of its expertise, prestige, or power (Fiske, 2004) versus ignoring them because of the source’s low standing on these particular dimensions. According to the unimodel, the evidential function of information derives from a fundamental syllogistic structure. For instance, a person may subscribe to a (major) premise that ‘‘experts are correct’’ (or to the belief that ‘‘if expert, then correct’’) and then assume that a position expressed by a source was valid because ‘‘she is an expert’’ (a minor premise). Similarly, a person may assume that ‘‘anything that promotes a clean environment should be supported’’ (a major premise) and then proceed to accept the statement that ‘‘electrically powered automobiles should be supported’’ because of the belief that ‘‘electrically powered automobiles preserve a clean environment’’ (a minor premise). Note that, whereas the former statement about expertise represents a peripheral or heuristic ‘‘cue,’’ and the latter represents a ‘‘message argument,’’ both are seen to function in a fundamentally identical syllogistic manner. The same would be true of a case where a source’s characteristic, e.g., physical attractiveness, functioned as a ‘‘message argument,’’ that is, as a ‘‘minor premise’’ that in conjunction with a major premise whereby ‘‘physical attractiveness of a cosmetic product’s user attest to its eVectiveness’’ yields the conclusion that the product, indeed, is eVective. Finally, the case where a source characteristic (e.g., ‘social power’) prompts an individual to closely attend to the source’s argument may be

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

351

based on a conviction (major premise) that ‘‘the views of powerful persons are worth considering.’’ Thus, the unimodel stresses the similarities between source and message eVects based on their similar syllogistic functioning as evidence for various conclusions.

A. ON THE UNIQUENESS OF SOURCE EFFECTS: THE CONCEPT OF EPISTEMIC AUTHORITY In contrast to the foregoing emphasis on commonalities that source eVects may share with other ‘‘peripheral cues’’ (in the ELM), other ‘‘types of heuristic information’’ (in the HSM) or other minor premises (in the unimodel) the present chapter highlights ways in which source eVects are unique. As shown subsequently, this framework oVers a novel perspective on source eVects, aVording fresh insights and ways of thinking about this important domain of phenomena. Our central construct is that of ‘‘epistemic authority.’’ This construct, originally proposed by Kruglanski (1989), denotes a source on whom an individual may rely in her or his attempts to acquire knowledge on various topics. This concept is akin to the omnibus notion of source’s credibility (apart from the source characteristics, such as expertise or intent, that may imply credibility to diVerent persons), and it addresses the extent to which an individual is prepared to rely on a source’s information and to accept it. The epistemic authority of various sources may vary and the authority of a given source may vary across domains as well as across lifespan developmental phases. The characteristics that serve as the basis for identifying a source as an epistemic authority can be general, like seniority (for example, of an elder), a role (for example, of a priest, a leader, or a teacher), or level of education (for example, of a Ph.D degree holder), an appearance in print (for example, in a book or a newspaper), or specific, as in assigning epistemic authority to a particular person or to a particular newspaper. A source may exert influence in numerous life domains, serving as a generalized epistemic authority; alternatively; it may influence only a specific area (for example, cardiology, statistics, or auto mechanics) wherein it is considered to possess valid knowledge. In the former role we may find priests, therapists, or parents, whereas in the latter we may find specialists in certain well-defined fields. Individuals may diVer widely in their reliance on various epistemic authorities and in their extent of such reliance across domains. Whereas some people may accept the judgment of a source (a rabbi, a priest, a psychiatrist, or a teacher) on any issue in their lives, hence treating him/her as a generalized epistemic authority, others may consult with a source with regard to more circumscribed issues related to the

352

ARIE W. KRUGLANSKI et al.

source’s specific domain of competence, and to seek knowledge from other specific sources in other life-domains. In counterdistinction to the dual-process models that viewed sources’ characteristics as somewhat inferior in persuasive impact to message arguments2, in the present framework epistemic authority of some sources (e.g., religious prophet, parent, political leader, or the printed word) might be extremely powerful—so powerful, in fact, that it may override all else and exert a determinative influence on the individual’s judgments and correspondent behavior. In Kruglanski’s (1989) lay epistemic theory, epistemic authority functions as a ‘‘stopping mechanism’’ analogous in its eVects to the need for closure (Kruglanski, 2004; Kruglanski & Webster, 1996; Webster & Kruglanski, 1998): In the same way that a heightened need for cognitive closure may eVect a ‘‘seizing and freezing’’ on a given judgment, so may a particularly high degree of epistemic authority conferred upon a source (e.g., an adored political leader, an admired scientist, or a revered religious pontiV). Thus, even though individuals’ accuracy motivation may be high and their cognitive resources ample, they may discontinue their epistemic search and instead accept (i.e., ‘‘seize and freeze’’ upon) the pronouncement of a high-authority source, whose statements simply are perceived as beyond reasonable doubt. In its role of a ‘‘stopping mechanism,’’ a source’s authority plays, therefore, a motivational role in information processing determining the amount of energy the individual is prepared to devote to continued epistemic activity in a domain. Finally, whereas in prior treatments of source credibility eVects, the discussion centered on the credibility of sources external to the self (cf. Chaiken et al., 1989; Hovland et al., 1953; Kruglanski & Thompson, 1999a,b; Petty & Cacioppo, 1986), i.e., of communicators whose messages one may be exposed to, the present theory considers the self as a particularly important target of epistemic authority assignments and investigates how these may change in the course of individuals’ social-emotional development. In accordance with the foregoing analysis, our analytic framework focuses on (1) the unique developmental aspect of source eVects, (2) assessment of stable individual diVerences in source eVects, and (3) the role of self as a source of information.

2

In the ELM, for instance, persuasion accomplished via such peripheral cues as source expertise is assumed to be less persistent, less resistant to counterpersuasion, and less linked to behavior than persuasion based on message arguments; Similarly, in the HSM, persuasion based on heuristics is assumed to aVord lower confidence and to be resorted to when the ‘‘suYciency threshold’’ for confidence is low.

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

353

1. Social Development of Epistemic Authority One way in which source authority is special relates to its inevitable developmental dimension. To the newborn baby, the world may well present a ‘‘buzzing confusion,’’ to use William James’ famous turn of phrase. Whereas perceptual clarity may be gaining in sharpness as a matter of physiological maturation, the development of conceptual knowledge may depend to a large extent on the child’s adult caregivers, in most instances its parents, and other family members that interpret for the child what diVerent informational patterns may mean and in what conceptual categories they should be apprehended. Indeed, developmental psychologists have long recognized (Erikson, 1950; Freud, 1965; Kagan, 1958, Ruble & Seidman, 1996; Spiegel, 1958; Staton, 1963; Winnicott, 1965) that the young child’s world-knowledge is shaped by the various ‘‘socialization agents’’ or ‘‘significant others’’ with whom it is in frequent contact. It seems plausible to assume then, that for the young child these adult caregivers may represent the ultimate epistemic authority, whose views on various matters are accepted as the indisputable truth. In the course of normal socio-cognitive development, the epistemic authority of the parents may wane, and that of other sources may increase. For instance, one’s schoolteachers are likely to be assigned some specific authority in their domains of competence (see, e.g., Raviv, Bar-Tal, Raviv, Biran, & Sela, 2003). Thus, a diVerentiation begins wherein the semiabsolute authority of one’s caregivers is gradually replaced by a more distributed pattern of attributions in which diVerent sources are relied on for diVerent types of knowledge. In adolescence, for example, the peer group typically is assigned considerable epistemic authority. This may represent a way in which the child is beginning to assert her or his unique identity, independent of its parents. In other words, bestowing respect and admiration on figures similar to oneself (i.e., one’s peers) could be a way of indirectly asserting one’s own self-worth as a source of knowledge. More directly, too, the self gradually gains in ascribed epistemic authority at least in some domains. One begins to trust one’s ability to make sense of various types of information and to draw valid conclusions from the available evidence. A normally functioning adult may acquire a balanced mental representation of epistemic authorities in various regions of her or his life space, having developed good ideas about whom to trust with respect to what, when, and how much. The development of such a representation may be aVected and distorted by the individual’s unique history of socialization. For instance, attachment theorists (see Mikulincer & Shaver, in press) have found that individuals with secure attachment tend to trust external sources (other people) more than do individuals with anxious or avoidant attachment, and hence bestow them with higher epistemic authority (Collins & Read, 1990;

354

ARIE W. KRUGLANSKI et al.

Mikulincer, 1998). Moreover, individuals with avoidant attachment tend toward what Bowlby (1988) has called ‘‘compulsive self-reliance,’’ that is, an exaggerated ascription of epistemic authority to the self (Mikulincer & Shaver, 2003). As a consequence of their unique developmental course, as well as unique socialization and learning experiences, adult individuals may vary considerably in their distributive profiles of epistemic authority (i.e., in the epistemic authority they may assign to various sources). The shape of such distributive profiles may be determined by various informational and motivational factors. In other words, whereas the young infant may instinctively orient toward adult caregivers as epistemic authorities, the subsequent source diVerentiation is probably governed by the same informational and motivational factors that determine the formation of all beliefs (Kruglanski, 1989). As far as information is concerned, sources may ‘‘prove’’ their knowledgeability in a domain in various ways, e.g., by making predictions and suggestions that pan out. Subsequently, trusted epistemic authorities may ‘‘confer authority’’ on other sources (as when a knowledgeable friend recommends an expert in a domain, e.g., a physician, a car mechanic, or a professor). As for motivation, a source that gratifies some of the individual’s needs may be perceived as capable of gratifying other needs as well, including the epistemic needs of knowledge formation. For instance, one might be motivated to view as an epistemic authority, and to accept the influence of, a caring parent, fostering a secure attachment (Bowlby, 1988) on the basis of the desire to believe that a parent who wishes one well is knowledgeable and right. For very diVerent reasons, the seeking of approval, one may bestow epistemic authority upon a distant and uncaring parent on the basis of the hope that such admiration will ensure the parent’s attention and goodwill. Though these two motivations might function similarly in the bestowal of authority on one’s parents, they may markedly diVer with regard to the development of one’s own authority. The aVection of a caring parent may be perceived as unconditional, allowing the child to explore its environment and develop its own sense of epistemic authority. By contrast, the (hoped for) attention of a distant parent may appear as contingent on one’s adoration of the parent; this may encourage the clinging to the parent’s ascribed authority and a reluctance to develop one’s own epistemic independence, potentially threatening the loss of the parent’s approval. 2. Further Determinants of Epistemic Authority Assisgnments Beyond aspects of one’s socio-emotional development, various motivational and cognitive factors may play a role in shaping the profile of one’s epistemic authorities. For instance, the need for cognitive closure

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

355

(Kruglanski & Webster, 1996; Webster & Kruglanski, 1998) may prompt one to diVerentiate sharply among epistemic authorities assigned to various sources. Accordingly, persons with a high need for closure may attribute either a more extremely high or a more extremely low epistemic authority to a given source than persons with a low need for closure. For reasons of closure, too, one might confer a higher degree of epistemic authority on individuals whose opinions are similar to one’s own, allowing one to preserve one’s closure (Kruglanski & Webster, 1996). Such relations might be complicated by one’s view of oneself as an epistemic authority: If one’s self-perception as an epistemic authority was high, agreement from another would be self-validating, in addition to safeguarding closure. This would render agreement highly motivationally desirable, increasing the likelihood of conferring a high degree of epistemic authority on the agreeing other. By contrast, if one’s self-perception as an epistemic authority was low, agreement from another would tend to invalidate one’s (negative) selfconcept while safeguarding one’s closure, hence possibly evoking ambivalence and a resistance to conferring a high epistemic authority on the agreeing other. Finally, because motivations as well as information can be induced situationally, the perception of epistemic authority too may vary across situations. For instance, source statements that the recipient perceived as intelligent and compelling may elicit ascriptions of higher epistemic authority than statements perceived as less intelligent or compelling (Erb et al., 2003). Similarly, momentarily induced motivations may aVect the assignments of epistemic authority. For instance, momentarily induced need for cognitive closure (De Grada, Kruglanski, Mannetti, & Pierro, 1999) may increase the diVerentiation between sources in terms of perceived epistemic authority, correspondingly increasing the influence of some sources and lowering the influence of other sources. As with other beliefs, then, the assignment of epistemic authority may involve the joint influence of informational and motivational factors. Thus, the mere presence of relevant information may not suYce to produce an impression of epistemic authority. In addition to the information being ‘‘given,’’ one would need to be motivated to ‘‘take it,’’ that is, to use that information to form such an impression (Mazliach-Liberman, 2003). Thus, the construct of epistemic authority has both nonunique and unique aspects. On the nonunique side, epistemic authority ascribed to a source represents a meta-cognitive belief about the degree to which it can be relied on for veridical opinions about some aspects of the world. In that sense, epistemic authority beliefs are like all other beliefs both in terms of the ways in which they are formed (based on the impact of motivational and informational conditions of their construction) and in terms of the ways they

356

ARIE W. KRUGLANSKI et al.

function (Erb et al., 2003) to aVord the drawing of inferences and the arrival at conclusions. Nonetheless, epistemic authority beliefs are in an important sense quite special and unlike other possible beliefs. That is true because they are a quintessential part and parcel of our social nature. Epistemically, they are inextricably tied to our informational dependence on other people and to the relative independence of others we may acquire while growing up. Epistemic authority beliefs are developmentally fundamental, and they constitute part and parcel of the infant’s fundamental orientation toward its adult caregivers. And in adult individuals, a diVerentiated hierarchy of epistemic authorities assigned to various sources may constitute a stable system of epistemic guides indispensable to people’s self-regulation toward their objectives. Thus, the topic of epistemic authority may be fruitfully approached from developmental, diVerential, as well as situational perspectives. 3. Epistemic Authority EVects on Information Processing and Decision Making We are presently assuming that an assessment of the epistemic authority of sources present in a given situation constitutes a preliminary phase in all information processing. For example in a psychological experiment, by delivering the instructions to participants, the experimenter is implying that the latter possess suYcient epistemic authority to make sense of the instructions. By contrast, during a medical consultation among physicians, the patient is usually left out of the ‘‘loop,’’ the implication being that he or she does not have suYcient epistemic authority to be part of the conversation. Subsequently, the patient’s primary physician may often present a ‘‘watered down’’ version of the conversation to the patient, which the patient presumably has suYcient self-ascribed epistemic authority to evaluate. We are assuming that once the individual has determined (implicitly) who in a given situation has the dominant epistemic authority for making sense of ‘‘the information given,’’ this source is (1) sought out and (2) attended to. Furthermore, (3) its conclusions tend to be accepted, resulting in (4) considerable epistemic confidence on the part of the recipient and (5) leading to behavioral decisions consistent with the source’s recommendations. In the remainder of this chapter, we describe some initial empirical research related to our conceptual analysis. A major innovative aspect of our proposal is that a distributional profile of epistemic authorities across diVerent knowledge domains can vary developmentally as well as individually and that it can pertain to the self as well as to others. Furthermore, because epistemic authority beliefs may determine the influence that various sources exert on individuals’ judgment and decisions, mapping out such beliefs in a

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

357

specific context can be of considerable applied significance (e.g., in the realms of education, politics, or economics). In what follows, we consider some ways in which these notions have been studied thus far and the evidence that has accumulated for the utility of the epistemic authority notion as a social psychological construct.

VI. Empirical Research in the Epistemic Authority Framework A. DEVELOPMENTAL DIFFERENCES: WHO KNOWS BEST? Prior developmental work (e.g., Berndt, 1979; Floyd & South, 1972; Hartup, 1983; Wintre, Hicks, McVey, & Fox, 1988) has primarily focused on confronting peers’ versus parents’ influence and did not use direct measures of ascribed epistemic authority. Our own research, based on specific such measures, included a more varied array of potential sources including the self. In an early study, Raviv, Bar-Tal, Raviv, and Houminer (1990a) assessed children’s attribution of epistemic authority to their mothers, fathers, teachers, and friends. They investigated kindergarten children (4–5 years old), first graders (6–7 years old), and third graders (8–9 years old). To that end they used a questionnaire in a semistructured interview. In the first phase of questionnaire construction, researchers attempted to determine what verb best represents to children the concept of ‘‘epistemic authority.’’ Preliminary studies related to this research revealed that kindergarten children do not understand the meaning of the phrase ‘‘to rely on,’’ that they tend to invest the verb ‘‘to trust’’ with moral significance, and that they respond to the question ‘‘whom do you ask?’’ on the basis of the source’s availability. It was furthermore determined that the question ‘‘who knows best’’ avoids these pitfalls and expresses most accurately the perception of epistemic authority. Consequently, the epistemic authority of each of the four sources was evaluated in seven knowledge areas, again gleaned on the basis of extensive pretests. These areas were: (1) pastime (e.g., tapped in items such as ‘‘who knows best, in your opinion, what games is it fun to be playing in the afternoon?’’), (2) social relations (e.g., ‘‘who knows best, in your opinion, whether you should be friends with a certain boy or girl?’’), (3) rules and laws (e.g., ‘‘who knows best, in your opinion, things that are allowed and forbidden to do?’’), (4) personal feelings (e.g., ‘‘who knows best, in your opinion, what you should do when you are sad or when you are insulted?’’), (5) science (e.g., ‘‘who knows best, in your opinion, all sorts of interesting things, such as why the sun disappears at night or why it is cold and rainy in the winter?’’), (6) future planning (e.g., ‘‘who knows best,

358

ARIE W. KRUGLANSKI et al.

in your opinion, how to help you to decide what you will be when you grow up?’’), and (7) physical appearance (e.g., ‘‘who knows best, in your opinion, what clothes you should wear when you want to get dressed up?’’). Reponses were given on a four-point scale. Several significant trends appeared in these data, yielding the following pattern of interest: during childhood (i.e., during the ages 4–10), (a) the perception of parents as epistemic authorities remains relatively stable, with decreases in a few knowledge areas, (b) the perception of the teacher as an epistemic authority remains stable with an increase in the area of science, (c) the perceived epistemic authority of friends increases in the social domain. Detailed results for the domains of Pastime and Science are presented in Fig. 1A, B. Raviv et al. (1990a) also looked at the diVerentiation in perceived epistemic authority of various sources across domains of knowledge and as a function of age. Four diVerentiation scores were defined for each participant such that for each source the standard deviation was computed between rankings given this source in the seven knowledge areas. The results indicated that, across the age groups, the diVerentiation in epistemic authority assigned to Father and Mother was less than that assigned to Teachers and Friends. That is, the perception of teachers and friends varied more as a function of knowledge areas than the perception of parents. The children

Fig. 1. Mean epistemic authority of various sources in the (A) domain of pastime and (B) domain of science for 3 age groups (Raviv et al., 1990a).

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

359

selected teachers and friends as epistemic authorities in certain knowledge areas only, whereas the parents were perceived more as overall authorities across domains, possibly as a function of continued material dependence on the parents inducing a motivation to view them as all-powerful in the epistemic, and probably also in other, realms. Two studies investigated the changes in perception of epistemic authorities between the ages of 9 and 18 (Bar-Tal, Raviv, & Raviv, 1991a; Raviv, Bar-Tal, Raviv, & Peleg, 1990b). Raviv et al. (1990b) investigated three groups of children/adolescents, fourth graders (ages 9–10), eighth graders (ages 13–14), and twelfth graders (ages 17–18). As in the research described earlier, the targeted sources consisted of mothers, fathers, teachers, and friends. Their epistemic authority was evaluated separately in regards to the following nine knowledge areas: (1) school studies, (2) politics, (3) science, (4) pastime, (5) physical appearance, (6) social relations, (7) future planning, (8) values, and (9) personal feelings. The opening question regarding each source was, for instance, ‘‘To what extent do you rely3, that is, you believe and trust, what your mother tells you on the following subjects?’’ The same question was presented with reference to ‘‘your father,’’ ‘‘your teacher,’’ and ‘‘your friends.’’ Then, for each source, nine questions followed, pertaining to each of the nine knowledge domains, for instance, ‘‘To what extent do you rely on the things your mother says regarding your school studies (for example, preparation of homework or how to study and prepare for examinations?’’) Answers were recorded on a five-point scale anchored at the ends with (1) ‘‘do not rely’’ and (5) ‘‘rely’’ with the intermediate points labeled as (2) ‘‘generally do not rely,’’ (3) ‘‘sometimes rely and sometimes not,’’ and (4) ‘‘generally do not rely.’’ Based on appropriate factor analyses, the original nine areas were integrated into three general domains of general interest to investigators of adolescent behavior (e.g., Sebald, 1986) namely those of Formal Knowledge (including the specific areas of School Studies, Politics, and Science), Social Knowledge (including Pastime, Physical Appearance, and Social Relations), and General Life Knowledge (including the specific areas of Future Planning, Values, Personal Feelings, and General Knowledge). The results showed that for all the sources, and across all three domains, the fourth graders assigned greater epistemic authority to sources than the older two age groups. The only exception to this trend existed for the Social Knowledge domain of Friends, where no significant diVerences obtained across the age-groups. It is also of interest that in the domains of Formal 3

The term ‘‘rely on’’ was used after a pretest had indicated that youngsters of the ages investigated took the phrase ‘‘to rely on someone’s knowledge’’ to mean ‘‘to believe in’’ and ‘‘to trust,’’ hence in present terms to regard as an epistemic authority.

360

ARIE W. KRUGLANSKI et al.

Fig. 2. Mean epistemic authority of various sources in the (A) domain of formal, (B) domain of social and (C) domain of general life knowledge for 3 age groups (Raviv et al., 1990b).

Knowledge and General Life Knowledge, adolescents continued to consider one or both parents as the single most important epistemic authority. For example, in the domain of Formal Knowledge, and across the three age groups, Father received higher scores than Mother and Teachers, both of whom received higher scores than Friends (see Fig. 2A). In the twelfth Grade, Teachers received a higher score on Formal Knowledge than Mother, who in turn was evaluated higher than Friends. In the realm of General Knowledge too, it appears that Father and Mother remain the most important epistemic authorities throughout adolescence (see also Hunter, 1985). As for Friends, they received the highest relative scores in the Social Knowledge domain for all three age groups. Also, across the age groups, teachers received their highest score in the Formal Knowledge domain, followed by

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

361

the General Knowledge domain, and the Social Knowledge domain, where their relative scores were the lowest. In the same study, Raviv et al. (1990b) investigated the diVerentiation in epistemic authority assigned each source across domains and as a function of age. A diVerentiation score in this case was defined as the standard deviation of evaluations given to the nine content categories with regard to each source. The overall diVerentiation diVered significantly by source, the Teacher being the most diVerentiated of all the four sources considered. Furthermore, the order of source diVerentiation varied with age (see Fig. 3). Thus, for the fourth graders, the order of diVerentiation from highest to lowest was Teachers, Friends, Father, and Mother. The order for the eighth graders was Teachers, Father, Friends, and Mother, and for the twelfth graders it was Teachers, Father, Mother, and Friends. These findings imply that, whereas for the fourth graders, Parents were perceived as sources of generalized epistemic authority; for the twelfth graders, such generalized authority was reserved for Friends. Fourth graders (ages 9–10), eighth graders (ages 13–14), and twelfth graders (ages 17–18) were also investigated in research by Bar-Tal, Raviv, Raviv, and Brosh (1991b). This study went beyond the Raviv et al. (1990) research in two important respects: (1) It included the self among the investigated knowledge sources, phrased in terms of questions with reference to ‘‘Myself,’’ and (2) it tapped participants’ stated reasons for epistemicauthority assignments. The main result of interest was that in the eighth and twelfth grades the importance of Myself as epistemic authority was greater than the epistemic importance of all the other sources (see Fig. 4). By contrast, in the fourth grade, Myself was not assigned a higher epistemic authority than Mother and Father. Overall, whereas the epistemic authority of both the Father and the Mother declines across the age groups, the epistemic authority assigned to Myself significantly increases.

Fig. 3. Mean epistemic authority diVerentiation scores of diVerent sources for three age groups (Raviv et al., 1990b).

362

ARIE W. KRUGLANSKI et al.

Fig. 4. Mean number of domains in which a given source was selected as a top epistemic authority for three age groups (Bar-Tal et al., 1991b).

The perception of Friend as an epistemic authority increased with age in the domain of School Work, Physical Appearance, and Personal Feelings. The perception of Teacher as an epistemic authority increased with age in the domain of School Work but decreased in the domain of Science. In the latter domain, two new sources appeared, Expert and Media, which become especially important in the twelfth grade. As for the stated reasons for epistemic authority assignments, Father, Teacher, and Media were selected as epistemic authorities mainly because of their Knowledge, Myself was selected mainly because of Familiarity (the source knows the person, his/her taste, and what he/she likes or wants), Mother was selected mainly because of Helpfulness (the source wants to help, wants to benefit, and is empathetic), and Friend was selected because of Helpfulness, Knowledge, and Similarity (the source is similar to the participant in traits, attitudes or age). It is noteworthy that these reasons are classifiable into those tapping the source’s motivation to provide valid knowledge (Helpfulness) and its ability to do so in terms of having the relevant background knowledge and evaluative criteria (Knowledge, Familiarity, and Similarity) and more generally recall the distinction between ‘‘trustworthiness’’ and ‘‘expertise’’ drawn by Hovland and Weiss (1951).

B. POP IDOLS AS EPISTEMIC AUTHORITIES The work by Bar-Tal et al. (1991b) demonstrates the emergence of novel sources of epistemic authority in the course of social development, such as of Experts and of Media. Whereas Experts and Media may constitute stable sources of epistemic authority throughout one’s adult life, other novel epistemic sources may be more transient in nature, their waxing and waning reflecting the vicissitudes of shifting developmental phases. An intriguing case of such a transiency is a study of the epistemic authority assigned to

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

363

pop singers (Raviv, Bar-Tal, Raviv, & Ben-Horin, 1996). Participants, male and female adolescents from three age groups (10–11, 13–14, and 16–17) were asked to list in order of importance three pop singers they idolized the most. Then they proceeded to answer a variety of questions concerning the singer who represented their first choice. Among others, participants answered questions concerning their potential reliance on the singer’s recommendations and opinions in (1) personal matters (including the domains of pastime, future planning, physical appearance, social relations, personal feelings, and general reliance), and (2) overall knowledge and values (including school studies, politics, values, and science). This research yielded three findings of interest: (1) reliance on the pop idol in regard to personal matters is considerably greater than reliance with regard to overall knowledge and values, (2) it declines with age, being most pronounced in the 10–11 age group, and lower subsequently, and (3) it is more prevalent among girls than among boys, particularly in the early age group (see Fig. 5). The authors conclude that ‘‘. . . during late childhood and early adolescence the phenomenon of pop star idols reaches its maximum intensity. In this period, boys and girls begin separation from their parents and start to join the youth culture, forming their own peer group. This is the beginning of the maturation process during which adolescents establish their self-identity and independence. At this time they look for other identification figures than their parents. The media, especially television radio and youth magazines, provide them with alternative objects of admiration, among them pop singers. Through these channels they are not only exposed to melodies and lyrics, but also to information that serves as a basis for idolization . . .’’ (Raviv et al., 1996, pp. 664–645).

Fig. 5. Mean reliance on pop singer in personal matters as a function of age and gender (Raviv et al., 1996).

364

ARIE W. KRUGLANSKI et al.

C. SUMMARY The studies reviewed above attest to the utility of the epistemic authority construct as a ‘‘window’’ to understanding the course of social development. In general, the results show a developmental trend involving a decline in authority assigned to the primary caregivers, and an increase in the epistemic authority attributed to the self. Evidenced also is a developmental trend toward diVerentiation and specificity. Whereas in early childhood the list of epistemic authorities is limited and those authorities are rather general across domains of knowledge, with age individuals begin to assign diVerent sources to diVerent knowledge domains and add novel sources to their ‘‘list.’’ However interesting and suggestive, studies considered thus far are limited in two respects. First, the upper age limit they tapped was that of late adolescence, and they did not investigate epistemic authority assignments by adult individuals. Second, and perhaps more importantly, they pertained exclusively to perceptions of epistemic authority and did not address the cognitive and behavioral impact that epistemic authority assignments may exert on perceivers. Both concerns were addressed in research described below.

D. INDIVIDUAL DIFFERENCES IN THE DISTRIBUTION OF EPISTEMIC AUTHORITY ASSIGNMENTS ACROSS SOURCES An early study by Bar (1983) devised a Hierarchy of Epistemic Authorities Test designed to investigate the epistemic authority assigned by Israeli college students to various sources. Participants answered 33 hypothetical questions regarding the following seven diVerent life domains identified by Jourard (1959): life style, social activities, interpersonal relations, child education, work, finances, and mental or physical problems. Regarding each question, participants were required to rank the following eight sources of authority (derived from the appropriate pilot research) in terms of their reliance upon them for advice: (1) majority opinion, (2) the person’s own reference group, (3) a domain-specific expert, (4) a general expert, (5) self authority – logical, (6) self authority – sensory, (7) self authority – aVective, and (8) self authority – intuitive. Participants also rated the first ranked authority in regard to each question on a scale from 1 to 10 regarding the degree to which they relied on its recommendations in each domain. Bar’s (1983) test had a good reliability (Cronbach  values for the diVerent sources of authority ranged from .82 to .94). It also exhibited considerable variability

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

365

in how diVerent persons distributed their assignment of epistemic authority across diVerent sources. This was true both when aggregating across domains, as well as within domains. Bar (1983) found intriguing gender diVerences in epistemic authority assignments. Specifically, in domains prototypically classified as masculine (such as work and finances), women viewed their peer group as a more dominant epistemic authority than did men, whereas in domains prototypically classified as feminine (social life, interpersonal relations, children’s education), men endowed their peer group with greater epistemic authority than did women. Apparently, where one’s own epistemic authority is low (as was the case for men in the feminine domains and for women in masculine domains), one’s reference group gains in epistemic dominance compared with domains where one’s self-ascribed epistemic authority is high. But the main concern in Bar’s (1983) research was whether assigned epistemic authority predicted behavior. To that end, she had participants engage in an Information Purchasing Test wherein participants’ role was to ‘‘purchase’’ information from six sources in order to reach a correct decision regarding a presented problem. The assumption was that the greater the epistemic authority assigned to a given source, the more resources participants should be willing to spend in order to purchase information from that source. The results indeed demonstrated that the Hierarchy of Epistemic Authorities Test was highly successful in predicting participants’ purchasing behavior. In all seven domains investigated, participants were willing to pay more (in hypothetical money represented by chips) for information from their highest domain-specific authority than for information from the second highest (or any other) authority. Bar (1999; Study 1) later abbreviated and computerized the ‘‘Hierarchy of Epistemic Authorities Test’’ and tested its implications in several subsequent studies. The Cronbach  values in this later research ranged between .72 and .87. In her first experiment, participants chose among eight products in each of four domains, namely dieting, cars, hair shampoo, and dryers. Participants were first instructed to open windows corresponding to the eight epistemic authorities in order to discover the corresponding authority’s recommendation with regard to one of the products. Participants were allowed to open as many windows as they felt were needed for a confident choice of the highest-quality product, and they subsequently did so for their second-highest-quality product. It was found that participants tended overwhelmingly to open first the window corresponding to their dominant epistemic authority and to choose the product recommended by that authority. Whereas by chance alone no more than 12.5% of the participants should open a given window, and/or choose its recommended product, the

366

ARIE W. KRUGLANSKI et al.

Fig. 6. Percentage of participants who first opened window corresponding to top epistemic authority and who chose product that it recommended (Bar, 1999; Study 1).

corresponding figures obtained in Bar’s study were in the 40%–60% range (see Fig. 6). It was also found that the choice of the ‘‘highest quality’’ product was determined by the dominant epistemic authority to a greater extent than the choice of the second-most-desirable product. Participants also expressed significantly greater confidence in their product choice when it was recommended by a dominant versus a nondominant epistemic authority (i.e., an epistemic authority other than that first-ranked one). Finally, for participants who chose the product recommended by their dominant epistemic authority—those who rated that authority as ‘‘strong’’ (that is, in the upper third of the ratings distribution) expressed greater confidence in their choice than those who rated that authority as ‘‘weak’’ (lower third of the distribution). 1. Epistemic Authority or Heuristic Cue? The above main eVects of epistemic authority were replicated and extended in Bar’s (1999; Study 2) second study that superimposed on the original procedures, orthogonal manipulations of time pressure (high vs. low) and evaluation apprehension (high vs. low). Specifically, half the participants were given 30 seconds to complete each decision problem, whereas the remaining half were given 10 minutes to do so. Similarly, half the participants were led to believe that the computer would record all of their choices in order to assess their decision-making ability and that feedback to that eVect would be provided them at the end of the experiment. The remaining participants were led to believe that the computer program was still in its testing phase and that their responses would not be recorded or evaluated in any way. Just as in the first study, it was found that participants tended

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

367

to first open the window pertaining to their dominant (vs. nondominant) epistemic authority, that they tended to do so to a greater extent if their dominant epistemic authority was rated as ‘‘strong’’ versus ‘‘weak,’’ that they were more confident in their decisions if those were based on the recommendations of dominant (vs. a nondominant) epistemic authority, and that this was true to a greater extent where the dominant epistemic authority was rated as ‘‘strong’’ versus ‘‘weak.’’ Moreover, participants tended to spend more time on information contained in a window belonging to their dominant (vs. nondominant) epistemic authority. Finally, participants whose first window corresponded to their dominant epistemic authority tended to open significantly fewer windows subsequently and to do so to an even greater extent if their dominant epistemic authority was rated as ‘‘strong’’ vs. ‘‘weak.’’ These findings held even when controlling for the amount of time participants spent on their first window, contrary to the alternative hypothesis that these findings were due to the fact that spending a great deal of time on the first window simply did not leave enough time for subsequent windows. This finding suggests that in addition to the ‘‘seizing’’ process aVected by the dominant epistemic authority, information provided by such an authority tends to induce a ‘‘freezing’’ on its recommendations so that subsequent information search is deemed less necessary and hence it is curtailed (Kruglanski & Webster, 1996). The manipulations of time pressure and accuracy referred to earlier were specifically intended to elucidate the nature of epistemic authority as a social psychological construct. One possibility to consider is that this particular construct may represent a peripheral ‘‘cue’’ or a ‘‘heuristic,’’ reminiscent of the way source characteristics were treated in much persuasion research (cf. Petty & Cacioppo, 1986; Chaiken et al., 1989). If so, one might assume that the impact of the dominant epistemic authority would be reduced under high-accuracy instructions, representing a high degree of processing motivation, and increased under time pressure, known to induce a high need for cognitive closure. In other words, if deferring to one’s dominant epistemic authority represents primarily a ‘‘quick and easy,’’ though a tad less reliable, way of forming judgments (and making the corresponding decisions) and if one is less inclined to choose such peripheral (or heuristic) processing when the elaboration likelihood is high (Petty and Cacioppo, 1986), then the influence of the dominant epistemic authority should be more pronounced under time pressure and in the absence of the accuracy instructions (representing a low elaboration likelihood) than in their presence. Similarly, if time pressure lowers the ‘‘elaboration likelihood,’’ individuals should be more likely to rely on their dominant epistemic authority in the presence (vs. absence) of time pressure. In other words, in the absence of time pressure and/or under accuracy instructions, individuals should open the

368

ARIE W. KRUGLANSKI et al.

informational windows relatively leisurely and process the information that they contained more extensively. This, however, did not happen. Instead, the epistemic authority eVects (on number of informational windows opened first and on confidence in recommendations of a given source) held sway across variations in time pressure and in accuracy concerns (see Figs. 7 and 8). In this connection, the fact that participants had higher confidence when their judgments were based on the recommendations of their dominant epistemic authority also questions the notion that source expertise functions merely as a heuristic (as in the HSM) that generally aVords a low judgmental confidence (appropriate only when the ‘‘suYciency threshold’’ is low). Apparently, the degrees of judgmental

Fig. 7. Mean number of informational windows opened first, corresponding to strong vs. weak dominant epistemic authority under two degrees of time pressure and accuracy concerns (Bar, 1999; Study 2).

Fig. 8. Mean degree of confidence in chosen products recommended by dominant vs. non dominant epistemic authority under two degrees of time pressure and of accuracy concerns (Bar, 1999; Study 2).

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

369

confidence conferred by sources may vary in accordance with their perceived authority, just as the degree of confidence aVorded by message arguments may depend on their perceived cogency (cf. Kruglanski & Thompson, 1999a,b; Erb et al., 2003). Thus, some sources (i.e., ones assigned high epistemic authority) might be able to bestow higher degrees of confidence on judgments based on their recommendations than some message arguments (i.e., those of relatively low cogency), whereas other sources (i.e., ones assigned low epistemic authority) may bestow lower degrees of confidence than other message arguments (i.e., those of relatively high cogency). These findings suggest that one’s dominant epistemic authority may represent a trusted source of knowledge, more so in some cases than one’s own selfascribed authority in a domain (that was nondominant for numerous participants in various domains). Hence, the dominant authority tends to be relied upon regardless of situational constraints such as those having to do with the amount of time at one’s disposal and the pressure to be accurate. In her third experiment, Bar (1999; Study 3) adopted a continuous measure of influence exerted by various epistemic authorities. Participants were presented with numerical estimates ascribed to various sources, e.g., ‘‘according to most people, a dryer activated twice a week is likely to last 12 years,’’ or ‘‘according to a mechanical expert, a gas-driven engine produces 300 sparks per second.’’ Participants had to form their own judgment on the issue addressed in the statement. It was found that participants’ judgments were ‘‘anchored’’ to a significantly greater extent on judgments of the source if it represented their dominant versus nondominant authority and when their dominant epistemic authority was rated as strong versus weak. Also, participants’ judgmental latencies were significantly briefer when the recommending source was a dominant (vs. nondominant) epistemic authority and when the recommending dominant source was rated as ‘‘strong’’ vs. ‘‘weak’’ in epistemic authority.

E. SUMMARY Bar’s (1983, 1999) studies reviewed above attest that the hierarchy of epistemic authorities matters. Individuals apparently diVer systematically in their distributional profiles of epistemic authority assigned to various sources, and these diVerences, in turn, are related systematically to individuals’ search for, and use of, information. Specifically, this research shows that individuals first turn to information provided by sources whom they regard as highest (vs. lower) in relative epistemic authority, they process such information more extensively, derive from it greater confidence, and tend more to act in accordance with its perceived implications.

370

ARIE W. KRUGLANSKI et al.

F. DIFFERENTIATION BETWEEN INTERNAL AND EXTERNAL EPISTEMIC AUTHORITIES A particularly important aspect of the epistemic authority construct is its diVerentiation that occurs over the course of socialization and development. And a particularly interesting aspect of such a diVerentiation is the distinction between one’s own self-ascribed authority and the authority ascribed to external agents. One implication of such a diVerentiation is that individuals for whom external authority is dominant would be more sensitive to external cues in general, more so than individuals for whom the dominant epistemic authority is internal. Specifically, individuals for whom external (vs. internal) authority is dominant may tend more to direct their attention externally and be particularly vigilant to changes in the external environment. Two studies provide evidence in support of this possibility. In Bar’s (1999; Study 3) research with numerical anchors, she again orthogonally manipulated time pressure and accuracy instructions. And again, the dominant epistemic authority exerted strong main eVects and did not interact with the time pressure and accuracy variables. However, there appeared an interesting diVerence between participants whose dominant epistemic authority was external vs. those whose authority was internal, or self-ascribed. Specifically, under high time pressure and in the absence of accuracy instructions (i.e., under situational conditions known to enhance the need for cognitive closure), those with an external epistemic authority tended to bring judgments significantly closer to the anchors mentioned regardless of the source they were mentioned by. In other words, participants assigning a dominant authority to external sources tended more to ‘‘freeze’’ on judgments of the source, that is, behave like high-need-forclosure persons, than individuals whose dominant epistemic authority was internal (!). It appears, then, that individuals for whom the dominant epistemic authority is external (vs. internal) are more aVected by external cues, at least those known to induce a heightened need for closure (Kruglanski & Webster, 1996). A conceptually similar eVect appeared in Bar’s (1999) fourth study using an entirely diVerent procedure based on Payne’s (1976) information board. On this particular board, listed on the y-axis are the various choice alternatives (in this particular case, various products) and on the x-axis the various features of those alternatives. It is generally assumed that opening windows along the x-axis represents a conflict-prone search strategy (because the various features represent divergent criteria, e.g., cost vs. durability vs. appearance vs. design) on which the diVerent alternatives can be judged. On the other hand, opening windows along the y-axis represent a simpler, less conflictual strategy—because one would simply choose for each feature

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

371

Fig. 9. Choice of conflict avoidant strategy as a function of externality of dominant epistemic authority and time pressure (Bar, 1999; Study 4).

the one alternative highest on that feature. We know that under high-needfor-closure, individuals are driven by a sense of urgency and immediacy, and they try to avoid cognitive conflicts that necessitate lengthy deliberations (Kruglanski, 2004; Kruglanski & Webster, 1996). Bar (1999; Study 4) found that participants with an external (vs. internal) epistemic authority, under closure-prompting situational conditions (high time pressure, no accuracy instructions) opted to a significantly greater extent for the simpler strategy (opening windows along the y-axis), i.e., they seemed to be more responsive to situational cues suggesting closure. Just like in the prior study, then, individuals with dominant external epistemic authority appeared to be more sensitive to situational or external cues (implying the need for closure) than individuals with a dominant internal authority (see Fig. 9).

G. EFFECTS OF SELF-ASCRIBED EPISTEMIC AUTHORITY 1. Self-Ascribed Epistemic Authority and External Information Search Under Need for Closure A major unique aspect of the epistemic authority construct is that it treats analogously external sources of information and the self. Indeed, several recent studies looked at informational eVects as a function of the selfascribed epistemic authority. In one such study, Pierro and Mannetti (2004) measured the strength of individuals’ self-ascribed epistemic authority in the specialized domain of cell phones. To that end, they constructed a 13-item scale including questions such as ‘‘I truly have considerable knowledge about diVerent types of cell phones,’’ ‘‘I can say a great many things about technical specs of diVerent cell phones,’’ ‘‘I can oVer people useful

372

ARIE W. KRUGLANSKI et al.

advice regarding the purchase of a cell phone,’’ etc. Reliability of this measure was quite high (Cronbach  ¼ .91). Pierro and Mannetti (2004) also assessed their participants’ dispositional need for cognitive closure using the Italian version of Webster and Kruglanski’s (1994) Need for Closure Scale. The main dependent variable of interest was participants’ readiness to search for information from external sources in case they entertained the possibility of purchasing a cell phone. This was measured, again, via several pertinent items, e.g., ‘‘If I were to buy a cell phone I would seek information from a large number of sources,’’ ‘‘If I were to buy a cell phone, I would visit a large number of retailers,’’ ‘‘If I were to buy a cell phone, I would seek advice from friends, relatives, and neighbors,’’ etc. The Cronbach  of this scale was .83. It was found that the greater the individuals’ self-ascribed epistemic authority in a domain, the less external information they indicated they would seek. This finding was not at all surprising. After all, having high self-ascribed epistemic authority in a domain is nearly tantamount to having less need of advice from others. What was more interesting, however, is that the tendency to seek such external information was moderated by the need for cognitive closure. For individuals with a low self-ascribed epistemic authority in the specific domain of cell phones, the higher their need for closure, the stronger their tendency to engage in an external information search. By contrast, for individuals with a high self-ascribed epistemic authority in this domain, the higher their need for closure, the lower their tendency to engage in an external search, and presumably the greater their tendency to rely on their own experience and expertise (see Fig. 10). In other words, under the pressure for closure, individuals seem to adhere particularly strongly to their dominant epistemic authority in a domain, relying on the source they trust the most.

Fig. 10. External information search as a function of Need for Closure (NFC) and Self-Ascribed Epistemic Authority (SAEA) (Pierro & Mannetti, 2004; Study 1).

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

373

These results demonstrate, additionally, that source authority doesn’t constitute a mere peripheral cue or heuristic used as a ‘‘quick and ready’’ manner for reaching judgments when the elaboration likelihood is low (Petty & Cacioppo, 1986) but rather constitutes a factor aVecting the direction of the informational search. That is, significance of the source’s perceived epistemic authority is not that this information is necessarily easy to process (for conditions where it may be diYcult to process, see Kruglanski and Thompson, 1999a, b) but rather that it represents the best source of information to be processed. 2. Informational Impact as a Function of a Self-Ascribed Epistemic Authority The degree of one’s self-ascribed epistemic authority should determine the kind of information likely to exert an impact on an individual’s judgments and decisions. Pierro and Mannetti (2004; Study 2) examined the degree to which more vs. less relevant product information would be recognized as such and would exert an impact on attitudes toward the product by individuals with diVerent degrees of self-ascribed epistemic authority in a domain. Participants, students at the University of Rome, ‘‘La Sapienza,’’ filled out a scale of self-ascribed epistemic authority in the domain of cell phones as described earlier. Then, in the second research phase administered a month later, they were presented with product information about a cell phone of Brand X that a well-known manufacturer was allegedly planning to introduce into the market and that a market research firm was investigating. Participants were then presented with a description of this telephone in comparison with its two competitors, Brand A and Brand B. In all cases, Brand X was portrayed as superior to those competing brands, but the features on which it was portrayed as superior varied. Half the participants read a description in which the comparison features of X with A and B were on dimensions found (in a pretest conducted with a random sample of students) to be highly relevant to the overall quality of a cell phone. These features included: (1) aesthetic quality of the design, (2) number of available functions, and (3) capacity to send text and video messages (a feature ostensibly lacking in A and B), and (4) an integrated dictionary for MMS writing of messages (lacking in A and B). The remaining half of the participants read a description in which X was compared with A and B on features rated by pretest participants as largely irrelevant to the overall quality of the phone. These included: (1) a color display (whereas A and B had only a black and white display, X had a color display), (2) an additional keyboard (present in X and lacking in A and B), (3) a special cover for the keyboard (present in X but lacking in A and B), and (4) an easier-to-use menu.

374

ARIE W. KRUGLANSKI et al.

Fig. 11. Attitude toward telephone as a function of feature relevance and Self-Ascribed Epistemic Authority (SAEA) (Pierro & Mannetti, 2004; Study 2).

It was found that participants with a high self-ascribed epistemic authority in the domain of cell phones were more capable than their counterparts of appropriately discriminating between the less and the more relevant telephone features. More importantly, they were more aVected in their ultimate attitudes toward Brand X than their low-self-authority counterparts by its superiority on the relevant vs. the irrelevant features (see Fig. 11). The above findings suggest that one’s self-perception of epistemic authority may be veridical and be related to actual expertise. A self-ascription of a low epistemic authority may mean that one is truly unable to discern what is relevant in a given domain. Whereas this is not particularly surprising, a more interesting implication is that persons with a low (vs. high) self-ascribed epistemic authority may be more swayed in their attitudes and opinions by irrelevant features and exogenous arguments. Evidence consistent with this possibility was provided by Ellis (1996). 3. Persuasion by Exogenous Features Imagine a personnel manager who, in order to persuade a promising potential candidate to take the job, oVers her or him a trip abroad or a fancy car. This manager may be said to ‘‘decorate’’ the job with exogenous features not really relevant to the job’s contents. Such tactics are often used in other life domains as well. For instance, a caring mother might try to persuade her child to eat its meal by using colorful plates adorned with funny cartoons and/or attractive, animal-shaped eating utensils. A student might try to compensate for a lack of original ideas in a composition by an impeccably neat handwriting and lovely drawings, etc. Ellis (1996) hypothesized that such exogeneous characteristics of the attitude object are likely to exert greater persuasive impact on individuals whose self-ascribed epistemic authority in a domain is low vs. high. In a

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

375

study designed to investigate this issue, senior master’s students at the organizational behavior program at the Faculty of Management at Tel-Aviv University, and about to enter the job market, were asked to express their opinion regarding one of two job oVers: that of a human resource manager and that of an organizational consultant. Previously, these participants responded to a questionnaire designed to assess their self-ascribed epistemic authority in one of these two domains. For instance, one of the 26 items in the organizational domain questionnaire stated, ‘‘The decision has been made to establish a professional team in your organization to help workers in distress. Do you believe that you can be one of the team members?’’ Another item stated, ‘‘To what extent can one rely on your knowledge of management-worker bargaining?’’ Answers were recorded on a 5-point scale with the response alternatives labeled appropriately for each item, e.g., ‘‘absolutely yes’’ ¼ 1, and ‘‘absolutely no’’ ¼ 5 for the first item and ‘‘to a very large extent’’ ¼ 1, and ‘‘to a very slight extent’’ ¼ 5 for the second item, respectively. The Cronbach  for this scale was .92. Thus, one of the 22 items in the personnel management questionnaire read, ‘‘The personnel manager in a big organization has decided to establish a special team to develop and monitor career paths. To what extent do you think that you can be a member of this professional team?’’ Another item stated, ‘‘An industrial manager is having some diYculties in selecting good criteria for middle-level manager promotion. To what extent do you believe you could help him solve his problems?’’ Answers were again recorded on appropriate 5-point scales, and the Cronbach  for this scale was .90. The self-ascribed epistemic authority questionnaire (SAEA) was administered at each participant’s home. Subsequent to a completion of this questionnaire by the participant, the researcher said that he was also working for a ‘‘Student Vocational Guidance Service’’ and that he was conducting a survey on ‘‘how job oVers were perceived by candidates.’’ He further stated that because he knew that he would be meeting organizational behavior students about to graduate, he had brought with him a real and actually available (or ‘‘hot’’) job oVer. He then asked the participant whether he or she would like to look it over with an eye of possibly taking the job. Not surprisingly, all participants agreed to do so. A description of the job requirements was attached to each of the positions presented. In order to reach a decision regarding the position, one of the versions included information endogenous to the position, whereas the other included both endogenous and exogenous features. The job content characteristics for each job were selected to describe the central aspects of the relevant position. For the organizational consultant’s position, these included individual consulting for managers, team building, and leading sensitivity training workshops. For the personnel manager’s position these included personnel recruitment,

376

ARIE W. KRUGLANSKI et al.

Fig. 12. Job attitudes as function of Self-Ascribed Epistemic Authority (SAEA) and job description (content only versus content + context) (Ellis, 1996).

selection and assessment, and monitoring career paths. The job context features included a company car, an executive club membership, spacious oYce, etc. The results demonstrated that participants with a low self-ascribed epistemic authority invested more eVort in processing exogenous information in their attempt to reach a decision. Additionally, they retained the irrelevant information better and perceived the position as more professional than did participants with a higher degree of self-epistemic authority. Those participants with a higher level of self-epistemic authority needed more time to process information regarding only those characteristics that were relevant to the position and retained this information better than the irrelevant information. Most importantly, in the condition where only job-endogenous features were presented, individuals’ SAEA scores were positively and significantly correlated with positive attitudes to the job oVer. This correlation was significantly lowered (and rendered nonsignificant) in conditions where both job content and job context features were presented. These findings suggest that individuals with a high SAEA base their attitudes on job-endogenous features, whereas those with a low SAEA base them on jobexogenous features. The pattern of means shown in Fig. 12 in fact demonstrates that (across the two job oVers) for high-SAEA participants (above the median of the SAEA score distribution), the addition of attractive job context features tended to lower these individuals’ positive attitudes toward the job oVers, whereas for low-SAEA participants, the addition of those features tended to enhance their attitudes to the job oVers. H. SUMMARY The findings described above attest to major diVerences in information processing of individuals with high levels of self-ascribed epistemic authority vs. individuals with low levels of self-ascribed epistemic authority.

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

377

These diVerences are reflected in processing duration, memory for endogenous and exogenous information, as well as in the ability to discern, and be appropriately aVected by, more vs. less relevant information.

I. SELF-ASCRIBED EPISTEMIC AUTHORITY AND LEARNING FROM EXPERIENCE Among the more interesting implications of the concept of epistemic authority is the phenomenon of learning from experience. It will be noted that the concept of ‘‘experience’’ has long been privileged in psychological theory. The use of experiential learning in training and education has been inspired by John Dewey’s (1916, 1958) instructional philosophy, Carl Rogers’ (1951, 1967) person-centered approach to therapy, and by humanistic psychology more generally (e.g., Shafer, 1978). In social psychology, Fazio and Zanna (1981) have suggested that attitudes acquired via direct experience with the attitude object are the strongest and are most tightly related to behavior. Yet, these last authors also hinted at the possibility of moderators that may qualify the power of experience to mold attitudes. As they put it, ‘‘An attitude formed by indirect means could conceivably also be held with extreme confidence, and, hence, be more predictive of behavior than a direct-experience attitude. For example, a child’s attitude toward members of a given ethnic or racial group may be held with great confidence, even though formed indirectly because of his or her parents’ extreme credibility (Fazio & Zanna, 1981, p. 184). The foregoing quote indeed suggests that experience may not constitute a superior base of knowledge under all conditions, but it stops short of identifying the critical determinant of whether it may or may not be capable of fulfilling this function. The concept of self-ascribed epistemic authority may well constitute such a determinant. In other words, whether or not personal experience would be regarded as a reliable knowledge base may depend on one’s self-ascribed capability to draw reliable conclusions from the experience or on one’s self-ascribed epistemic authority in a domain. In absence of such an authority, a person may fail to utilize the experience to derive confident knowledge. An individual may speak English all her life without deriving the principles of English grammar from this experience, she may drink a wide range of wines over the years without forming notions about the diVerent varietals or vintages, or she may play tennis on a weekly basis without forming notions about the proper strategies and tactics of this game. More generally, we assume that the extent to which individuals tend to draw confident conclusions from information is related to their authority assignment to the information source. When the information consists of

378

ARIE W. KRUGLANSKI et al.

one’s own experience, the source simply is oneself. In these circumstances, the higher one’s self-ascribed epistemic authority, the more readily one may trust one’s own interpretation of information and the more might one be able to ‘‘benefit’’ from the experience. When the information is interpreted by an external communicator (e.g., a teacher or a parent), however, the individual’s tendency to accept the interpretation may partially depend on the perceived gap in epistemic authority between the source and the self. When the authority imputed to the source is considerably higher than that imputed to the self, the source’s pronouncements are likely to be attended closely and/or be assigned considerable weight. However, when the assigned authorities are more nearly equal, the source’s statements might not be taken as seriously because of a sense that there is little the source could contribute to one’s own ability to process the information. Thus, a ‘‘reverence eVect’’ is hypothesized whereby interpretations from an external source will have greater impact on persons whose perceived authority gap between themselves and the source is large vs. small. In a study designed to investigate these notions, Ellis and Kruglanski (1992) assessed their participants’ self-ascribed epistemic authority in mathematics. This was accomplished via a 14-item questionnaire inquiring into participants’ perception of their expertise and confidence in understanding and being able to conceptualize mathematical material. For instance, one item on this questionnaire read, ‘‘After completing a math exam, I can tell how well I have done.’’ Answers were recorded on a five-point scale with the response alternatives ranging from ‘‘always’’ to ‘‘never.’’ Another relevant item was phrased, ‘‘Your fourteen-year-old son is having diYculties with mathematical material with which you have not dealt in a long while. To what extent would you prefer to assist him yourself as opposed to getting him outside help?’’ Response alternatives in this case ranged from ‘‘very large extent’’ to ‘‘very slight extent.’’ Cronbach  of this SAEA scale was .89, attesting to a relatively high degree of internal consistency. Participants also responded to the numerical aptitude test (Cattell & Epstein, 1975) to serve as a control measure for their actual math ability and to a postexperimental questionnaire designed to assess their perceptions of their own and the instructor’s (whenever appropriate) epistemic authority in mathematics. The mathematical learning task employed in this research consisted of multiplication exercises in which some numbers were replaced by letters, as in the following example: The numbers 9, 6, 7, 5, and 3 in the following exercises have been replaced by letters. Find the missing numbers:

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

379

22a bce 202a 1ba laba 1hh2ba To successfully solve this kind of exercise, participants need to use five arithmetic principles: (a) multiplication of x by 0 ¼ 0; (b) multiplication of x by 1 ¼ x, (c) multiplication of odd numbers by 5 ¼ a number whose last digit is 5; (d) multiplication of an even number by 6 ¼ an even number whose last digit is the same as the multiplicand.4 Participants were introductory psychology or education students at Tel-Aviv University. They were randomly assigned to one of three experimental conditions: In the experiential condition, participants were given self-instruction booklets with six exercises related to the five arithmetic rules above. In the instructional principles condition, the experimenter was introduced by the participants’ professor as a Ph.D. in mathematics conducting research on various ways of teaching the subject. For this condition, he stated the various mathematical principles explicitly. In the intermediate instructional-examples condition, the instructor was introduced in the same way as before, but now he both solved on the board the very same exercises contained in the ‘‘experiential booklet’’ and stated the arithmetic principle underlying each solution. Following these procedures, participants took a performance test on the principles they had just been taught. Participants in the two instructional conditions were additionally asked to estimate the gap in ability between themselves and the instructor. Specifically, they answered the question: ‘‘In your opinion, how large is the gap between your and the instructor’s ability to solve exercises of the present type?’’ Responses were recorded on a 5-point scale anchored at the ends with ‘‘a very large gap’’ and ‘‘a very small gap.’’ The results of this research indicated that the method of instruction entered into a significant interaction with participants’ SAEA (see Fig. 13). Controlling for participants’ actual mathematical ability, in the experimental condition, participants with a high SAEA did significantly better than participants with a low SAEA. In the instructional principles condition, the low-SAEA participants tended to do better than their high-SAEA 4 Thus, in the present example, it is easy to discover that a ¼ 5 and e ¼ 9. Hence, e  a ¼ 45 resulting in the digit 5 (or a) in the first row of the multiplication. If a ¼ 5 and a þ 2 ¼ b (the second to last digit of the first row plus the last digit of the second row), then b must be 7, and so on.

380

ARIE W. KRUGLANSKI et al.

counterparts, and in the intermediate instructional-examples condition, the high-and low-SAEA participants did about the same. In the two instructional conditions, it was furthermore found that participants with a high SAEA perceived the gap between their own and the instructor’s ability as significantly lower than did participants with a low SAEA. Of greater interest, in both instructional conditions, participants who perceived a large gap between themselves and the instructor did better in both instructional conditions than participants who perceived a smaller gap (see Fig. 14). A large gap indicates that the source’s relative epistemic authority (compared with one’s own) is considerable. This may turn the recipient into a ‘‘true believer,’’ enhancing her or his readiness to accept the source’s conclusions and recommendations.

Fig. 13. Mean math performance as a function of self-ascribed epistemic authority and learning condition (Ellis and Kruglanski, 1992).

Fig. 14. Mean math performance in instructional conditions as function of perceived gap in epistemic authority between self and instructor (Ellis and Kruglanski, 1992).

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

381

J. SUMMARY The findings above identify an important boundary condition on the eYcacy of experience as a mediator of learning. It appears that in order to be able to learn from experience, individuals need to believe in their ability to draw inferences in the experiential domain in question, captured by our construct of self-ascribed epistemic authority. It is of particular interest that self-ascribed epistemic authority is empirically distinct from actual ability in a domain. In the present study, the correlation between the two, though significant, was relatively low (r ¼ .36), and the interaction between SAEA and method of instruction was significant, controlling for actual mathematical ability. Finally, yet of considerable interest, in the instructional learning conditions it was participants whose perceived gap between own and instructor’s ability was large (vs. small) who did significantly better attesting to a ‘‘reverence eVect’’ whereby the impact of an external source is greater if its authority is high relative to one’s own self-ascribed authority.

K. EPISTEMIC AUTHORITY BELIEFS IN REAL-WORLD CONTEXTS Beliefs about epistemic authorities of various sources constitute lay theories that people may hold and that exert an important influence on how they think, feel, and act in given situations (Dweck, 1999). Identification of people’s lay theories about epistemic authority may thus provide a useful window into their reactions to information and communication. This may prove of considerable value in real-world contexts (e.g., in education, politics, and commerce), where understanding the underpinnings of people’s attitude change is of critical importance. 1. Political Socialization in College In a study on the change in students’ political attitudes in the course of their college experience, Guimond and Palmer (1996) examined the epistemic authority that students ascribed to their peers, professors, and courses to investigate whether the attitude change shown to occur during the college years, first demonstrated in Newcomb’s (1943) classic study in Bennington college, is ascribable to the college experience as a whole, compared to the academic major. To answer that question, Guimond and Palmer (1996) conducted a three-year longitudinal study at a Canadian university with commerce and social science majors as participants. In the first phase, participants answered questions about the causes of unemployment and

382

ARIE W. KRUGLANSKI et al.

poverty, including items about the degree to which they blamed the poor and the unemployed for their condition, and other items about the degree to which they assigned blame for poverty and unemployment to the political and economic system. In a second phase conducted three years later, these questions were repeated. In addition, participants reported how much their ideas were influenced by the course contents, professors, fellow students, and other sources of information. Guimond and Palmer (1996) found intriguing diVerences between the commerce and the social science majors in the relative epistemic authority they ascribed to various sources. In commerce, peers were generally perceived as the most important source of influence. The perceived influence of professors and course contents were lower in comparison. In contrast, the social science students perceived course contents and peers to be at a similar level of influence. It was also found that in both the commerce and social science, peers had generally conservative eVects; that is, their perceived epistemic authority was correlated with a conservative shift in the students’ attitudes. Professors and courses in the social sciences had a liberal eVect, in that their perceived epistemic authority was correlated with an attitudinal shift in the liberal direction. By contrast, in commerce, professors and courses had a conservative eVect on students’ attitudes. Possibly as a consequence of these patterns of ascribed epistemic authority, commerce students became less likely over time to attribute poverty and unemployment to the political system, thus increasing their tendency to engage in ‘‘system justification’’ (Jost & Banaji, 1994) and to exhibit political conservatism. By contrast, social science students for whom the conservative influence of peers tended to be counteracted by the liberalizing influence of professors and courses tended less over time to engage in system justification, or attribute unemployment internally to the poor. Guimond and Palmer (1996, p. 2009) conclude that ‘‘This provides considerable support to theory and research related to the concept of epistemic authority . . . [and that] results concerning the role of perceived source of influence are noteworthy and represent a potentially fruitful area for future research.’’ 2. Extra-Academic Sources of Political Attitudes A critical issue in democratic political systems is persuading voters to support candidates’ positions on various issues. But political communications often ‘‘fall on deaf ears,’’ and recipients may reject or reframe arguments to the extent that these are delivered by sources of low ascribed epistemic authority. In an attempt to identify who such authorities may be for members of the Israeli electorate, Bar-Tal et al. (1991) investigated the relation between similarity of positions on political issues and the ascriptions of epistemic authority in the political domain. Israeli university students

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

383

were asked to identify their political orientation and to respond to a questionnaire addressing nine diVerent and controversial political issues related to the Arab-Israeli conflict (e.g., establishment of an autonomous rule in the West Bank and the Gaza Strip, building Jewish settlements in the occupied territories, death penalty to terrorists, etc). With respect to each issue, participants were presented with a list of three prominent sources: a politician, a professor, and a journalist. In total, participants evaluated their reliance on 27 sources (9 from each professional sector). Three selected sources were previously identified as having a leftist political orientation, three as having a rightist political orientation, and three as having a centrist orientation. Participants were asked to evaluate on a 5-point scale the degree to which they would rely on the information from each source regarding each of the nine issues. The investigators’ point of departure was that individuals will bestow higher epistemic authority on sources sharing their political orientation. Several reasons combine to suggest this prediction. First, the selective exposure to information hypothesis (Sears & Freedman, 1967; Frey, 1986) suggests that people will seek out sources of information likely to provide information with which the recipients can agree. Such information may forestall situations of aversive dissonance (Festinger, 1957), or epistemic uncertainty (Kruglanski, 1989) and allow individuals to perceive their opinions as a stable and reliable basis for action. Second, similarity has been known to promote liking (Byrne, 1971), and liking for a communicator has been known to increase persuasion (Petty & Cacioppo, 1986; Chaiken, et al., 1989). In turn, the self-perception of persuasion might lead one to infer that the source had considerable epistemic authority, and that is why it was persuasive. Third, the causal relation between similarity and perceived authority might be such that the latter determined the former, that is, that perceived authority of the source induced persuasion, which made the recipients’ opinions and attitudes similar to those of the source. Consistent with the above analysis, it was found that students bestowed greater epistemic authority on politicians whose opinions they agreed with, suggesting that political communicators might be ‘‘preaching to the converted,’’ i.e., to recipients that a priori agree with their conclusions. Of greater local interest, there obtained some intriguing diVerences between recipients with a leftist (or dovish) vs. a rightist (or hawkish) political orientation, and in accordance with the sources’ profession. Specifically, the leftist students tended to rely more exclusively on leftist sources than did the hawkish students who were more open to sources of the opposing political persuasion (namely leftist university professors and politicians). Bar-Tal et al. (1991a, p. 11) interpreted this to mean ‘‘that dovish students are more closed in their views than hawkish students.’’ However, subsequent

384

ARIE W. KRUGLANSKI et al.

research by Raviv, Bar-Tal, Raviv, and Abin (1993; Study 1) found that students with a rightist political orientation reported, in general, a higher reliance on their leaders and especially perceived them more as generalized epistemic authorities across knowledge domains than did leftist students. This finding suggests that the rightist students tend to perceive their leaders less critically and more unquestioningly than do the leftist students. It is thus possible that rightist recipients are less critical of authoritative sources in general, including sources preaching the opposite political views (Bar-Tal et al., 1991a) and that the leftist students are generally more critical of authoritative sources including sources preaching views with which they agree. These finding accord with a recent meta-analysis (Jost, Glaser, Kruglanski, & Sulloway, 2003) whereby conservatives, or rightleaning individuals, tend to have a higher need for closure (Kruglanski & Webster, 1996) and in this sense are more authority oriented than are liberals, or left-leaning individuals. It was also found that all participants, regardless of their political orientation, perceived centrist journalists as the most reliable epistemic authorities. Generally, university professors were perceived as important epistemic authorities, possibly due to their assumed extent of knowledge and commitment to objectivity. Finally, leftist journalists were perceived as least reliable and were endowed with the least epistemic authority in comparison with the other sources. In this connection, Bar-Tal et al. (1991a, p. 12) comment on ‘‘the prevailing perception in Israel that leftist journalists are not to be trusted and [that they] stand outside the main consensus [and] are generally considered as opinionated and extreme in their positions.’’ In general, the findings by Bar-Tal et al. (1991a) suggest the utility of locally mapping the patterns of epistemic authority and identifying whose opinion matters to what recipients. Such information could serve as an important basis for launching eVective political discussions in which the issues at hand are considered relatively open-mindedly and in which they receive the attention they deserve. 3. Parents’ Epistemic Authority in the Realm of Political Knowledge Political psychologists have often commented on the intergenerational continuity in voting patterns in the United States (Sears, 1969). In this vein, Shapira (1993) conducted a study exploring the role of the parents’ epistemic authority in fostering such a continuity. Specifically, Shapira carried out telephone interviews with a sample of 16- to 18-year-old Israelis as well as either their father or their mother. The interview included questions pertaining to the parent’s epistemic authority in politics as well as questions tapping the adolescents’ and their parents’ political opinion.

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

385

Shapira found a significant relation between the father’s perceived epistemic authority in politics and the similarity between his and his son’s or daughter’s political views. In other words, the more the father was perceived as an epistemic authority, the smaller was the gap between his and his children’s political views. By contrast, there was no relation between the mother’s perceived epistemic authority and the similarity between her and her children’s views. In part, this may have been due to the fact that the mother’s perceived epistemic authority in politics was substantially lower than the father’s, hence restricted to a portion of the epistemic authority range where the source exercises little influence on the recipients. 4. Epistemic Authority of Professors in Statistics and Psychology DiVerent academic disciplines may be accorded diVerent amounts of prestige in society, and it is therefore of interest to ask whether experts in those domains are bestowed with correspondingly diVerent degrees of epistemic authority and whether such authority is reserved for the specific domains of the experts’ competence or whether it is generalized to other domains of knowledge as well. As a first step toward investigating these issues, Raviv et al. (1993; Study 2) carried out an investigation of the epistemic authority accorded to professors of statistics and psychology by their respective students. It might be assumed that as a branch of mathematics, statistics is perceived as a more exact and specialized science than psychology. Statistical knowledge may be regarded as more remote from common sense and hence accorded greater reverence than psychological knowledge assumed by many to be close to common sense. Thus, it might be hypothesized that statistical experts would be accorded a higher degree of epistemic authority than psychological experts. To investigate these matters, Raviv et al. (1993; Study 2) had statistics and psychology majors evaluate the epistemic authority in domain-specific or general knowledge of professors in their respective departments. It was found that the statistical majors ascribed greater domain-specific authority to their professors than did the psychology students but that the latter ascribed greater epistemic authority to their professors in the domain of general knowledge (see Fig. 15). These findings support the prediction that statistical experts are perceived as greater authorities in their domain of competence than psychological experts but that their authority is perceived as more circumscribed than that of psychological experts. One limitation of the Raviv et al. (1993) research is that the evaluation of the diVerent targets was carried out by participants from diVerent populations (namely, of statistics vs. psychology students). Thus, one cannot be sure whether the eVects are target driven or participant driven. Further research is needed wherein the same

386

ARIE W. KRUGLANSKI et al.

Fig. 15. Mean ascribed epistemic authority to professors of statistics and psychology regarding specific and general knowledge domain (Raviv et al., 1993).

type of participants evaluate the epistemic authority of various sources to investigate the possibility that in highly prestigious academic domains, the experts’ perceived epistemic authority may spill over to domains beyond their area or competence, such as politics, education, or ethics.

L. SUMMARY The results reviewed in the preceding sections attest to the utility of the epistemic authority construct as a tool for mapping the patterns of potential informational influence (French & Raven, 1959) in real-world contexts. Such mappings may be useful in predicting given sources’ influence in a given situation, with respect to a given topic and for a given category of recipients. Such mapping could also serve to appropriately associate communications with sources for maximal persuasive impact in real-world interventions.

VII. Summary and Conclusions The present chapter featured the concept of ascribed epistemic authority (Kruglanski, 1989) oVered as a unique perspective on source eVects in social judgment. In contrast to prior approaches that viewed the source of communication as external to the self, the present framework assumes that both the self and external sources may be assigned diVerent degrees of epistemic authority in diVerent domains and that such assignment aVects various aspects of individuals’ information-processing activities, including the search for information and the readiness to base decisions on the information given.

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

387

The present framework claims a central role for the information’s source in suggesting that an evaluation of its epistemic authority (however implicit) constitutes an essential preliminary phase in individuals’ approach to information. For instance, a careful processing of a persuasive message may occur only after a prior determination has yielded that one possesses a suYcient epistemic authority to make sense of the message and hence is not dependent on an external source for interpretation. In general, it could be the perceived gap in epistemic authority ascribed to oneself versus an external source that determines the source’s influence. A failure to take one’s self-ascribed epistemic authority into account might thus lead to inaccurate estimations of the potential impact of external sources’ authority. A source with relatively high ascribed authority may be quite persuasive with respect to a recipient of much lower self-ascribed authority, but much less persuasive with respect to a recipient of a high self-ascribed authority. We thus conceive of epistemic authority ascriptions as meta-cognitive beliefs about a source of information. This perspective highlights the developmental, individual diVerences, self-related, and applied aspects of source phenomena. The developmental aspect pertains to the fact that in the course of ‘‘growing up,’’ the initial generalized epistemic authority accorded by the child to its adult caregivers is gradually distributed over a variety of sources in the individual’s environment, including the self. The individual diVerence aspect pertains to the fact that individuals’ disparate socialization histories may foster the development of correspondingly diVerent hierarchies of epistemic authorities for diVerent individuals. In turn, these may eVect behavioral diVerences in the search for and reliance on information for decision making and action. The self-relevant aspect relates to the role of experience in attitude and opinion formation and the moderating role that self-ascribed epistemic authority may play in one’s ability to draw strong conclusions from experience. Finally, the applied aspect pertains to the fact that identification of a distributional pattern of epistemic authorities assigned by a specific group of individuals to specific sources in specific real-world contexts (in politics, education, and health domain, among others) may allow one to make specific predictions and plan specific interventions designed to influence those individuals in a desirable direction. For instance, Internet-savvy, educated consumers may weigh heavily their own epistemic authority in health-related domains and expend considerable eVorts on analyzing information relevant to their condition before opting for a course of treatment. By contrast, persons with low self-ascribed authority may be guided primarily by their family physician’s recommendations in all medical matters. The unfolding historical trends in these regards, and their impact on physicians’ and patients relations, could be of considerable real-world interest.

388

ARIE W. KRUGLANSKI et al.

Research conducted in the epistemic authority paradigm supported various aspects of the foregoing analysis with respect to the developmental, individual diVerence, self-related, and applied aspects of source eVects. Yet a great deal of further work is needed to fully understand the development and functioning of epistemic authority in social judgment. As already noted, the relation of epistemic authority ascriptions to attachment patterns (Mikulincer & Shaver, in press) suggests intriguing questions: Would secure attachment to one’s parents contribute to a generalized trust in others and hence to the development of a strong emphasis on external epistemic authorities, and would it mean a retardation in the development of one’s self as an epistemic authority? Or would this contribute, instead, to a more nuanced diVerentiation of epistemic authorities, with the self being ascribed authority in specific domains and others being conceded authority in other domains? Similarly, would an avoidant attachment style contribute to the development of an exaggerated and overgeneralized sense of epistemic authority, and the tendency of excessive self-reliance in domains where one’s objective epistemic competence may be limited? These questions and others may be fruitfully pursued in future research contributing to the integration of attachment and social influence theories. A diVerent set of issues may arise in research on the situational determinants of epistemic authority ascriptions. Social cognition researchers have been emphasizing how people’s attitudes, judgments, and beliefs may be situationally constructed (cf. Bem, 1972; Kruglanski & Stroebe, in press). Insofar as epistemic authority ascriptions also constitute beliefs, they too should be susceptible to various situational influences. In this vein, Brinol and Petty (2004) review evidence attesting that various experimental manipulations may aVect individuals’ confidence in their cognitive responses to a communication topic, and hence, presumably, their self-ascribed epistemic authority in a domain. It would be of interest to investigate whether similar eVects may be obtained in regard to external sources as well. For example, nodding one’s head, an assured tone of voice, the use of definite and unqualified terminology, or the speed of speech (Miller, Maruyama, Beaber, & Valone, 1976) may convey a source’s confidence, leading to an ascription of a high epistemic authority and enhancing the source’s persuasive impact upon a recipient. Furthermore, whereas the situational manipulations of confidence reviewed thus far pertained to possible informational eVects on epistemic authority, it is likely that various motivations would exert such eVects as well. In this vein, Bar (1999) reported that participants’ need for cognitive closure eVected more extreme assignments of epistemic authority, such that the high-authority sources were accorded an even higher authority by individuals high (vs. low) on the need for cognitive closure. Conversely, low-authority sources were accorded an even lower authority by high

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

389

(vs. low)-need-for-closure individuals. Whereas Bar’s findings were based on a dispositional need for closure assessed by means of a scale (Webster & Kruglanski, 1994), it should be the case that situational inducements of such a need, e.g., via noise, fatigue, or time pressure, should also aVect an increased diVerentiation in epistemic authority assignments. In addition to nondirectional motivational eVects (such as the need for closure eVects described above), it is also likely that various directional motivations (Kunda, 1990; Kruglanski, 1996) would similarly exert an impact on epistemic authority assignments to self and others. For instance, in a situation where ascribing epistemic authority to a source would seem highly desirable, e.g., where the source was in a leadership position or where one was informationally dependent on the source for important outcomes, one might assign the source a greater epistemic authority than one would in other circumstances. These possibilities and many others could be profitably explored in subsequent research conducted within the epistemic authority framework.

References Allyn, J., & Festinger, L. (1961). The eVectiveness of unanticipated persuasive communications. Journal of Abnormal and Social Psychology, 62, 35–40. Bar, R. (1983). Hierarchy of Epistemic Authority Test. Unpublished master’s Thesis. Tel-Aviv University. Bar, R. (1999). The Impact of Epistemic Needs and Authorities on Judgment and Decision Making. Unpublished doctoral dissertation. Tel Aviv University. Bar-Tal, D., Raviv, A., & Raviv, A. (1991a). The concept of epistemic authority in the process of political knowledge acquisition. Representative Research in Social Psychology, 19, 1–14. Bar-Tal, D., Raviv, A., Raviv, A., & Brosh, M. (1991b). Perception of epistemic authority and attribution for its choice as a function of knowledge area and age. European Journal of Social Psychology, 21, 477–492. Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 6, pp. 1–62). San Diego, CA: Academic Press. Bowlby, J. (1988). A Secure Base: Clinical Applications of Attachment Theory. London: Routledge. Brinol, P., & Petty, R. E. (2004). Self-validation process: The role of thought confidence in persuasion. In G. Haddock & G. Maio (Eds.), Contemporary Perspectives on the Psychology of Attitudes (pp. 205–226). New York: Psychology Press. Berndt, T. J. (1979). Developmental changes in conformity to peers and parents. Developmental Psychology, 15, 608–616. Byrne, D. E. (1971). The Attraction Paradigm. San Diego, CA: Academic Press. Chaiken, S. (1979). Communicator physical attractiveness and persuasion. Journal of Personality and Social Psychology, 37, 1387–1397. Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic processing within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended Thought (pp. 212–252). New York: Guilford.

390

ARIE W. KRUGLANSKI et al.

Chun, W. Y., Spiegel, S., & Kruglanski, A. W. (2002). Assimilative behavior identification can also be resource dependent: A unimodel-based analysis of dispositional attribution phases. Journal of Personality and Social Psychology, 83, 542–555. Cohen, A. R. (1964). Attitude Change and Social Influence. New York: Basic Books. Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58, 644–663. De Grada, E., Kruglanski, A. W., Mannetti, L., & Pierro, A. (1999). Motivated cognition and group interaction: Need for closure aVects the contents and processes of collective negotiations. Journal of Experimental Social Psychology, 35, 346–365. Dewey, J. (1916). Democracy and Education. New York: Free Press. Dewey, J. (1958). Experience and Nature. New York: Collier Macmillan. Dweck, C. S. (1999). Self-Theories: Their Role in Motivation, Personality, and Development. New York: Psychology Press. Ellis, S. (1996). Selecting information on job content or job context: The moderating eVect of one’s own epistemic authority. Journal of Applied Social Psychology, 26(18), 1643–1657. Ellis, S., & Kruglanski, A. W. (1992). Self as an epistemic authority: EVects on experiential and instructional learning. Social Cognition, 10(4), 357–375. Erikson, E. H. (1950). Childhood and Society. New York: Norton. Erb, H. P., Kruglanski, A. W., Chun, W. Y., Pierro, A., Mannetti, L., & Spiegel, S. (2003). Searching for commonalities in human judgment: The parametric unimodel and its dual mode alternatives. European Review of Social Psychology, 14, 1–47. Fazio, R. H., & Zanna, M. P. (1981). Direct experience and attitude-behavior consistency. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 14, pp. 161–203). New York: Academic Press. Festinger, L. (1957). A Theory of Cognitive Dissonance. Stanford, CA: Stanford University Press. Fiske, S. T. (2004). Social Beings: A Core Motives Approach to Social Psychology. New York: Wiley. Floyd, H. H., & South, D. R. (1972). Dilemma of youth: The choice of parents or peers as a frame of reference for behavior. Journal of Marriage and the Family, 34, 627–634. French, J. R. P., & Raven, B. H. (1959). The bases of social power. In D. Cartwright (Ed.), Studies in Social Power (pp. 150–167). Institute for Social Research, Ann Arbor (MI). Freud, A. (1965). Normality and Pathology in Childhood: Assessments of Development. New York: International Universities Press. Frey, D. (1986). Recent research on selective exposure to information. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 19, pp. 41–80). San Diego, CA: Academic Press. Guimond, S., & Palmer, D. L. (1996). The political socialization of commerce and social science students: Epistemic authority and attitude change. Journal of Applied Social Psychology, 26(22), 1985–2013. Hartup, W. W. (1983). Peer relations. In P. H. Mussem (Ed.), Handbook of Child Psychology (4th ed., Vol. 4, pp. 103–196). New York: John Wiley & Sons. Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and Persuasion: Psychological Studies of Opinion Change. New Haven, CT: Yale University Press. Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication eVectiveness. Public Opinion Quarterly, 15, 635–650. Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system-justification and the production of false consciousness. British Journal of Social Psychology, 33, 1–27. Jost, J. T., Glaser, J., Kruglanski, A. W., & Sullaway, F. J. (2003). Political conservatism as motivated social cognition. Psychological Bulletin, 129(3), 339–375.

EPISTEMIC AUTHORITY EFFECTS IN SOCIAL JUDGMENT

391

Jourard, S. M. (1959). Disclosure and other cathexes. Journal of Abnormal and Social Psychology, 59, 428–431. Kagan, J. (1958). The concept of identification. Psychological Review, 65, 296–305. Kelley, H. H., & Thibaut, J. W. (1969). Group problem solving. In G. Lindzey & E. Aronson (Eds.), The Handbook of Social Psychology (Vol. 4, pp. 1–101). Reading, Mass: Addison Wesley. Kelman, H. C., & Hovland, C. I. (1953). ‘‘Reinstatement’’ of the communicator in delayed measurement of opinion change. Journal of Abnormal and Social Psychology, 48, 327–335. Kruglanski, A. W. (1989). Lay Epistemics and Human Knowledge: Cognitive and Motivational Bases. New York: Plenum. Kruglanski, A. W. (2004). The Psychology of Closed-Mindedness. New York: Psychology Press. Kruglanski, A. W., Fishback, A., Erb, H. P., Pierro, A., & Mannetti, L. (2004). The unimodel as a theory of persuasion. In G. Haddock & G. R. Maio (Eds.), Contemporary Perspectives on the Psychology of Attitudes (pp. 399–422). New York: Psychology Press. Kruglanski, A. W., & Thompson, E. P. (1999a). Persuasion by a single route: A view from the unimodel. Psychological Inquiry, 10, 83–109. Kruglanski, A. W., & Thompson, E. P. (1999b). The illusory second mode, or the cue is the message. Psychological Inquiry, 10, 182–193. Kruglanski, A. W., & Stroebe, W. (in press). Attitudes, goals and beliefs: Issues of structure, function and dynamics. In. D. Albarracin, B. Johnson, & M. P. Zanna (Eds.), Handbook of Attitude Research. New York: Guilford Kruglanski, A. W., & Webster, D. M. (1996). Motivated closing of the mind: ‘‘Seizing’’ and ‘‘freezing.’’ Psychological Review, 103, 263–283. Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480–498. Mackie, D. M. (1987). Systematic and nonsystematic processing of majority and minority persuasive communications. Journal of Personality and Social Psychology, 53, 42–52. Mazliach-Liberman, G. (2003). The Role of Cognitive Structuring as a Moderating Variable among Expertise, Self Epistemic Authority, Self EYcacy and Performance. Unpublished doctoral dissertation. Tel-Aviv University. McGuire, W. J. (1969). The nature of attitudes and attitude change. In G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology (2nd ed., Vol. 3, pp. 136–314). Reading, Mass: Addison-Wesley. Mikulincer, M. (1998). Attachment working models and the senses of trust: An exploration of interaction goals and aVect regulation. Journal of Personality and Social Psychology, 74, 1209–1224. Mikulincer, M., & Shaver, P. R. (2003). The attachment behavioral system in adulthood: Activation, psychodynamics, and interpersonal processes. In M. P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 35, pp. 53–152). San Diego: Academic Press. Mikulincer, M., & Shaver, P. R. (in press). Contributions of attachment theory and research to motivation science. In J. Shah & W. Gardner (Eds.), Handbook of Motivation Science. New York: Guilford Miller, N., Maruyama, G., Beaber, R. J., & Valone, K. (1976). Speed of speech and persuasion. Journal of Personality and Social Psychology, 34, 615–624. Newcomb, T. M. (1943). Personality and Social Change. New York, NY: Dryden. Payne, J. W. (1976). Task complexity and contingent processing in decision making. An information search and protocol analysis. Organizational Behavior and Human Performance, 16, 366–387. Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 19, pp. 123–205). San Diego, CA: Academic.

392

ARIE W. KRUGLANSKI et al.

Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a determinant of argument-based persuasion. Journal of Personality and Social Psychology, 41, 847–855. Pierro, A., & Mannetti, L. (2004). Motivated Consumer Search Behavior: The EVects of Epistemic Authority. Unpublished manuscript. University of Rome ‘‘La Sapienza.’’ Pierro, A., Mannetti, L., Kruglanski, A. W., & Sleeth-Keppler, D. (2004). Relevance override: On the reduced impact of ‘‘cues’’ under high motivation conditions of persuasion studies. Journal of Personality and Social Psychology, 86, 251–264. Putnam, R. D. (2000). Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster. Raviv, A., Bar-Tal, D., Raviv, A., & Abin, R. (1993). Measuring epistemic authority. European Journal of Personality, 7, 119–138. Raviv, A., Bar-Tal, D., Raviv, A., & Ben-Horin, A. (1996). Adolescent idolization of pop singers: Causes, expressions and reliance. Journal of Youth and Adolescence, 25(5), 631–650. Raviv, A., Bar-Tal, D., Raviv, A., & Houminer, D. (1990a). Development of epistemic authority perception. British Journal of Developmental Psychology, 8, 157–169. Raviv, A., Bar-Tal, D., Raviv, A., & Peleg, D. (1990b). Perception of epistemic authorities by children and adolescents. Journal of Youth and Adolescence, 19, 495–510. Raviv, A., Bar-Tal, D., Raviv, A., Biran, B., & Sela, Z. (2003). Teachers’ epistemic authority: Perceptions of students and teachers. Social Psychology of Education, 6, 17–42. Rogers, C. R. (1951). Client-Centered Counseling. London: Constable. Rogers, C. R. (1967). On Becoming a Person. London: Constable. Ruble, D. N., & Seidman, E. (1996). Social transitions: Windows into social psychological processes. In E. T. Higgins & A. W. Kruglanski (Eds.), Social Psychology: A Handbook of Basic Principles (pp. 830–856). New York: Guilford. Sears, D. O. (1969). Political behavior. In G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology (Vol. 5, pp. 315–458). Reading: Mass: Addison-Wesley. Sears, D. O., & Friedman, J. L. (1967). Selective exposure to information: A critical review. Public Opinion Quarterly, 31, 194–213. Sebald, H. (1986). Adolescents’ shifting orientation toward parents and peers: A curvilinear trend over recent decades. Journal of Marriage and the Family, 48, 5–13. Shafer, J. B. P. (1978). Humanistic Psychology. Englewood CliVs, NJ: Prentice Hall. Shapira, B. (1993). The Development of Political Preferences among Youths: The EVects of Parents’ Perception as Epistemic Authorities. Unpublished doctoral dissertation. Tel-Aviv University. Spiegel, L. A. (1958). Comments on the psychoanalytic psychology of adolescence. Psychoanalytic Study of the Child, 13, 296–308. Staton, T. F. (1963). Dynamics of Adolescent Adjustment. New York: Macmillan. Webster, D. M., & Kruglanski, A. W. (1994). Individual diVerences in need for cognitive closure. Journal of Personality and Social Psychology, 67, 1049–1062. Webster, D. M., & Kruglanski, A. W. (1998). Cognitive and social consequences of the need for cognitive closure. European Review of Social Psychology, 8, 133–161. Weiss, W., & Fine, B. J. (1955). Opinion change as a function of some intrapersonal attributes of the communicates. Journal of Abnormal and Social Psychology, 51, 246–253. Weiss, W., & Fine, B. J. (1956). The eVect of induced aggressiveness on attitude change. Journal of Abnormal and Social Psychology, 52, 109–114. Winnicott, D. W. (1965). The Maturational Processes and the Facililtating Environment. New York: International Universities Press. Wintre, M. J., Hicks, R., McVey, G., & Fox, J. (1988). Age and sex diVerences in choice of consultant for various types of problems. Child Development, 59, 1046–1055.

INDEX

conceptual objections to, 14–35 contributions of, 6 failure to address explanations, 17–19 importance of addressing controversies in, 9–11 insight into cognitive processes through, 1 necessity of component assessment, 56 political objections to, 29–30 value of, 5–9 Adolescence, sources of epistemic authority in, 353 Adult caregivers children’s attribution of epistemic authority to, 361 and developmental epistemic authority, 357–362 as ultimate epistemic authorities, 353 Affective mediators, 322 of contact effects, 284–285, 308–309 Affirmative action, 114 contemporary initiatives to outlaw, 96 Agreement among multiple perceivers as basis of accuracy, 47 behavior and, 46 with experts, 47–48, 49 with independent judges, 48 interaction with level of self-ascribed epistemic authority, 355 limitations to, 49–52 with models, 48 with nonindependent judges, 48–49 with other perceivers, 49 with others, 47–52 agreement with others, 39 with targets, 52–55 with targets’ self-perceptions/self-reports, 53–54

A Accommodation, as mediator of contact effects, 304 Acculturation orientations, and intergroup attitudes, 315–317 Accuracy arousal of political motivations by, 30–32 assessing with statistical techniques, 41 as cause of agreement, 47 coexistence with bias, 16–17, 53 common criticisms of, 76 componential vs. noncomponential approaches to, 74 –75 defined, 4 and discrimination, 33–35 empathic, 18 enhanced by bias, 35 integration with bias, 80–81 models of, 14–15 non-absolute nature of, 5 noncomponential approaches to assessing, 66–74 overall levels of, 41–42 phenomenology of, 12–13 probabilistic conception of, 1 of social perception, 1–4, 36–40 of stereotypes, 29–30 vs. morality and legality, 13 vs. process, 62, 77 vs. self-fulfilling prophecy, 19–23 vs. social justice, 35 Accuracy interventions, 82–84 Accuracy research alleviation of social problems via, 33 blossoming of, 79–85 common criticisms of, 10 component assessment in, 62–66 393

394

INDEX

American exceptionalism, 138 Anger, moving to cooperative behaviors from, 190–193 Anglo-Japanese attitudes study, 288 ANOVA factors, 61 Antecedent conditions, 227, 232 for groupthink, 219, 220, 222, 238, 244 Anti-Black affect, blend with traditional conservative values, 118, 135 Anti-Semitism, 256 Anxious attachment, trust and epistemic authority in, 353–354 Assimilation as acculturation strategy, 315–317 avoiding via Hewstone-Brown intergroup contact model, 266 as preference of majority groups, 316 Attachment theory, and development of epistemic authority, 353–354 Attention-grabbing power, of error/bias vs. accuracy, 15–16 Attitude change, role of trustworthiness vs. expertness in, 347 Attitudinal homogeneity, 242 Authority eight sources of, 364 emergence and effect on cooperation, 180 leftist-rightist leanings and criticism of, 384 legitimizing effect of trust on, 183–184 relational model of, 160–161, 191–192 role of procedural justice in gaining voluntary cooperation for, 202 Authority assignments, distribution across sources, 345–346 Authority gap, 378 Aversive racism, 97 Avoidant attachment, trust and epistemic authority in, 353–354, 388

Belief, correspondence with reality, 3–4, 11–12 Belongingness, 173, 196, 206 degree of, 162 moderating effect in context of respect/ disrespect, 183 and procedural fairness, 164–166 and respect/disrespect, 182 Bias, 7, 12, 15–16, 53, 54 accuracy enhanced by, 35, 77 coexistence with accuracy, 16–17, 53, 77 integration with accuracy, 80–81 in judgment, 64 and person perception, 24–26 power of, 80 relative importance of, 14–16 and stereotypes, 31 trustworthiness and procedural fairness in, 184 validity of perspectives emphasizing, 8 Big five factors questionnaire, 173 Blackindividualism, 107, 119, 121, 134, 136, 137 Black political candidates, 100 Blacks, racial prejudice and symbolic racism, 95–97 Blacksheep effect, 241 Blaming the victim, 31, 32 Bookkeeping model of stereotype change, 268 Brewer-Miller decategorization model, 256, 262–264, 326 evidence from educational settings, 309 role of cross-group friendships in, 291 in studies of children with disabilities, 313 Brown, Rupert, 255 Brunswik’s Lens Model, 69–71, 75. See also Lens Model Bush Doctrine, 242 Bush White House, invasion of Iraq by, 221 Busing, 96, 114

B

C

Bar-Tal, Daniel, 345 Baron, Robert S., 219 Bayes’ theorem, 17 Bayesian inference, 50–52 Behavior agreement with targets’ self-reports of, 53 defined, 45 regulating effectively in groups, 204 Behavior criteria, 39–40

Category salience, 319–320 effect of intergroup anxiety on, 287 moderation effects of, 283 Causality, flow from target behavior to perceiver beliefs, 21–22 Children with disability, 321 intergroup contact studies of, 311–315 Civic apathy, 346 Classical liberal values, 121

INDEX Close-mindedness, in groupthink, 221 Closure. See also Need for closure need for, 354–355, 370–371 Cognition, relationship to motivation, 158, 197 Cognitive conflicts, avoidance by need-for-closure individuals, 371 Cognitive empathy, 323 Cognitive mediators, of contact effects, 284–285 Cognitive processes insights provided by accuracy research, 1, 11 relative importance of, 14–16 Cognitive Zeitgeist, 348 Coherence, vs. correspondence, 11–12 Cohesion, 226, 227. See also Group cohesion measures of, 224 Collective amnesia, 145 Collective goals, 156 motivating individuals to adopt, 197 Collective guilt, 323 Collective interest, 174 Collective-oriented goals, 174 College education, correlation with symbolic racism, 95–96, 116–118 Commitment, stimulating, 191 Common in-group identity model, 256. See also Gaertner-Dovidio common in-group identity model advantage over decategorization model, 265 and personalization, 326–329 Communicator credibility, 347 Communicator influence incentives for, 347 in learning theory paradigm, 346 Competitive actions cooperation restraining, 155 vs. prosocial and individualistic orientations, 177 Component assessment, 56 necessity in accuracy research, 62–65 Componential approaches, 56, 64, 77 Cronbach’s components, 56–58 Judd and Park’s full accuracy design for research on stereotypes, 60–61 Kenny’s social relations model, 58–59 measuring and eliminating noise via, 61–62 vs. noncomponential approaches, 74–75 Componential models, 14–15 use for criteria of self-reports and self-perceptions, 66

395

Components, avoiding reification of, 64–65 Compulsive self-reliance, in avoidant attachment patterns, 354 Conceptual knowledge, child’s development of, 353–354 Concurrence seeking, 221, 227, 232, 236 and hidden profile paradigm, 233, 234 Conditioned stimuli, in context of groupthink, 240 Conflict-prone search strategy, 370 Conflict resolution, 158 subjective elements of, 330–331 Conformity, 239 in development of groupthink, 228–231 Conservative traditional values, 96, 134 blend with anti-Black affect, 118–119 and need for closure, 384 Consistency, rule of, 164 Construct validity. See also Scientific construct validity and correlational approaches to accuracy, 69–71 and social perceptual accuracy, 37–39 Contact cognitive analysis of intergroup, 267 dimensions of, 318–319 effects of quality and quantity of, 280–283 opportunities for, 297, 318–319 Contact avoidance, and intergroupanxiety, 286 Contact effects cognitive and affective mediators of, 284–285 intergroup anxiety as mediator of, 285–291 mediators of, 298–301 miscellaneous negative and positive mediators of, 292–302 moderated mediation of, 310–311 simultaneous moderation and mediation of, 302–308 Contact Hypothesis, 256, 257, 315 and extended contact effects, 289–291 history of, 258–261 Northern Ireland challenges to, 288 theoretical developments of, 261–267 Conversion model of stereotype change, 268 Cooperation instrumental approaches to motivating, 204–205 intrinsic motivations for, 194–195 and leader procedural fairness, 182

396

INDEX

Cooperation (continued ) leadership strategy designed to encourage, 201 managing with procedural justice, 151–154 motivating, 152, 204–205 as productive effort on behalf of groups, 191 real-world, 203–204 relationship to procedural fairness, 154–158 role of trust in facilitating, 176 as rule following, 191 self-based model of, 196 social psychology literature on, 196 stimulating, 191 voluntary, 200–201 Cooperation research, 180–181 procedural fairness and prosocial goals, 181–183 procedural fairness and trust, 183–185 Cooperative interaction, Contact Hypothesis and, 258 Correlational approaches, 67–69 and construct validity, 69–71 in intergroup contact studies, 275–278 Correspondence, 4 vs. coherence, 11–12 Credibility, role of expertness and trustworthiness in, 347 Crisis emotions associated with, 246 and low situational self-efficacy, 246 need for rapid and decisive action in, 245 role in development of groupthink, 227 Criteria for accuracy, 76 agreement with self-perceptions, 53–54 agreement with targets’ self-descriptions, 52–53 four classes of, 39–40 imperfect, 43 suboptimal weighting of, 74 for validity of social perception, 10 Critical realism, 2 Critics, group insulation from, 221 Cronbach’s componential model, 14, 56–58, 62, 63, 65, 68 Cross-group friendships, 259, 289, 292, 294, 297, 318, 319, 328. See also Outgroup friendships in Northern Ireland, 291 Cross-sectional studies

children with disabilities, 311–315 in intergroup contact research, 281 Crossed categorization, research on effects of, 262–263 Cue function, in informational sources, 349 Cues in Brunswik’s Lens Model, 69–71 in Dawes’ improper linear models, 73–74 D Dawes’ improper linear models, 73–74, 75 De Cremer, David, 151 Decategorization model, 256. See also Brewer-Miller’s decategorization model correspondence of marginalization strategy to, 315–317 limitations for prejudice reduction, 264 in studies of children with disability, 313 Decategorized contact, 262 Deception, underestimation of, 42 Deception-detection literature, 42 Decision making defective processes in groupthink, 221, 245 epistemic authority effects on, 356–357 groupthink in, 219–227 role of procedural fairness in, 185 and social cognition, 6 Decision polarization, 232 Decision quality, 226, 228 relationship of crisis and cohesion to, 223 Defection, vs. collective interest, 175 Democratic processes, risk during crises, 245 Denial-of-difference ideology, 31, 32 Desirable behaviors, strategies for stimulating, 190–193 Differential accuracy, 57, 63 Differential elevation, 57, 63 Dimensions of contact, 318–319 Direct contact, 285 mediation of contact effects by, 294–296 Directive leadership, 226, 227 in crisis situations, 246 impact on decision making, 223–224 Discrimination denial of continuing, 102, 109, 113, 114 as result of accuracy, 1, 33–35 Dispersion inaccuracies, 61 Disrespect, within groups, 165

INDEX Dissent, suppression of, 226, 228, 229, 231, 235, 236 Distributive justice, 188 shift to procedural justice from, 186–187 Diversity, as accuracy intervention, 82 Dual-identity strategies, 328 correspondence of integration acculturation model to, 315–317 outcomes in intergroup contact, 273 Dynamic interactionist approach, 163 E Economic individualism, vs. Black individualism, 119–120 Education as accuracy intervention, 82 contribution to racial prejudice, 116–118 Educational settings, intergroup contact research from, 309–315 Efficient markets theory, 49–50 Egalitarian values, vs. symbolic racism, 137–141 Elaboration likelihood model, 348–349 Elevation, 57, 58, 59, 63, 67 Elevation inaccuracy, 61 Emotion, as integral facet of motivation, 198–200 Emotional empathy, 323 Empathic accuracy, 18 Empathy, 322 and manipulation of salience, 273–275 as mediator of contact effects, 292–293, 297, 306 Endogenous features, in persuasion, 374–376 Endogenous switching regression models, 277 Enduring personality traits, 43 self-perceptions of, 66 Entitativity, 238 Epistemic authority, 346, 351–352 and ability to learn from experience, 377–380 adult caregivers as ultimate sources of, 353 ascriptions in political domain, 382–384 and attachment theory, 353–354 beliefs in real-world contexts, 381–386 children’s attribution of, 357–359 developmental differences research in, 357–362 differentiation between internal and external sources of, 370–371

397

effects of self-ascribed, 371–376 effects on information processing and decision making, 356–357 empirical research in, 357–386 gender differences in assignment of, 365 generalized vs. specialized, 351 individual differences in assignment of, 364–369 internal vs. external, 370–371 motivational and cognitive factors in assignment of, 354–356 nondirectional motivational effects, 389 of parents in political knowledge, 384–385 perceived gap in, 380 of pop idols, 362–363 of professors in statistics and psychology, 385–386 reliance on external vs. internal, 388 and seize-and-freeze mechanism, 352, 367 self as target of, 352 self-ascribed, 370–371, 377–380 situational determinants of ascriptions of, 388 social development of, 353–354 as stopping mechanism, 352 vs. heuristic cues, 366–369 Epistemic authority assignments determinants of, 354–356 individual differences in, 364–369 Epistemic authority effects, in social judgment, 345–346 Equal opportunity programs, 141 Error, 7, 12, 15–16, 53, 54 power of, 80 relative importance of, 14–16 and stereotypes, 31 Excessive demands, role in symbolic racism, 103, 107–108, 109 Expectancy effects, 16 Experience, efficacy as mediator of learning, 381 Expertness as communicator incentive, 347 persuasive effect under low issue involvement conditions, 349 and veridical self-perception of epistemic authority, 374 Experts agreement with, 47–48 limitations to agreement with, 49

398

INDEX

Explanations, accuracy of, 17–19 Extended contact, 285, 289–291 assessing benefits of, 319 as mediator of contact effects, 296, 302 External rewards, vs. intrinsic motivation to cooperate, 197 External threats, as condition for groupthink, 221 F Fairness heuristic theory, 156, 184 False consensus effect, 16 Fathers children’s attribution of epistemic authority to, 357–359 epistemic authority in political arena, 385 Fear, reducing influence in social dilemmas, 175, 176 Forgiveness, 324 and collective guilt, 323 as moderator of intergroup contact effects, 307 Formal knowledge, children’s attribution of epistemic authority in domain of, 360 Friends, children’s attribution of epistemic authority to, 357–359, 360, 361, 362 Friendship, 326 association with reduced anxiety, 286 and contact effects, 285 Future planning, children’s attribution of epistemic authority in domain of, 357–358, 359–361 G Gaertner-Dovidio common in-group identity model, 256, 264–265, 328 correspondence of assimilation strategy to, 315–317 Gay men empathy and group salience in attitudes towards, 273–275 intergroup anxiety in contact with, 289 mediator and moderator effects on intergroup contact with, 307–308 Gender individualism, vs. Black individualism, 121 General life knowledge, children’s attribution of epistemic authority in domain of, 360

General Social Survey (GSS), 105 Generalized epistemic authority, 351 Generalized outcomes for intergroup relations, 318 maximizing, 322 Goal/expectation theory, 175 Goal transformation, trust and, 174–180 Government, loss of reliability as informational source, 346 Government antiracial policies, 114–115 Greed, reducing influence in social dilemmas, 175 Group behavior, managing with procedural justice, 151–154 Group cohesion, 220, 222, 223 and suppression of dissent, 235 Group conflicts, failure of distributive justice to resolve, 186 Group consciousness and symbolic racism, 105 of Whites, 129–132 Group differences awareness as mediator of contact effects, 284 in perception and outcomes, 35 Group discussion and development of groupthink, 232 and self-censorship, 232–235 Group engagement model, 152, 191, 192, 193 of procedural fairness, 161–162 Group feedback, cooperation as function of, 180 Group identification, 178 cooperation as function of, 179, 180 in groupthink, 224 as moderator of intergroup contact effects, 307 Group insulation, 227 from outside influences, 221 Group interaction, 151 Group judgments, polarization due to intergroup anxiety, 286 Group member insecurity, 224, 246 and development of groupthink, 222 Group membership, 151, 320 interdependence of, 151 and self-definition, 154 and social identity development, 262 Group norms, 242–243 belief in, 229 and social identification, 240

399

INDEX Group position, sense of, 128–129, 138 Group process, social influence and, 220 Group rules, 184 Group salience, 276, 302, 307, 318, 329 benefits for intergroup contact, 317–318 in children with disabilities, 314 empirical evidence for moderator variables, 278–282 experimental studies of, 270–275 in Gaertner-Dovidio common ingroup identity model, 264–265 Hewstone-Brown model of, 258, 266–267, 312 intergroup anxiety as by-product of, 258 with interpersonal vs. intergroup contact, 327 manipulating conditions of, 271–273, 275–276 moderating effects of, 283 moderation of empathy by, 323 as moderator of effects of contact on prejudice reduction, 257, 275–283 predicted by intergroup anxiety, 288 role in prejudice reduction, 264 and tendencies toward intergroup differentiation/assimilation, 262 timing of introduction, 271–272 Group-value model, 191 of procedural fairness, 159–160 Group value model, of procedural fairness, 152 Groupthink antecedent conditions for, 219 factors amplifying, 246–247 history of, 219–227 negative stereotyping in, 221 outcome variables of, 220 paucity of empirical studies on, 221–222 pervasiveness of, 238 in polarized group decision making, 219 symptoms of, 220 ubiquity model of, 219, 227–238, 238–239, 245–247 H Henry, P. J., 95 Heuristic cues, 366–369 Heuristic mode, 349 Heuristic-Systematic Model, 348–349, 349–350 Heuristics, role in increasing accuracy, 81

Hewstone, Miles, 255 Hewstone-Brown intergroup contact model, 258, 266–267, 328 correlation of separation strategy to, 315–317 evidence from educational settings, 309, 313 and high-salience conditions, 312 Hidden profile effect, 232–235 Hierarchical regressions, in symbolic racism literature, 127 Hierarchy of Epistemic Authorities Test, 364–365 High attribute elevation, 61 Homogeneity of background, 222 perceived, 319–320 I Ideal self, 199 Identity-based needs, 154 Identity judgments, 161–162 in group engagement model, 192 influence of outcomes on, 170–171 Ideological conservatism, 99 Illusion of consensus, 226, 227, 236–237 Immigrants, effects of contact on judgments regarding, 303, 306, 321 Impartial leadership, 221 Imperfect criteria, 43, 48 Implicit beliefs, accuracy of, 81–82 Implicit racism, 97 and social desirability, 143–144 Incentives, limitations of, 194, 195 Inclusion, messages of, 161, 164 Independent judges, agreement with, 48 Independent of influence, 4 Indirect contact. See Extended contact Individual difference scale, 168 need to belong measured in, 164 Individual differences, as moderators of accuracy, 79 Individualism, 107 beyond, 137 black, 96 blend with anti-Black affect, 135 racialized, 145 vs. prosocial orientation, 177 Individuating information, 17, 79, 329 as cognitive mediator of contact effects, 284–285

400

INDEX

Individuation, 304, 305 Inegalitarianism, 139 Information processing, epistemic authority effects on, 356–357 Information search, inadequate in groupthink, 221 Informational dependence, 345–346 Informational impact, as function of selfascribed epistemic authority, 373–374 Informational source effects, 346 elaboration likelihood model of, 348–349 heuristic-systemic model of, 349–350 in social development stages, 353–354 unimodel of, 350–357 Ingroup, overestimation in groupthink, 221 Ingroup affinity, 129, 132, 133 vs. Black antagonism, 134 Ingroup bias effect, 61, 237, 246, 272–273 manipulation of group salience and, 275–276 pervasiveness of, 262 Ingroup defensiveness, 227 Ingroup deviance, 241 Ingroup favoritism, reduction through personal friendships with outgroup members, 263 Ingroup insulation, 224 impact on decision quality, 223 Ingroup messages, 239, 240 Ingroup-outgroup boundaries, erosion vs. maintenance of, 266 Injustice, emotions of, 199 Instrumental needs, role of groups in meeting, 154 Integration, as acculturation strategy, 315–317 Intelligence tests. See IQ tests Intensification of attitude, 232 Interaction, socioemotional aspects of, 188 Interdependent self-construal (ISC), 169, 173, 195, 206 Intergroup anxiety, 257, 277, 302, 307, 322 as by-product of heightened group salience, 258, 297 correlation with high group salience, 304 and encounters with gay men, 307–308 in Hindu-Muslim intergroup contact, 295 as mediator of intergroup contact effects, 285–291 moderated by interpersonal vs. intergroup contact, 327

reduction by subgroup salience, 328 research on, 287–291 Intergroup attitudes and acculturation orientations, 315–317 indirect measures of, 324 Intergroup conflict and acculturation orientations, 315–317 and correlational studies of group salience, 275–283 evidence from educational settings, 309–315 and experimental group salience studies, 270–275 and experimental studies of stereotypedisconfirming information, 267–270 history of contact hypothesis in, 258–261 integrative theory of, 255–258, 317–331 and mediators of contact effects in, 284–308 theoretical development of contact hypothesis, 261–267 Intergroup contact, 330 affective processes in, 258 cognitive analysis of, 267–270 experimental studies of, 278 importance of conditions of, 261 limitations in ultimate impact of, 330 mediation of effects of, 322–323 moderation hypothesis of, 276 moderational role of, 319–322 quality of, 280, 282–283 quantity of, 280, 282–283 with support of institutional authority, 258 under varying conditions of typicality and outgroup homogeneity, 271 vs. interpersonal contact, 287, 326 Intergroup contact theory, 318–326 Intergroup salience, leading to favorable intergroup attitudes, 275 International conflict, 256 Interpersonal accuracy, 1 Interpersonal contact, 287 benefits of, 329 vs. intergroup contact, 326 Interpersonal perceptions, 16 Interpersonal relatedness, 155 coexistence with intergroup salience, 291 Interpersonal treatment, focus in justice research, 187–190 Interpretation, 46 Intragroup homogeneity, 262

401

INDEX Intrinsic motivation, among group members, 154 IQ tests, 43, 44 Iraq, invasion by Bush White House, 221 Irrationality, 50, 51 in human nature, 6–7 validity of perspectives emphasizing, 8 Islamophobia, 256 J Jim Crow racism, 96, 97, 98, 124 Journalists, ascriptions of epistemic authority to, 383–384 Judd and Park, 65 componential approaches of, 64 full accuracy design for research on stereotypes, 60–61 Judgment polarization, 232 Judgmental confidence, degrees correlated to epistemic authority attribution, 368–369 Judgments, using components to measure and eliminate noise in, 61–62 Jussim, Lee, 1 Justice research, 152, 196 focus on interpersonal treatment issues, 187–190 movement from anger/negative behaviors to cooperative behaviors, 190–193 self and, 186 shift from distributive to procedural justice, 186–187 K Kenny’s Social Relations Model, 65, 66 Kruglanski, Arie W., 345 L Laissez-faire racism, 97 Leader Fairness Theory (LFT), 200–201 Leader procedural fairness, 181–183 Leadership literature, future directions for, 200–201 Learning from experience, and self-ascribed epistemic authority, 377–380 Learning theory paradigm, of source effects, 346–348 Legal and moral issues, 13

Length of relationship, as moderator of accuracy, 79–80 Lens Model, 14, 69–71 Levels of analysis, in studies of social perceptual accuracy, 25 Liberal racial policies, 139 Likability as communicator incentive, 347 as factor in increasing persuasion, 383 Logical coherence, 11–12 distinguishing from accuracy, 1, 11 Longitudinal contact. See also Pettigrew’s theory of longitudinal contact theory of, 256 Longitudinal evidence, wanting in symbolic racism studies, 135–137 Longitudinal studies, in intergroup contact research, 277–278, 281–282 Los Angeles County Social Surveys (LACSS), 105, 115, 130, 139 two-factor model dominant ideology and political resentment, 112 Low situational self-efficacy, 243–244, 247 as condition for groupthink, 221 M Majority groups, acculturation strategies of, 316–317 Mannetti, Lucia, 345 Marginalization, as acculturation strategy, 315–317 Material needs, role of groups in meeting, 154 Mean differences, between judgments and criteria, 67 Measurement to establish accuracy of social perceptions, 55 unreliability in, 44 Mediational analyses, 284–308, 318 Mediator variables, in intergroup contact studies, 276 Message arguments, 349, 350 Methodological techniques for distinguishing accuracy from self-fulfilling prophecy, 20–21 in intergroup contact studies, 277–278 Minority views expression of, 235 favoring integration strategy, 315–316

402

INDEX

Mixed-motive conflicts, 152, 175, 193, 203 predicting cooperation in, 175–176 Models of accuracy, 14–15 agreement with experts’, 48 Cronbach’s components, 56–58 Judd and Park’s full accuracy design for research on stereotypes, 60–61 Kenny’s social relations model, 58–59 limitations to agreement with, 49 noncomponential, 66–75 prescriptive, 75 Moderated mediation effects, 320 correlational evidence of, 302–308 Moderators, 79–80 vs. mediator variables in intergroup contact, 276 Modern racism, 97 Moral and legal issues, 13, 51 Mothers children’s attribution of epistemic authority to, 357–359, 360 epistemic authority in political knowledge, 385 Motivation defined, 152, 153 emotion as integral factor of, 198–200 implications of social self for, 156, 158–159 intrinsic among group members, 154 relationship to cognition, 158, 197 role in epistemic authority attribution, 362 toward group engagement, 161 Motives, integrating into social self, 162 Multiculturalism as accuracy intervention, 82 and stereotype accuracy, 31–32 N National Election Studies (NES), 105, 115, 138 Near misses, 4 Need for closure, 389 among conservatives vs. liberals, 384 external information search under, 371–373 momentarily induced, 355–356 and reliance on external source of epistemic authority, 370, 371 role in epistemic authority assignments, 354–355 Need for Closure Scale, 372

Negative affect, transference to outgroup members, 286 Negative behaviors, moving to cooperative behaviors from, 190–193 Negative stereotyping, in groupthink, 221 Neuroticism, role in procedural fairness issues, 173 Noise, eliminating in judgments, 61, 63 Noncomponential approaches, 66–67, 77 Brunswik’s Lens Model, 69–71 construct validity and correlational approaches, 68–69 correlational approaches, 67–69 Dawes’ improper linear models, 73–74 realistic accuracy model, 71–73 Nonindependent judges, agreement with, 48–49 Nonracial issues, racialization of, 142–143 Nonracial political attitudes, 118–119, 137 Nonverbal behaviors, 46 Nonverbal sensitivity, test of, 42 Northern Ireland conflict resolution in, 331 direct contact effects in, 294–296 extended contact and intergroup relationships in, 290 intergroup anxiety in, 288 outgroup friendships in, 260 Numerical aptitude test, 378–380 O Objective criteria, 40 independent and standardized, 43–45 one-shot hit or miss, 40–41 overall levels of accuracy, 41–42 Observable action, as basis of social perceptions, 41 Old-fashioned racism factor analyses of, 125 vs. symbolic racism, 123–124 One-shot hit or miss criteria, 40–41 Opinion polarization, 236 Ought self, 199 Outcomes group differences in valued societal, 34 managing uncertainty about, 170–173 measures in intergroup contact research, 324–326 predictions of, 73–74

INDEX Outgroup antagonism, 132, 133, 241 in ordinary and minimal groups, 237–238 vs. White group consciousness, 128–132 Outgroup attributes, stereotyping of, 246 Outgroup biases, 237, 272–273 moderated by category salience or typicality, 275 Outgroup evaluation, 307 Outgroup friendships, 296 and intergroup contact with gay men, 307–308 in Northern Ireland, 260 and prejudice reduction, 259, 264 Outgroup members achieving greater differentiation among, 262 decategorizing and personalizing interactions with, 262–264 individuation of, 284–285 trust issues regarding, 178 Outgroup stereotypes, 226, 228 Outgroup trust and forgiveness, 276 Outgroups, generalized ratings of, 271 P Pancritical realism, 2 Parental authority. See also Adult caregivers epistemic authority in realm of political knowledge, 384–385 Party identification, 142. See also Republican party Pastimes, children’s attribution of epistemic authority in domain of, 357–358, 359–361 Peer influence and epistemic authority, 357–372 gender differences in epistemic authority assignment to, 365 Perceived collective threat, 129–130 as mediator of contact effects, 292 Perceived typicality, 269, 319–320, 321. See also Typicality Perceiver judgments, impossibility of evaluating accuracy of, 27 Peripheral cues, 350 Person perception, 25–28 Bayesian inference vs. stereotype biasing, 50–52 and beliefs about groups, 52

403

Personal feelings, children’s attribution of epistemic authority in domain of, 357–358, 359–361, 362 Personality traits, 43 Personalization in Brewer-Miller decategorization model of intergroup contact, 262 and common ingroup identity, 326–329 Perspective-taking, 304, 305, 322, 329 Persuasion and attitude change, 346–347 as cognitive activity, 348 by exogenous features, 374–376 and issue involvement conditions, 349 models of, 346 modes of, 350 Pettigrew’s theory of longitudinal contact, 256 Philosophical homogeneity, 242 impact on decision quality, 223 Physical appearance, children’s attribution of epistemic authority in domain of, 358, 359–361, 362 Pierro, Antonio, 345 Pluralistic ignorance, 236–237 Polarization of attitude, 228 Political attitudes extra-academic sources of, 382–384 parents’ epistemic authority in realm of, 384–385 Political disengagement, 346 Political effects, of symbolic racism, 109–118 Political ideology, vs. symbolic racism, 124–126 Political objections, 84–85 to accuracy of stereotypes, 29–30, 52 Political science, importance of motivating cooperation to, 2004 Political socialization, studies of epistemic authority beliefs in, 381–382 Politicians, ascriptions of epistemic authority to, 383–384 Politics, children’s attribution of epistemic authority in domain of, 359–361 Pop idols, as epistemic authorities, 362–363 Postdecision voice, 160 Preadult socialization, in symbolic racism, 136 Predecision voice, 160 Prejudice reduction and Contact Hypothesis, 258 effect of intergroup contact on, 259 and group salience, 257

404

INDEX

Prevention focus, 199–200 Prior expectations, 51 Privilege legitimatization of, 97, 129 maintenance of, 101 threatened sense of White, 105 Probabilism, 1, 72 Probabilistic realism and Realistic Accuracy Model, 71–73 and social perceptual accuracy, 2–4 Procedural fairness future research directions for, 198–200 as group tool for managing cooperation, 151 interpersonal aspects of, 188 motivational analysis, 159 and prosocial goals, 181–183 relationship to cooperation, 154–158 and trust, 183–185 Procedural fairness models, 152, 158–159 group-engagement model, 161–162 group-value model, 159–160 including emotional components in, 199 relational model of authority, 160–161 Procedural fairness research, 163–164 belongingness and procedural fairness, 164–166 self-definition and procedural fairness, 168–170 self-uncertainty and procedural fairness, 170–173 social reputation and procedural fairness, 166–168 Procedural justice and regulation, 201–202 role in shaping motivation, 153 shift from distributive justice to, 186–187 Procedural justice literature, 203 Procedures, link to self-related consequences, 162 Process, 84 vs. accuracy, 62–63, 77 Process control literature, 188 Professors, ascriptions of epistemic authority to, 383–384 Promotion focus, 199–200 Proself orientation, 177 Prosocial goals, 176 and procedural fairness, 181–183 Prototype-subtyping model of stereotype change, 269

Psychotherapy, as accuracy intervention, 83 Public good dilemmas, 176 games of, 178, 182 Q Quality of contact, 280, 281–282, 283, 287, 319 Quantity of contact, 280–283, 287, 319 and realistic threat in Hindu-Muslim intergroup relations, 295 R Race-based policies popular and unpopular, 141–142 White opposition to, 114–115, 117 Racial ambivalence, 97, 134 Racial conservativism, and symbolic racism, 95 Racial equality, 145 vs. social dominance, 137–141 Whites’ continuing resistance to, 96 Racial policy, predicting White opposition to, 131 Racial policy preferences, 122 accounted for by racial and nonracial attitudes, 127 effect of Black individualism on, 121 outgroup vs. ingroup factors in, 133 of Whites, 100 Racial politics, effects of symbolic racism on, 95 Racial prejudice, 96, 101 symbolic racism grounded in, 137 Racial status quo, legitimization of, 101 Racial stereotypes, accuracy research and, 30–32 Racialization, of nonracial issues, 142–143 Racism, implicit, 97 Raviv, Alona, 345 Raviv, Amiram, 345 Realism, defined, 2–3 Realistic Accuracy Model, 15, 71–73, 75 Realistic group conflict theory, 128–129, 138 Realistic threats, 292 Reality, relationship to belief, 3–4 Reciprocal self-disclosure, 296–297 Reflection-Construction Model, 15 Reframing, as accuracy intervention, 82 Regression principles, 17, 52 Regression to the mean, 49–50 Regulatory focus theory, 199, 200

INDEX Relatedness, need for, 162 Relational model of authority, 152, 191, 192 in procedural fairness models, 160–161 Religious conservativism, and symbolic racism, 137 Republican party, 143 shift of White Southerners to, 101 Respect, 167 central to status messages, 190 within groups, 165 as quality of procedural fairness, 182, 183 Reverence effect, 378, 381 Rewards, role in motivating cooperative behavior, 154 Rule following, 191, 193, 201–202 role of trust in, 184 Rules and laws, children’s attribution of epistemic authority in domain of, 357–358 S Salience manipulation, in intergroup contact, 271 Salience moderation, 278–282 Sanctions failure to motivate intrinsic cooperation, 194 inefficiency of, 195 role in motivating cooperation, 154 in structural leadership solutions, 181 Scaling discrepancies, 68 School studies, children’s attribution of epistemic authority in domain of, 359–361, 362 Science, children’s attribution of epistemic authority in domain of, 357–358, 359–361, 362 Scientific construct validity, 1, 11 and social perceptual accuracy, 36–39 Sears, David O., 95 Secure attachment, trust and epistemic authority in, 353–354, 388 Seize and freeze, 352, 367 Selective exposure to information hypothesis, 383 Selective group attention, 235 Self, as target of epistemic authority, 352, 361 Self-ascribed epistemic authority (SAEA), 370–371 and ability to learn from experience, 377–380

405

differences in information processing by, 376–377 and external information search under need for closure, 371–373 informational impact as function of, 373–374 in mathematics, 378 and method of instruction, 379–380 and need of advice from others, 372 and persuasion by exogenous features, 374–376 and persuasiveness of external authority figures, 387 Self-based model of cooperation, 196 Self-censorship, 235 in group discussion, 232–235 in groupthink, 227 Self-concept clarity scale, 172, 206 Self-construal, 196. See also Interdependent self-construal (ISC) Self-definition, 163, 240 and procedural fairness, 168–170 Self-descriptions agreement with targets’, 52 and procedural fairness, 168 Self-determination theory, 197 Self-disclosure, 304, 321, 322, 326, 329 as mediator in intergroup contact effects, 295–296 reciprocal, 296–297 Self-doubt scale, 172, 206 Self-efficacy, 243–244 Self-esteem, 240 procedural fairness contributions to, 161 Self-fulfilling prophecy, 81 and the criterion problem, 38–39 distinguishing from accuracy, 1, 11, 19–23 Self-interest divergence from social interest, 155 sanctions on, 181 short-term, 203 vs. group goals, 175 Self-perceptions agreement with targets’, 53–54 limitations to agreement with, 54–55 use of componential models for, 66 Self-regulation, 202 in social dilemmas, 174 Self-reliance, in avoidant attachment patterns, 354, 388

406

INDEX

Self-reporting, 44, 52 agreement with targets’, 52–53 limitations to agreement with, 54–55 of opportunity for intergroup contact, 307 of racial attitudes, 144 use of componential models for, 66 Self-uncertainty, 163, 173, 196 degree within group, 162 and procedural fairness, 170–173 Sense of self, 151 Separation, as acculturation strategy, 315–317 Shared ideology, and development of groupthink, 227 Shared information, and hidden profile effect in decision making, 234 Sharvit, Keren, 345 Short-term self-interest, 203 Similarity as communicator incentive, 347 as factor in increasing persuasion, 383 Situational forces, as moderators of accuracy, 80 Small groups, accuracy and bias in, 24–26 Social allegiances, 240 Social attraction, in groupthink, 224 Social beliefs, probabilistic nature of, 3–4 Social cognition, 6 Social desirability, and implicit prejudice, 143–144 Social dilemma research, 153, 196 and procedural fairness, 174–180 structural solutions, 180 Social dominance, 97, 129, 139 theory of racism, 136, 138 Social emotions, 199 Social evaluation, role in self-construction, 167 Social identity theory, 237, 238, 239–242 Social inequalities, 79 exacerbated by accuracy research, 30, 31 Social influence, with low situational selfefficacy, 244 Social interest, divergence from self-interest, 155 Social judgment, 57, 58 componential approaches to, 65 epistemic authority effects in, 345–356 Social justice, vs. goals of accuracy, 35 Social needs, 154 Social perception accuracy in, 1–4

based on observable action, 44 criteria for validity of, 10, 36–40 measurements to establish accuracy of, 54 Social perceptual accuracy and construct validity, 37–39 defined, 2 levels of analysis in, 25 and probabilistic realism, 2–4 Social problems, created/alleviated by accuracy research, 1, 33, 79, 84 Social psychology literature, on procedural fairness, 196–198 Social reality, probabilistic nature of, 3–4 Social rejection, 240 Social relations, children’s attribution of epistemic authority in domain of, 357–358, 359–361 Social Relations Model (SRM), 15, 58–59, 63, 64 Social reputation, 162, 163, 173, 196, 206 and procedural fairness, 166–168 Social self, 155, 195 dynamic nature of, 156 and expectations of ingroup members, 178 future research directions for, 198–200 and leadership, 200–201 positively reinforcing in groups, 175 procedures directly affecting, 173 Social structural theories, 128–129 Social value orientation, 178 Source effects dual-mode perspective, 348–349 elaboration-likelihood model of, 348–349 heuristic-systematic model of, 348–349, 349–350 similarities with message effects, 351 unimodel of, 340–357 uniqueness of, 351 Specialized epistemic authority, 351 Standardized criteria, 43–45 Standardized test scores, 43 Statistical experts, epistemic authority of, 385–386 Statistical techniques, 64 for assessing accuracy, 39 for distinguishing accuracy from self-fulfilling prophecy, 20–21 Status recognition, 189 Stereotype accuracy, 15, 16, 19, 50–52, 57, 63, 67, 77

407

INDEX Judd and Park’s full accuracy design for research on, 60–61 and levels of analysis, 23–29 Stereotype change, 324 cognitive models of, 268 Stereotype-disconfirming information, experimental studies of, 267–270 Stereotypes active preservation via subtyping, 269 effect of emotions on, 246 as hierarchical structures, 268 in intergroup conflict, 255 of outgroups in groupthink situations, 226 as perceptions of populations, 24 and person perception, 24–26 weak effect of contact on, 324 Subgroup salience, 265, 328 Subtyping model of stereotype change, 268 Symbolic predispositions, 136 Symbolic racism among college-educated whites, 116–118 among White Southerners, 143 antecedents of, 121 belief system of, 106–109 conceptual inconsistencies of, 106–107 consistency across diverse racial policies, 116 across themes of, 109–113 of political effects of, 109–118, 134 controversies in literature of, 95–97,101–106 distinctiveness from alternative constructs, 123–128 as dominant predictor of racially relevant political attitudes, 104 as expression of White group interests, 141 factor analyses of, 120 inconsistent measurement, 107–108 as independent construct, 128 internal consistency of belief system, 108–109 life span approach to origins of, 135–137 origins of, 118–123 overlap with policy attitudes, 113–116 as predictor of White opposition, 98 as predictor of White racial policy preferences, 140 and social dominance, 140 theory of, 95, 98–101 vs. old-fashioned racism, 123–124 vs. other constructs, 126–128

vs. political ideology, 124–126 Symbolic Racism 2000 Scale, 108, 110–111 Symbolic threats, 292 System justification, 382 Systematic mode, 349 T Targets agreement with, 39, 52–55 limitations to agreement with, 54–55 Teachers, children’s attribution of epistemic authority to, 357–359, 360 Thibaut, John, 158 Time pressure, impact on group decision quality, 224 Trust, 196 and attachment theory, 353–354 enhanced by group identification, 179 and goal-transformation, 174–180 informational, 346 legitimizing effect on authority, 183–184 links to sense of moral commitment, 176 and procedural fairness, 183–185 role in facilitating cooperation, 176 in self-based model of cooperation, 196 Trustworthiness, as communicator incentive, 347 Truth, with a small ‘‘t,’’ 37, 55 Tyler, Tom R., 151 Typicality, 319–320 association with stereotyping, 269 and degree of inference from member to group, 267 in Hewstone-Brown intergroup contact model, 266–267 limits to importance of, 273–275 moderating role of, 268–269, 270 and outgroup homogeneity, 271 of outgroup members, 270 in stereotype attitude change, 255 U Ubiquity model, of groupthink, 227–238, 238–239 Uncertainty management, 156, 170–173 Undeserved advantage, in symbolic racism, 103, 109

408

INDEX

Unimodel of source effects, 350–351 concept of epistemic authority in, 351–357 Unobservable personal characteristics, 73 V Valence inaccuracies, 61 Validity, 5 of beliefs about populations, 24 of criteria for social perception, 10 of stereotypes, 31 Values, children’s attribution of epistemic authority in domain of, 359–361 Variability differences and intergroup contact, 269 between judgments and criteria, 68

Victims, blaming, 31 Voice role in procedural fairness, 160, 167, 172 value-expressive worth of, 189 Voting patterns, intergenerational continuity in, 384–385 W Walker, Laurens, 158–159 White group consciousness, 133 vs. outgroup antagonism, 128–132 White self-interest, 129 Who knows best study, 357–359 Work ethic, in symbolic racism, 102, 107, 109, 114

CONTENTS OF OTHER VOLUMES

Volume 1 Cultural Influences upon Cognitive Processes Harry C. Triendis The Interaction of Cognitive and Physiological Determinants of Emotional State Stanley Schachter Experimental Studies of Coalirion Formation William A. Gamson Communication Networks Marvin E. Shaw A Contingency Model of Leadership Effectiveness Fred E. Fiedler Inducing Resistance to Persuasion: Some Contemporary Approaches William J. McGuire Social Motivation, Dependency, and Susceptibility to Social Influence Richard H. Walters and Ross D. Purke Sociability and Social Organization in Monkeys and Apes William A. Mason Author Index—Subject Index Volume 2 Vicarious Processes: A Case of No-Trial Learning Albert Bandura Selective Exposure Jonathan L. Freedman and David O. Sears Group Problem Solving L. Richard Hoffman Situational Factors in Conformity Vernon L. Allen Social Power John Schopler

From Acts to Dispositions: The Attribution Process in Person Perception Edward E. Jones and Keith E. Davis Inequality in Social Exchange J. Stacy Adams The Concept of Aggressive Drive: Some Additional Considerations Leonard Berkowitz Author Index—Subject Index Volume 3 Mathematical Models in Social Psychology Robert P. Abelson The Experimental Analysis of Social Performance Michael Argyle and Adam Kendon A Structural Balance Approach to the Analysis of Communication Effects N. T. Feather Effects of Fear Arousal on Attitude Change: Recent Developments in Theory and Experimental Research Irving L. Janis Communication Processes and the Properties of Language Serge Moscovici The Congruity Principle Revisited: Studies in the Reduction, Induction, and Generalization of Persuasion Percy H. Tannenbaum Author Index—Subject Index Volume 4 The Theory of Cognitive Dissonance: A Current Perspective Elliot Aronson

409

410

CONTENTS OF OTHER VOLUMES

Attitudes and Attraction Donn Byrne Sociolinguistics Susan M. Ervin-Tripp Recognition of Emotion Nico H. Frijda Studies of Status Congruence Edward E. Sampson Exploratory Investigations of Empathy Ezra Stotland The Personal Reference Scale: An Approach to Social Judgment Harry S. Upshaw Author Index—Subject Index Volume 5 Media Violence and Aggressive Behavior: A Review of Experimental Research Richard E. Goranson Studies in Leader Legitimacy, Influence, and Innovation Edwin P. Hollander and James W. Julian Experimental Studies of Negro-White Relationships Irwin Katz Findings and Theory in the Study of Fear Communications Howard Leventhal Perceived Freedom Ivan D. Steiner Experimental Studies of Families Nancy E. Waxler and Elliot G. Mishler Why Do Groups Make Riskier Decisions than Individuals? Kenneth L. Dion, Robert S. Baron, and Norman Miller Author Index—Subject Index Volume 6 Self-Perception Theory Daryl J. Bem Social Norms, Feelings, and Other Factors Affecting Helping and Altruism Leonard Berkowitz The Power of Liking: Consequence of Interpersonal Attitudes Derived from a Liberalized View of Secondary Reinforcement Albert J. Lott and Bernice E. Lott

Social Influence, Conformity Bias, and the Study of Active Minorities Serge Moscovici and Claude Faucheux A Critical Analysis of Research Utilizing the Prisoner’s Dilemma Paradigm for the Study of Bargaining Charlan Nemeth Structural Representations of Implicit Personality Theory Seymour Rosenberg and Andrea Sedlak Author Index—Subject Index Volume 7 Cognitive Algebra: Integration Theory Applied to Social Attribution Norman A. Anderson On Conflicts and Bargaining Erika Apfelbaum Physical Attractiveness Ellen Bersheid and Elaine Walster Compliance, Justification, and Cognitive Change Harold B. Gerard, Edward S. Connolley, and Roland A. Wilhelmy Processes in Delay of Gratification Walter Mischel Helping a Distressed Person: Social, Personality, and Stimulus Determinants Ervin Staub Author Index—Subject Index Volume 8 Social Support for Nonconformity Vernon L. Allen Group Tasks, Group Interaction Process, and Group Performance Effectiveness: A Review and Proposed Integration J. Richard Hackman and Charles G. Morris The Human Subject in the Psychology Experiment: Fact and Artifact Arie W. Kruglanski Emotional Arousal in the Facilitation of Aggression through Communication Percy H. Tannenbaum and Dolf Zillman The Reluctance to Transmit Bad News Abraham Tesser and Sidney Rosen Objective Self-Awareness Robert A. Wicklund

CONTENTS OF OTHER VOLUMES Responses to Uncontrollable Outcomes: An Integration of Reactance Theory and the Learned Helplessness Model Camille B. Wortman and Jack W. Brehm Subject Index Volume 9 New Directions in Equity Research Elaine Walster, Ellen Berscheid, and G. William Walster Equity Theory Revisited: Comments and Annotated Bibliography J. Stacy Adams and Sara Freedman The Distribution of Rewards and Resources in Groups and Organizations Gerald S. Leventhal Deserving and the Emergence of Forms of Justice Melvin J. Lerner, Dale T. Miller, and John G. Holmes Equity and the Law: The Effect of a Harmdoer’s ‘‘Suffering in the Act’’ on Liking and Assigned Punishment William Austin, Elaine Walster, and Mary Kristine Utne Incremental Exchange Theory: A Formal Model for Progression in Dyadic Social Interaction L. Lowell Huesmann and George Levinger Commentary George C. Homans Subject Index Volume 10 The Catharsis of Aggression: An Evaluation of a Hypothesis Russell G. Geen and Michael B. Quanty Mere Exposure Albert A. Harrison Moral Internalization: Current Theory and Research Martin L. Hoffman Some Effects of Violent and Nonviolent Movies on the Behavior of Juvenile Delinquents Ross D. Parke, Leonard Berkowitz, Jacques P. Leyens, Stephen G. West, and Richard Sebastian

411

The Intuitive Psychologist and His Shortcomings: Distortions in the Attribution Process Less Ross Normative Influences on Altruism Shalom H. Schwartz A Discussion of the Domain and Methods of Social Psychology: Two Papers by Ron Harre and Barry R. Schlenker Leonard Berkowitz The Ethogenic Approach: Theory and Practice R. Harre On the Ethogenic Approach: Etiquette and Revolution Barry R. Schlenker Automatisms and Autonomies: In Reply to Professor Schlenker R. Harre Subject Index Volume 11 The Persistence of Experimentally Induced Attitude Change Thomas D. Cook and Brian F. Flay The Contingency Model and the Dynamics of the Leadership Process Fred E. Fiedler An Attributional Theory of Choice Andy Kukla Group-Induced Polarization of Attitudes and Behavior Helmut Lamm and David G. Myers Crowding: Determinants and Effects Janet E. Stockdale Salience: Attention, and Attribution: Top of the Head Phenomena Shelley E. Taylor and Susan T. Fiske Self-Generated Attitude Change Abraham Tesser Subject Index Volume 12 Part I. Studies in Social Cognition Prototypes in Person Perception Nancy Cantor and Walter Mischel A Cognitive-Attributional Analysis of Stereotyping David L. Hamilton

412

CONTENTS OF OTHER VOLUMES

Self-Monitoring Processes Mark Snyder Part II. Social Influences and Social Interaction Architectural Mediation of Residential Density and Control: Crowding and the Regulation of Social Contact Andrew Baum and Stuart Valins A Cultural Ecology of Social Behavior J. W. Berry Experiments on Deviance with Special Reference to Dishonesty David P. Farrington From the Early Window to the Late Night Show: International Trends in the Study of Television’s Impact on Children and Adults John P. Murray and Susan Kippax Effects of Prosocial Television and Film Material on the Behavior of Viewers J. Phillipe Rushton Subject Index

Cognitive, Social, and Personality Processes in the Physiological Detection of Deception William M. Waid and Martin T. Orne Dialectic Conceptions in Social Psychology: An Application to Social Penetration and Privacy Regulation Irwin Altman, Anne Vinsel, and Barbara B. Brown Direct Experience and Attitude–Behavior Consistency Russell H. Fazio and Mark P. Zanna Predictability and Human Stress: Toward a Clarification of Evidence and Theory Suzanne M. Miller Perceptual and Judgmental Processes in Social Contexts Arnold Upmeyer Jury Trials: Psychology and Law Charlan Jeanne Nemeth Index Volume 15

Volume 13 People’s Analyses of the Causes of AbilityLinked Performances John M. Darley and George R. Goethals The Empirical Exploration of Intrinsic Motivational Processes Edward I. Deci and Richard M. Ryan Attribution of Responsibility: From Man the Scientist to Man as Lawyer Frank D. Fincham and Joseph M. Jaspars Toward a Comprehensive Theory of Emotion Howard Leventhal Toward a Theory of Conversion Behavior Serge Moscovici The Role of Information Retrieval and Conditional Inference Processes in Belief Formation and Change Robert S. Wyer, Jr. and Jon Hartwick Index

Balance, Agreement, and Positivity in the Cognition of Small Social Structures Walter H. Crockett Episode Cognition: Internal Representations of Interaction Routines Joseph P. Forgas The Effects of Aggressive-Pornographic Mass Media Stimuli Neil M. Malamuth and Ed Donnerstein Socialization in Small Groups: Temporal Changes in Individual–Group Relations Richard L. Moreland and John M. Levine Translating Actions into Attitudes: An Identity-Analytic Approach to the Explanation of Social Conduct Barry R. Schlenker Aversive Conditions as Stimuli to Aggression Leonard Berkowitz Index

Volume 14

Volume 16

Verbal and Nonverbal Communication of Deception Miron Zuckerman, Bella M. DePaulo, and Robert Rosenthal

A Contextualist Theory of Knowledge: Its Implications for Innovation and Reform in Psychological Research William J. McGuire

CONTENTS OF OTHER VOLUMES Social Cognition: Some Historical and Theoretical Perspectives Janet Landman and Melvin Manis Paradigmatic Behaviorism: Unified Theory for Social-Personality Psychology Arthur W. Staats Social PsychologyfromtheStandpointofa StructuralSymbolicInteractionism: Toward anInterdisciplinarySocialPsychology Sheldon Stryker Toward an Interdisciplinary Social Psychology Carl W. Backman Index Volume 17 Mental Representations of Self John F. Kihlstrom and Nancy Cantor Theory of the Self: Impasse and Evolution Kenneth J. Gergen A Perceptual-Motor Theory of Emotion Howard Leventhal Equity and Social Change in Human Relationships Charles G. McClintock, Roderick M. Kramer, and Linda J. Keil A New Look at Dissonance Theory Joel Cooper and Russell H. Fazio Cognitive Theories of Persuasion Alice H. Eagly and Shelly Chaiken Helping Behavior and Altruism: An Empirical and Conceptual Overview John F. Dovidio Index Volume 18 A Typological Approach to Marital Interaction: Recent Theory and Research Mary Anne Fitzpatrick Groups in Exotic Environments Albert A. Harrison and Mary M. Connors Balance Theory, the Jordan Paradigm, and the Wiest Tetrahedon Chester A. Insko The Social Relations Model David A. Kenny and Lawrence La Voie Coalition Bargaining S. S. Komorita

413

When Belief Creates Reality Mark Snyder Index Volume 19 Distraction–Conflict Theory: Progress and Problems Robert S. Baron Recent Research on Selective Exposure to Information Dieter Frey The Role of Threat to Self-Esteem and Perceived Control in Recipient Reaction to Help: Theory Development and Empirical Validation Arie Nadler and Jeffrey D. Fisher The Elaboration Likelihood Model of Persuasion Richard E. Petty and John T. Cacioppo Natural Experiments on the Effects of Mass Media Violence on Fatal Aggression: Strengths and Weaknesses of a New Approach David P. Phillips Paradigms and Groups Ivan D. Steiner Social Categorization: Implications for Creation and Reduction of Intergroup Bias David A. Wilder Index Volume 20 Attitudes, Traits, and Actions: Dispositional Prediction of Behavior in Personality and Social Psychology Icek Ajzen Prosocial Motivation: Is It Ever Truly Altruistic? C. Daniel Batson Dimensions of Group Process: Amount and Structure of Vocal Interaction James M. Dabbs, Jr. and R. Barry Ruback The Dynamics of Opinion Formation Harold B. Gerard and Ruben Orive Positive Affect, Cognitive Processes, and Social Behavior Alice M. Isen

414

CONTENTS OF OTHER VOLUMES

Between Hope and Fear: The Psychology of Risk Lola L. Lopes Toward an Integration of Cognitive and Motivational Perspectives on Social Inference: A Biased Hypothesis-Testing Model Tom Pyszczynski and Jeff Greenberg Index Volume 21 Introduction Leonard Berkowitz Part I. The Self as Known Narrative and the Self as Relationship Kenneth J. Gergen and Mary M. Gergen Self and Others: Studies in Social Personality and Autobiography Seymour Rosenberg Content and Process in the Experience of Self William J. McGuire and Claire V. McGuire Information Processing and the Study of the Self John F. Kihlstrom, Nancy Cantor, Jeanne Sumi Albright, Beverly R. Chew, Stanley B. Klein, and Paula M. Niedenthal Part II. Self-Motives Toward a Self-Evaluation Maintenance Model of Social Behavior Abraham Tesser The Self: A Dialectical Approach Carl W. Backman The Psychology of Self-Affirmation: Sustaining the Integrity of the Self Claude M. Steele A Model of Behavioral Self-Regulation: Translating Intention into Action Michael F. Scheier and Charles S. Carver Index Volume 22 On the Construction of the Anger Experience: Aversive Events and Negative Priming in the Formation of Feelings Leonard Berkowitz and Karen Heimer Social Psychophysiology: A New Look John T. Cacioppo, Richard E. Petty, and Louis G. Tassinary

Self-Discrepancy Theory: What Patterns of Self-Beliefs Cause People to Suffer? E. Tory Higgins Minding Matters: The Consequences of Mindlessness-Mindfulness Ellen J. Langer The Tradeoffs of Social Control and Innovation in Groups and Organizations Charlan Jeanne Nemeth and Barry M. Staw Confession, Inhibition, and Disease James W. Pennebaker A Sociocognitive Model of Attitude Structure and Function Anthony R. Pratkanis and Anthony G. Greenwald Introspection, Attitude Change, and Attitude– Behavior Consistency: The Disruptive Effects of Explaining Why We Feel the Way We Do Timothy D. Wilson, Dana S. Dunn, Dolores Kraft, and Douglas J. Lisle Index Volume 23 A Continuum of Impression Formation, from Category-Based to Individuating Processes: Influences of Information and Motivation on Attention and Interpretation Susan T. Fiske and Steven L. Neuberg Multiple Processes by Which Attitudes Guide Behavior: The MODE Model as an Integrative Framework Russell H. Fazio PEAT: An Integrative Model of Attribution Processes John W. Medcof Reading People’s Minds: A Transformation Rule Model for Predicting Others’ Thoughts and Feelings Rachel Karniol Self-Attention and Behavior: A Review and Theoretical Update Frederick X. Gibbons Counterfactual Thinking and Social Perception: Thinking about What Might Have Been Dale T. Miller, William Turnbull, and Cathy McFarland Index

CONTENTS OF OTHER VOLUMES Volume 24 The Role of Self-Interest in Social and Political Attitudes David O. Sears and Carolyn L. Funk A Terror Management Theory of Social Behavior: The Psychological Functions of Self-Esteem and Cultural Worldviews Sheldon Solomon, Jeff Greenberg, and Tom Pyszczynski Mood and Persuasion: Affective States Influence the Processing of Persuasive Communications Norbert Schwarz, Herbert Bless, and Gerd Bohner A Focus Theory of Normative Conduct: A Theoretical Refinement and Reevaluation of the Role of Norms in Human Behavior Robert B. Cialdini, Carl A. Kallgren, and Raymond R. Reno The Effects of Interaction Goals on Person Perception James L. Hilton and John M. Darley Studying Social Interaction with the Rochester Interaction Record Harry T. Reis and Ladd Wheeler Subjective Construal, Social Inference, and Human Misunderstanding Dale W. Griffin and Lee Ross Index Volume 25 Universals in the Content and Structure of Values: Theoretical Advances and Empirical Tests in 20 Countries Shalom H. Schwartz Motivational Foundations of Behavioral Confirmation Mark Snyder A Relational Model of Authority in Groups Tom R. Tyler and E. Allan Lind You Can’t Always Think What You Want: Problems in the Suppression of Unwanted Thoughts Daniel M. Wegner Affect in Social Judgments and Decisions: A Multiprocess Model Joseph Paul Forgas

415

The Social Psychology of Stanley Milgram Thomas Blass The Impact of Accountability on Judgment and Choice: Toward a Social Contingency Model Philip E. Tetlock Index Volume 26 Attitudes toward High Achievers and Reactions to Their Fall: Theory and Research Concerning Tall Poppies N. T. Feather Evolutionary Social Psychology: From Sexual Selection to Social Cognition Douglas T. Kenrick Judgment in a Social Context: Biases, Short-comings, and the Logic of Conversation Norbert Schwarz A Phase Model of Transitions: Cognitive and Motivational Consequences Diane N. Ruble Multiple-Audience Problems, Tactical Communication, and Social Interaction: A Relational-Regulation Perspective John H. Fleming From Social Inequality to Personal Entitlement: The Role of Social Comparisons, Legitimacy Appraisals, and Group Membership Brenda Major Mental Representations of Social Groups: Advances in Understanding Stereotypes and Stereotyping Charles Stangor and James E. Lange Index Volume 27 Inferences of Responsibility and Social Motivation Bernard Weiner Information Processing in Social Contexts: Implications for Social Memory and Judgment Robert S. Wyer, Jr. and Deborah H. Gruenfeld

416

CONTENTS OF OTHER VOLUMES

The Interactive Roles of Stability and Level of Self-Esteem: Research and Theory Michael H. Kernis and Stephanie B. Waschull Gender Differences in Perceiving Internal State: Toward a His-and-Hers Model of Perceptual Cue Use Tomi-Ann Roberts and James W. Pennebaker On the Role of Encoding Processes in Stereotype Maintenance William von Hippel, Denise Sekaquaptewa, and Patrick Vargas Psychological Barriers to Dispute Resolution Lee Ross and Andrew Ward Index Volume 28 The Biopsychosocial Model of Arousal Regulation Jim Blascovich and Joe Tomaka Outcome Biases in Social Perception: Implications for Dispositional Inference, Attitude Change, Stereotyping, and Social Behavior Scott T. Allison, Diane M. Mackie, and David M. Messick Principles of Judging Valence: What Makes Events Positive or Negative? C. Miguel Brendl and E. Tory Higgins Pluralistic Ignorance and the Perpetuation of Social Norms by Unwitting Actors Deborah A. Prentice and Dale T. Miller People as Flexible Interpreters: Evidence and Issues from Spontaneous Trait Inference James S. Uleman, Leonard S. Newman, and Gordon B. Moskowitz Social Perception, Social Stereotypes, and Teacher Expectations: Accuracy and the Quest for the Powerful Self-Fulfilling Prophecy Lee Jussim Jacquelynne Eccles, and Stephanie Madon Nonverbal Behavior and Nonverbal Communication: What do Conversational Hand Gestures Tell Us? Robert M. Krauss, Yihsiu Chen, and Purnima Chawla Index

Volume 29 Counterfactual Thinking: The Intersection of Affect and Function Neal J. Roese and James M. Olson Terror Management Theory of Self-Esteem and Cultural Worldviews: Empirical Assessments and Conceptual Refinements Jeff Greenberg, Sheldon Solomon, and Tom Pyszczynski The Flexible Correction Model: The Role of Naı¨ve Theories of Bias in Bias Correction Duane T. Wegener and Richard E. Petty Self-Evaluation: To Thine Own Self Be Good, to Thine Own Self Be Sure, to Thine Own Self Be True, and to Thine Own Self Be Better Constantine Sedikides and Michael J. Strube Toward a Hierarchical Model of Intrinsic and Extrinsic Motivation Robert J. Vallerand Volume 30 Promotion and Prevention: Regulatory Focus as a Motivational Principle E. Tory Higgins The Other ‘‘Authoritarian Personality’’ Bob Altemeyer Person Preception Comes of Age: The Salience and Significance of Age in Social Adjustments Joann M. Montepare and Leslie A. Zebrowitz On the Perception of Social Consensus Joachim Krueger Prejudice and Stereotyping in Everyday Communication Janet B. Ruscher Situated Optimism: Specific Outcome Expectancies and Self-Regulation David A. Armor and Shelley E. Taylor Volume 31 Affect and Information Processing Robert S. Wyer, Jr., Gerald L. Clore, and Linda M. Isbell Linguistic Intergroup Bias: Stereotype Perpetuation through Language Anne Maass

CONTENTS OF OTHER VOLUMES Relationships from the Past in the Present: Significant-Other Representations and Transference in Interpersonal Life Serena Chen and Susan M. Anderson The Puzzle of Continuing Group Inequality: Piecing Together Psychological, Social, and Cultural Forces in Social Dominance Theory Felicia Prano Attitude Representation Theory Charles G. Lord and Mark R. Lepper Discontinuity Theory: Cognitive and Social Searches for Rationality and Normality— May Lead to Madness Philip G. Zimbardo

417

Ambivalent Sexism Peter Glick and Susan T. Fiske Videotaped Confessions: Is Guilt in the Eye of the Camera? G. Daniel Lassiter, Andrew L. Geers, Patrick J. Munhall, Ian M. Handley, and Melissa J. Beers Effort Determination of Cardiovascular Response: An Integrative Analysis with Applications in Social Psychology Rex A. Wright and Leslie D. Kirby Index Volume 34

Volume 32 The Nature and Function of Self-Esteem: Sociometer Theory Mark R. Leary and Roy F. Baumeister Temperature and Aggression Craig A. Anderson, Kathryn B. Anderson, Nancy Dorr, Kristina M. DeNeve, and Mindy Flanagan The Importance of Being Selective: Weighing the Role of Attribute Importance in Attitudinal Judgment Joop van der Pligt, Nanne K. de Vries, Antony S. R. Manstead, and Frank van Harreveld Toward a History of Social Behavior: Judgmental Accuracy from Thin Slices of the Behavioral Stream Nalini Amabady, Frank J. Bernieri, and Jennifer A. Richeson Attractiveness, Attraction, and Sexual Selection: Evolutionary Perspectives on the Form and Function of Physical Attractiveness Dianne S. Berry Index Volume 33 The Perception–Behavior Expressway: Automatic Effects of Social Perception on Social Behavior Ap Dijksterhuis and John A. Bargh A Dual-Process Cognitive-Motivational Theory of Ideology and Prejudice John Duckitt

Uncertainty Management by Means of Fairness Judgments Kees van den Bos and E. Allan Lind Cognition in Persuasion: An Analysis of Information Processing in Response to Persuasive Communications Dolores Albarracin Narrative-Based Representations of Social Knowledge: Their Construction and Use in Comprehension, Memory, and Judgment Robert S. Wyer, Jr., Rashmi Adaval and Stanley J. Colcombe Reflexion and Reflection: A Social Cognitive Neuroscience Approach to Attributional Inference Matthew D. Lieberman, Ruth Gaunt, Daniel T. Gilbert, and Yaacov Trope Antecedents and Consequences of Attributions to Discrimination: Theoretical and Empirical Advances Brenda Major, Wendy J. Quinton, and Shannon K. McCoy A Theory of Goal Systems Arie W. Kruglanski, James Y. Shah, Ayelet Fishbach, Ron Friedman, Woo Young Chun, and David Sleeth-Keppler Contending with Group Image: The Psychology of Stereotype and Social Identity Threat Claude M. Steele, Steven J. Spencer, and Joshua Aronson Index

418

CONTENTS OF OTHER VOLUMES

Volume 35

Volume 36

Social Identity and Leadership Processes in Groups Michael A. Hogg and Daan van Knippenberg The Attachment Behavioral System in Adulthood: Activation, Psychodynamics, and Interpersonal Processes Mario Mikulincer and Phillip R. Shaver Stereotypes and Behavioral Confirmation: From Interpersonal to Intergroup Perspectives Olivier Klein and Mark Snyder Motivational Bases of Information Processing and Strategy in Conflict and Negotiation Carsten K. W. De Dreu and Peter J. Carnevale Regulatory Mode: Locomotion and Assessment as Distinct Orientations E. Tory Higgins, Arie W. Kruglanski, and Antonio Pierro Affective Forecasting Timothy D. Wilson and Daniel T. Gilbert Index

Aversive Racism John F. Dovidio and Samuel L. Gaertner Socially Situated Cognition: Cognition in its Social Context Eliot R. Smith and Gu¨n R. Semin Social Axioms: A Model for Social Beliefs in Multicultural Perspective Kwok Leung and Michael Harris Bond Violent Video Games: Specific Effects of Violent Content on Aggressive Thoughts and Behavior Craig A. Anderson, Nicholas L. Carnagey, Mindy Flanagan, Arlin J. Benjamin, Jr., Janie Eubanks, and Jeffery C. Valentine Survival and Change in Judgments: A Model of Activation and Comparison Dolores Albarracı´n, Harry M. Wallace, and Laura R. Glasman The Implicit Volition Model: On the Preconscious Regulation of Temporarily Adopted Goals Gordon B. Moskowitz, Peizhong Li, and Elizabeth R. Kirk Index