ADVANCES IN
Experimental Social Psychology
VOLUME 33
This Page Intentionally Left Blank
ADVANCES IN
Experimental...
14 downloads
1136 Views
2MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
ADVANCES IN
Experimental Social Psychology
VOLUME 33
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 33
San Diego San Francisco London Sydney Tokyo
New York
Boston
∞ This book is printed on acid-free paper.
C 2001 by ACADEMIC PRESS Copyright
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. (222 Rosewood Drive, Danvers, Massachusetts 01923), 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-2001 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. 0065-2601/01 $35.00 Explicit permission from Academic Press is not required to reproduce a maximum of two figures or tables from an Academic Press chapter in another scientific or research publication provided that the material has not been credited to another source and that full credit to the Academic Press chapter is given.
Academic Press A Harcourt Science and Technology Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.academicpress.com
Academic Press Harcourt Place, 32 Jamestown Road, London NW1 7BY, UK http://www.academicpress.com International Standard Book Number: 0-12-015233-9 PRINTED IN THE UNITED STATES OF AMERICA 01 02 03 04 05 06 SB 9 8 7 6 5 4
3
2
1
CONTENTS
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
The Perception–Behavior Expressway: Automatic Effects of Social Perception on Social Behavior Ap Dijksterhuis and John A. Bargh I. II. III. IV. V. VI. VII. VIII. IX.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Relation between Perception and Behavior . . . . . . . . . . . . . . . . . . . . . . . . . The Direct Effect of Perception on Behavior Produces Imitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Three Musketeers of Social Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . Social Perception Elicits Corresponding Behavior . . . . . . . . . . . . . . . . . . . . . . . Mediators of the Perception–Behavior Link . . . . . . . . . . . . . . . . . . . . . . . . . . . . Behavioral Contrast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Moderation of the Perception–Behavior Link . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 3 6 8 10 22 24 28 32 33
A Dual-Process Cognitive-Motivational Theory of Ideology and Prejudice John Duckitt I. II. III. IV. V. VI. VII.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prejudice and Personality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Theoretical Model of Personality, Ideology, and Prejudice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Testing the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dual Dimensions of Ideology and Prejudice . . . . . . . . . . . . . . . . . . . . . . . . . . . Prejudice as an Individual and a Group Phenomenon . . . . . . . . . . . . . . . . . . . . Integrating Individual and Intergroup Dynamics of Prejudice . . . . . . . . . . . . . .
v
41 42 50 60 85 90 96
vi
CONTENTS
VIII.
Overview and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
105 106
Ambivalent Sexism Peter Glick and Susan T. Fiske I. II. III. IV. V. VI. VII. VIII. IX. X. XI.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attitudes toward Women: Ambivalent Sexism Theory . . . . . . . . . . . . . . . . . . . Measuring Ambivalent Sexism: The ASI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Initial Validation of the ASI in the United States . . . . . . . . . . . . . . . . . . . . . . . . Cross-Cultural Validation of the ASI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consequences of HS and BS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Where Is the Ambivalence? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attitudes toward Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Development of Gender Prejudice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Implications for Prejudice: A Theory of Stereotype Content . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
115 117 123 127 133 140 149 152 165 169 179 183
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 I. II. III. IV. V. VI. VII. VIII. IX. X.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Confession Evidence—Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Presentation Format of Confession Evidence: The Growing Emphasis on Videotape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of the Literature on Point-of-View/Salience Effects in Causal Attribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Implications of the Salience Effects Literature for the Use of Videotaped Confessions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Program of Research Investigating the Potential for Bias in Videotaped Confessions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stage One: Establishing the Existence and Robustness of the Camera Perspective Bias in Videotaped Confessions . . . . . . . . . . . . . . . . . . . . . . . . . . . Stage Two: Examining the Camera Perspective Bias in Videotaped Confessions in the Context of Elaborate Trial Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . Stage Three: Exploring Possible Mediators of the Biasing Effect of Camera Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical and Practical Implications of the Present Program of Research . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
189 191 194 195 197 198 199 218 231 241 247
CONTENTS
vii
Effort Determination of Cardiovascular Response: An Integrative Analysis with Applications in Social Psychology Rex A. Wright and Leslie D. Kirby I.
Psychological Determinants of Cardiovascular Response: Three Conceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Integrative Analytic Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contrast with the Alternative Views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evidence for the Primary Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evidence for Implications Deriving from the Ability Extension . . . . . . . . . . . . Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
256 258 265 266 279 290 299 299
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contents of Other Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
309 317
II. III. IV. V. VI. VII.
This Page Intentionally Left Blank
CONTRIBUTORS
Numbers in parentheses indicate the pages on which the authors’ contributions begin.
JOHN A. BARGH (1), Department of Psychology, New York University, New York, New York 10003 MELISSA J. BEERS (189), Paul Werth Associates, Columbus, Ohio 43215 AP DIJKSTERHUIS (1), Department of Social Psychology, University of Nijmegan, 6500 HE, Nijmegan, The Netherlands JOHN DUCKITT (41), Psychology Department, University of Auckland, Private Bag 92019, Auckland, New Zealand SUSAN T. FISKE (115), Department of Psychology, Princeton University, Princeton, New Jersey 08544 ANDREW L. GEERS (189), Department of Psychology, Ohio University, Athens, Ohio 45701 PETER GLICK (115), Psychology Department, Lawrence University, Appleton, Wisconsin 54912 IAN M. HANDLEY (189), Department of Psychology, Ohio University, Athens, Ohio 45701 LESLIE D. KIRBY (255), Department of Psychology, University of Alabama at Birmingham, Birmingham, Alabama 35294 G. DANIEL LASSITER (189), Department of Psychology, Ohio University, Athens, Ohio 45701 PATRICK J. MUNHALL (189), Department of Psychology, Ohio University, Athens, Ohio 45701 REX A. WRIGHT (255), Department of Psychology, University of Alabama at Birmingham, Birmingham, Alabama 35294
ix
This Page Intentionally Left Blank
THE PERCEPTION–BEHAVIOR EXPRESSWAY: AUTOMATIC EFFECTS OF SOCIAL PERCEPTION ON SOCIAL BEHAVIOR
Ap Dijksterhuis John A. Bargh
Each of us is in fact what he is almost exclusively by virtue of his imitativeness —William James (1890, p. 741)
I. Introduction Some years ago, one of the authors was driving on a highway at a speed of about 85 miles an hour. Not surprisingly, after a while a police car turned up and the driver was summoned to stop. One of the officers approached the car and asked, “What do you think you’re doing? Have you just been watching the Formula One Grand Prix on TV?” The driver pondered this question for a while and said, somewhat hesitantly, “Well, yes, as a matter of fact I was.” The officer, presumably a Formula One fan himself, nodded, smiled sympathetically, and gave the driver a steep fine. As William James noted in our opening quote, we have an innate tendency to imitate. We whisper to someone who is whispering; we start to speak much louder when others do so. We scratch our head upon seeing someone else scratch his or her head. We walk slower in the presence of the elderly, we cycle faster after we have seen a cycling race on TV, and, yes indeed, we get a fine for driving too fast after we have been watching a Formula One Grand Prix. In this chapter, we argue that this tendency to imitate is the consequence of the way we—or, rather, our brains—are shaped. We argue that social perception, defined here as the activation of a perceptual representation, has a direct effect on social behavior. Perceptual inputs are translated automatically into corresponding behavioral outputs. As a result, we often do what we see others doing. ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 33
1
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved. 0065-2601/01 $35.00
2
AP DIJKSTERHUIS AND JOHN A. BARGH
We must at the outset distinguish the present notion of a direct effect of perception on behavior from two major historical positions that are superficially similar. The first, the behaviorists’ thesis that responses follow directly from perceived stimuli, or S-R bonds, also holds that perception directly leads to action (e.g., Skinner, 1938; Watson, 1913). However, the mechanism proposed is quite different because these responses are not imitations of the perceived event but are stamped-in responses to stimuli based on one’s past reinforcement history. The second apparently similar theoretical position is Gibson’s (e.g., 1979; McArthur & Baron, 1983) notion of affordances. In this view, environmental stimuli directly suggest the appropriate behavioral response to them: the grilled lobster says “eat me” and the cold glass of beer says “drink me.” Both the behaviorist and the Gibsonian theorist would argue, as we do here, that behavioral tendencies are put into motion directly by perceptual activity, but unlike the present theme, they also argue (more or less) that these tendencies are learned responses over time based on one’s history of reward and punishment with those stimuli. The perception–behavior link argued for here, on the other hand, is the human (and basic animal) tendency to act in the same way as we see others act. We contend that this phenomenon flows directly from a fact of mental representation and organization—that perceptual and behavioral representations for the same action overlap. Thus the effect is a natural consequence of the automatic activation of the behavioral response by the perception of someone else doing the same thing. It is not necessary that the behavioral response be stamped in as a habit through reinforcement and it is not necessary for the response to be intended and strategic. The chapter is structured in the following manner: First, we start with a broad discussion on the functional relation between perception and behavior. This section is not restricted to imitation; instead, we present a more general perspective on perception and action. Second, we elaborate in more depth on the direct relation between perception and behavior and specifically on one consequence of this relation: imitation. Third, we define the core concepts of social perception. A distinction is drawn between observable behavior (such as gestures), inferences we make on the basis of the observed behaviors of others, and representations that become activated because of the social group membership of others. In the fourth section, evidence is discussed that indicates that all three of these forms of social perception lead directly to corresponding overt behavioral tendencies. We next review evidence concerning the mediators of the perception–behavior relation, as well as evidence regarding various moderators that are relevant for understanding the circumstances under which people do versus do not imitate. We close the chapter by discussing the perception–behavior link from a functional perspective.
THE PERCEPTION–BEHAVIOR LINK
3
II. The Relation between Perception and Behavior The cognitive approach that has dominated psychology for over 30 years has changed psychology’s perspective on perception. When asked what the most important function of perception is, most—if not all—people would presumably answer that perception provides us with an understanding of the world. We perceive because we want to know what is going on around us. Although this answer is compelling, it is also largely incomplete and to some extent plainly wrong. Certainly, perception is essential for us to comprehend our environment but that does not mean that this understanding is an end in itself. Rather, understanding is a means by which we act effectively. Adaptive perception is ultimately in the service of functional behavioral responding to the environment, and comprehension and understanding are only important means to that end. Another way to look at this is by taking an evolutionary perspective. In the course of our development as a species, perceptual abilities and functions developed because we started to behave, not because we started to understand. Humans and squirrels are able to perceive and to behave, while oak trees and stinging nettles are not able to perceive and not able to behave. Plants that are fixed in position and do not motorically navigate their environment did not develop mechanisms of perception, whereas animals that are able to move around in their world did. As Milner and Goodale (1995) noted: “Natural selection operates on the level of overt behavior; it cares little about how well an animal ‘sees’ the world, but a great deal about how well the animals forages for food, avoid predators, finds mates, and moves efficiently from one place in the environment to another” (p. 11). We are able to see, in other words, simply because we descend from individuals who could see and who were better at mating or better at avoiding falling trees or hungry lions than other, nonseeing individuals. In sum, perception is for doing. It is our best action guidance and control device. Especially in nonprimate animals there is often a one-to-one relation between a specific perceptual process and a specific form of action. Frogs, for instance, have two different perceptual systems. One system is responsible for detecting and hunting small prey objects, whereas the other is responsible for avoiding large objects. These systems function independent of each other and, importantly, were developed independent of each other. Evidence shows that destroying the system responsible for detecting prey objects has no detrimental effects whatsoever on the capacity to avoid larger objects and vice versa (Ingle, 1973). Another case in point is the small sea creature belonging to the order Balanomorpha. This creature leads a curious life. For a while, it does nothing but float with the currents. When it eventually reaches a solid surface, it performs the only action of its life—it attaches itself to this surface. Once the perceptual system has performed its function (detecting a solid surface), it ceases to function and dies. Action is not called for
4
AP DIJKSTERHUIS AND JOHN A. BARGH
anymore, so perception is thrown overboard (in fact, the creature devours its own brain at this point).
A. PERCEIVING LEADS TO DOING Among these more simple creatures, the same perceptual process always has the same behavioral consequence. For a frog, a large object above the surface means “flight,” while a small, irregularly moving object on the surface means “go for it.” There are no exceptions. The perception of a small object on the surface always prompts hunting behavior. Perception races right through the brain to evoke behavioral output. It does not stop somewhere, does not alter its course. Further animal evidence for the direct relation between perception and action comes from studies of the behavior of fish in shoals (e.g., Breder, 1976; Pitcher, 1979). Everyone has witnessed the impressive synchrony of movement that fish in shoals can display. They all move in the same direction, and then change direction, at the same time. This behavior is in harmony with the hypothesis of a direct link between perception and action. If a fish perceives the fellow fish in front of it change direction, it can do nothing but the same. Admittedly, or rather, fortunately, many species have a behavioral repertoire that is more flexible than that of fish, frogs, or Balanomorpha. Humans also prey (or at least recognize food), mate, and avoid large objects, but in humans a specific perceptual process does not always lead to the same specific act. Although certain stimuli possess strong affordances (McArthur & Baron, 1983), we are all able to look at a grilled buttery lobster without starting to eat it or look at a cold glass of beer without starting to drink it, though this may be harder for some than for others (e.g., Metcalfe & Mischel, 1999). And although perceiving a person behave in a certain way creates that same tendency in oneself, the doctor in the examining room does not undress along with his or her patient. The conclusion from this state of affairs is that whereas in some species perception always leads to action, in others (such as humans) it does not. As Buytendijk (1922) put it long ago, “In ourselves we notice that perceptual processes can occur independently of specific actions. However, with animals this is not the case. With animals, perception is always related to specific actions or, more precisely, perception always include the impetus to actions” (p. 24).
B. TWO POSSIBLE ROADS TO FLEXIBILITY How can we reconcile the fact that perception appears specifically designed for and directly leads to action tendencies with the fact that in humans (and only a limited number of other species) these action tendencies are not obligatory? In
THE PERCEPTION–BEHAVIOR LINK
5
other words, how can there be such rigid relations between perceptual processes and action tendencies and at the same time such flexibility? Although there are surely numerous possibilities, two more general “classes” of possibilities loom largest. The first possibility is that perception in itself is insufficient to elicit action and that an additional process is needed. In the absence of this additional facilitating mechanism, perception does not directly affect overt behavior. For example, it is possible that perception must be accompanied by a consciously made decision, some form of “express fiat” to be translated into overt behavior. Alternatively, it is possible that motivation functions as such a critical facilitator. A system shaped like this can certainly explain flexibility in behavior. Sometimes the facilitator is present, and sometimes it is absent; hence, sometimes perception leads to action, whereas on other occasions it does not. This possibility (which can be termed the facilitator option) is regarded by many as the most likely candidate. The second possibility is that perceptual activity is sufficient to create action but that it is sometimes inhibited. That is, the default option is that perception does lead to action (as in fish or frogs) but under some circumstances a “stop sign” is given in order to block the impulse from resulting in overt behavior (see Logan & Cowan, 1984). In concrete terms, we would see an aggressive act (for example) and the impulse or urge would be to act aggressively ourselves (see Berkowitz, 1984) but then control or inhibit this impulse from reaching behavioral fruition, for some reason. (We discuss what these reasons might be in a further section, but they mainly have to do with a conflict between the automatically suggested behavior and one’s current or chronic goals.) And if no inhibitors are present, we will indeed act on the perceptually instigated impulse. Of course, such a system can account for flexibility. Perception leads to action, but inhibitors (or acts of control) are able to block or prevent this from occurring. This possibility (which we call the “inhibitor option”) is not regarded by many as the most likely candidate, but as we see it, it is way ahead on points. First, the inhibitor option is the more likely candidate from an evolutionary perspective. When new species develop, this is done by adding new brain parts to existing old ones (see, e.g., Dennett, 1995; Milner & Goodale, 1995). Old modules do not suddenly cease to exist; it is rather that some new function is added. The frog and fish, in other words, are still in us. The advantage that humans have is that we also possess new inhibiting or moderating systems to the automatic perception–behavior effect. The inhibitor option is consistent with this principle of evolution. Direct perception–behavior links still exist, but can be moderated by newer systems that can exercise a certain degree of control over older ones. The facilitator option, on the other hand, is inconsistent with this truism. The assumption of some sort of facilitator itself is not problematic. It can simply be seen as a new system added somewhere along the way. However, requiring a facilitator (whether motivated or
6
AP DIJKSTERHUIS AND JOHN A. BARGH
not) would mean that the direct link between perception and behavior has somehow ceased to exist. The facilitator option, then, depends on the unlikely assumption that old modules are thrown away and fully replaced by new ones. In concrete terms, the frog or fish would have turned out to be useless, and the development of a new species would have to start from scratch. But this is not how evolution works. Other evidence in favor of the inhibitor option comes from studies of people with various disorders. Stronger than normal effects of perception on behavior can be observed in aphasia, apraxia, low-rate mental deficiency, epilepsy, and catatonic states (Prinz, 1990; Stengel, Vienna, & Edin, 1947), conditions in which the ability to control or inhibit thought and action is impaired. Frontal lobe damage is also associated with diminished inhibitory functioning (e.g., Passingham, 1993; Smith & Jonides, 1999) and, indeed, frontal lobe patients are characterized by relatively direct and uncontrolled effects of perception on behavior (Lhermitte, 1983). When they see water, they drink. When they see a grilled lobster, they eat, even when this is obviously inappropriate. In other words, removing the capacity for inhibition increases the effect of perception on behavior. These findings contradict the facilitator option because explaining them in terms of facilitation would require the absurd assumption that the effects are due to better facilitatory capacities among frontal lobe patients.
III. The Direct Effect of Perception on Behavior Produces Imitation As can be concluded from the above, there is an express connection between perceptual input and behavioral output. However, such a direct link between perception and action does not yet explain imitative behavior. That is, the assumption of such a link does not necessarily imply that perception leads to behavior that corresponds with perception, or that which resembles that what has just been perceived. The reason that this happens is that perception and action share neurological systems. This means that the translation of perception into corresponding action is a consequence of the way we are wired. In what follows, we review both neurological evidence and research on the “common-coding” hypothesis that support the view of shared neurological systems or shared mental representations.
A. NEUROPHYSIOLOGICAL EVIDENCE There is plenty of evidence for a direct relation between perception and behavior in animals other than fish or frogs. Various neurological studies with monkeys show that the same area of the premotor cortex becomes activated when the monkey
THE PERCEPTION–BEHAVIOR LINK
7
witnesses an action (the experimenter reaching for something) as when the monkey performs the same action (Di Pellegrino, Fadiga, Fogassi, Gallese, & Rizzolatti, 1992; Rizzolatti & Arbib, 1998). Thus in primates there is an overlap between the mental representations used in perceiving an action and those used to perform the same action. Thus, perception primes or activates the behavioral tendency itself. Such strong support for a direct relation between perception and action has also been obtained with human participants. Zajonc, Pietromonaco, and Bargh (1982) showed that participants instructed to try to remember each of a series of faces taken from a college yearbook spontaneously (and subtly) mimicked the facial expressions while they viewed each photograph; interfering with these slight muscle movements by having some participants chew gum while viewing the photographs interfered with later memory for the faces. Similarly, Fadiga et al. (1995) showed that watching an experimenter grasping an object leads to muscular responses that are (more or less) the same as the muscular responses participants displayed while grasping the object themselves. Long ago, Carpenter (1874) and James (1890) proposed the notion of ideomotor action—that merely thinking about doing something automatically makes it more likely that you will perform the action. James defined this principle as “every representation of a movement awakens in some degree the actual movement which is its object” (p. 396). James also emphasized the passive nature of the effect; he argued that an act of will was not necessary for the action impulse instigated by the thought to emerge in actual behavior. Recent neurophysiological evidence, as well as experimental evidence reviewed below, is in harmony with the principle of ideomotor action. Paus, Petrides, Evans, and Meyer (1993) found that thinking about a word or a gesture leads to the same activation in the anterior cingulate cortex as actually uttering the word or making the gesture. Jeannerod (e.g., 1994; see also 1997) showed that mentally simulating an action leads to activation of the same neurons in the premotor cortex as performing this action and concluded that “simulating a movement is the same thing as performing it, except that the execution is somehow blocked” (p. 1422). In their studies, Jeannerod and colleagues demonstrated that imagining complex actions (such as running, rowing, or weightlifting) has neurophysiological consequences that are largely comparable to those of actually engaging in those actions. In both cases, motor programs are active (Decety, Jeannerod, Germain, & Pastene, 1991; Jeannerod, 1994, 1997). The principle of ideomotor action, or “thinking is for doing” in James’s wellknown phrase, is consistent with the notion of a direct and unmediated effect of perception on behavior if it is assumed that perceptual activity is another source of behavior-relevant thought (see Berkowitz, 1984; Chartrand & Bargh, 1999). As we argue below, the key mediator of perception–behavior effects is the activation of the mental representation of the behavior, and this can occur through perceiving that behavior as well as thinking about it actively.
8
AP DIJKSTERHUIS AND JOHN A. BARGH
B. THE COMMON-CODING HYPOTHESIS There is other evidence for shared representational systems for perception and action. Prinz (1990) claimed that language comprehension and language production depend on the same representational systems. More generally, he proposed the idea of common-coding, that is, shared representational systems for perception and action. An interesting corrolary of this hypothesis is that performing an action at the same time as perceiving that action should be difficult if both activities require the same representation. M¨usseler and Hommel (1997) tested this idea. They presented participants with series of four left or right arrows (e.g., “<<><”). Participants were asked to read these series and to reproduce them by pressing on the corresponding arrow buttons on a computer keyboard. Later, participants were presented with a fifth arrow that was always presented exactly when the participant was pressing the key corresponding to the second arrow presented. Participants were asked to press the key corresponding to the fifth arrow immediately upon reponding to the first four. Of interest was the number of mistakes participants made with their responses to the fifth arrow. According to the logic of common coding, participants should make more mistakes in their reponses to the fifth arrow if this arrow was the same as the second (i.e., both were right arrows or both were left arrows), that is the one they were responding to while the fifth arrow was being presented. And this is indeed what happened. While pressing a certain arrow key, participants had more trouble perceiving this same arrow than the opposite arrow, as shown by greater error rates in reporting after the fact which arrow had been presented. The implications of the fact that activation of the mental representation of an action leads to actual engagement in this behavior is that people have a natural tendency to imitate (see also Greenwald, 1970; Wheeler, 1966). Perceiving an action activates the mental representation of this action, which in turn leads to performance of the action. In other words, our tendency to imitate others is a consequence of the way that behavior is represented mentally. It is not motivated (necessarily) or requiring of a choice to occur, but rather is a natural consequence of the way we are wired.
IV. The Three Musketeers of Social Perception The conclusion of the previous section is that we have a tendency to imitate others because perception automatically elicits corresponding behavior. If one wants to know what sort of behavior we tend to imitate, an easy way out would
THE PERCEPTION–BEHAVIOR LINK
9
be to say that—because perception leads to corresponding behavior—we imitate everything we can perceive. This is true, but then the need arises to first discuss what we can perceive. So what does a social perceiver perceive? First, social perceivers perceive what we may call observables. This class of behavior is easy to define. It involves behavior that we can literally perceive. We perceive gestures and movements of others. We can see someone wave, scratch her head, or wiggle his foot. Furthermore, we can perceive various facial expressions. We see people smile or frown, for instance. Also, we hear people speak. Not only do we listen to the contents of speech, we also perceive other variables such as accents or tone of voice. Second, we generate trait inferences on the basis of the behaviors of others. These inferences (e.g., honest, intelligent) are themselves not literally perceived, but are made upon the perception of behavior that is present and observed in the current environment. Such inferences are made spontaneously—that is, unintentionally and immediately—upon perception of the observable act (e.g., Gilbert, 1989; Winter & Uleman, 1984). If we learn that “Pam brought flowers when she picked up her boyfriend from the airport,” we spontaneously translate this concrete behavior into an abstract personality trait. Without being aware of it, we draw the conclusion that Pam is a nice and considerate person. We make trait inferences spontaneously, unconsciously, and constantly, and they are an integral part of everyday social perception (Higgins, 1989; Higgins & Bargh, 1987). Third, social perceivers also go beyond the information actually present in the current environment through the activation of social stereotypes based on easily detectable identifying features of social groups (Brewer, 1988). Stereotypes are integrated collections of trait concepts purportedly descriptive of the social group in question. Unlike trait inferences, however, stereotypes represent mental activation that does not have a one-to-one correspondence with current events being perceived. Upon seeing a person, we automatically categorize that person as a member of his or her group based on these characteristics, and also, often if not usually the stereotype associated with that group becomes active as well (Bargh, 1999; Devine, 1989; Greenwald & Banaji, 1995; Lepore & Brown, 1997). Merely seeing an African American face (even subliminally) is sufficient to cause the activation of the stereotype of African Americans in randomly selected White U.S. college students (Bargh, Chen, & Burrows, 1996; Chen & Bargh, 1997). Stereotype activation, like trait inferences, occurs as a natural and automatic part of the process of everyday social perception. In sum, we perceive more than is literally present. Apart from perceiving observables, we make trait inferences and activate social stereotypes. As is demonstrated in the next section, all three forms of social perception elicit the tendency to imitate in the social perceiver.
10
AP DIJKSTERHUIS AND JOHN A. BARGH
V. Social Perception Elicits Corresponding Behavior A. OBSERVABLES In the following paragraphs, evidence of imitation of observable behavior is reviewed. The research on imitation of observables can be divided into three domains. First, there is a large literature on imitation of facial expressions. In addition, others have investigated imitation of gestures and movements. Finally, there is evidence of imitation of various speech-related variables. The major findings of all three domains are discussed, starting with facial expressions.
1. Facial Expressions The evidence for imitation of facial expressions is abundant (e.g., Dimberg, 1982; Vaughan & Lanzetta, 1980; Zajonc et al., 1982). An example of a very contagious facial expression that is familiar to all of us is yawning. If, after a long car or train ride, a person starts to yawn, usually his or her travel companions start to yawn within a few minutes. This tendency to imitate yawning has also been demonstrated empirically. Provine (1986) asked participants to watch a 5-min videotape. In one condition, participants watched a video with yawning people, whereas in a control condition participants watched a video with smiling people. As expected, 55% of the participants in the experimental (i.e., yawn) condition started to yawn while watching the video, as opposed to only 21% in the control (i.e., smile) condition. Interestingly, Provine also obtained evidence supporting our claim that activation of the mental representation of an action (which can be the result of perception but also of, for instance, thought) is crucial in eliciting corresponding behavior. That is, one does not have to literally perceive a yawn to engage in yawning. Provine found that reading about yawning or thinking about yawning also caused participants to yawn. Finally, the fact that one of the authors of this chapter is yawning right now can be taken as anecdotal evidence that writing about yawning does the trick as well. Although no consensus emerged among researchers on the exact cause of the phenomenon, various investigators have studied imitation of facial expressions among newborns (Anisfield, 1979; Field, Woodson, Greenberg, & Cohen, 1982; Jacobsen & Kagan, 1979; Meltzoff & Moore, 1977, 1979, 1983). Meltzoff and Moore (1977, 1979) showed that even 1-month-old babies imitate facial expressions. If you look at a baby and open your mouth, the baby will open her mouth. If you stick out your tongue, the baby will often do the same. An interesting early demonstration of imitation of facial expression among adults can be found in an experiment by O’Toole and Dubin (1968). Their experiment was aimed at investigating mother–child interactions during feeding.
THE PERCEPTION–BEHAVIOR LINK
11
They had observed that a mother would usually open her mouth just prior to feeding her infant a spoonful of food. Their intuitive explanation for this finding was that a mother would open her mouth in the hope that her child would do the same and—most importantly—that the food would end up where it is supposed to end up. They put their ideas to a test by watching various mother–infant interactions and observed indeed that both mothers and infants open their mouths. Surprisingly however, in almost 80% of the cases, a mother opens her mouth only after the child does so. In other words, it is the mother who is imitating the child, not vice versa. The child is merely opening his or her mouth upon perceiving the food on its way. Another example of adult imitation of facial expressions comes from experiments carried out by Bavelas and colleagues (Bavelas, Black, Lemery, & Mullett, 1986, 1987). In their experiments, a confederate was the victim of a painful injury that occurred in the presence of the participants. As expected, the participants imitated the expressions of the confederate, which can best be described as a big wince. Interestingly, they also manipulated the visibility of the expression of the confederate. In one condition, the expression of the confederate was easier to see than in a second condition. As a result, the degree to which participants imitated the expression varied as well. More visible expression led to more imitation; that is, the easier it was to perceive the expression, the greater the effect on one’s own behavior. Zajonc and colleagues (Zajonc, Adelmann, Murphy, & Niedenthal, 1987) reasoned that couples who have lived together for a period of time should have often experienced the same emotions at the same times, and because frequent facial expressions eventually lead to changes in facial lines, they hypothesized that partners should start to look more like each other the longer they are together. In their experiment, they gave participants 24 photographs. These photographs were those of the partners of 12 married couples. Some photographs were made at the wedding, whereas others were made 25 years later. The task of the participants was to assess the degree of resemblance of various pairs of photographs. As predicted, partners who were together for 25 years resembled each other more than random pairs of the same age and than newlywed couples. Although Zajonc et al. (1987) interpreted these findings in terms of shared emotional experience, these findings are also consistent with the present hypothesis of a direct effect of perception on behavior; that is, it may be that frequent perception of the partner’s expression leads one to adopt that same expression repeatedly oneself, producing over time the similarity in facial lines between the two partners (see Bargh, 2000). Imitation of facial expressions has also been studied in the context of emotional contagion (see, e.g., Hatfield, Cacioppo, & Berntson, 1994). Our facial expressions affect our emotions through a process of feedback elicited by facial muscles (Strack, Martin, & Stepper, 1988). Imitation of facial expression therefore leads to shared emotions. In concrete terms, the perception of a sad face evokes a sad expression in the perceiver and the perceiver will actually begin to feel sad as well.
12
AP DIJKSTERHUIS AND JOHN A. BARGH
In the Zajonc et al. (1987) research, the relation between shared facial expressions and shared emotions was obtained in a follow-up study. They had observed variations in the degree of resemblance of life partners. This led to the intriguing hypothesis that partners who have grown to look like each other more may actually be happier together than those who have not because their resemblance is due to a greater history of shared emotions. And, in general at least, shared emotions lead to a stronger bond between partners. A questionnaire study indeed confirmed this hypothesis, with effects being impressive in size (with a correlation of .49 between resemblance and self-reported happiness). 2. Behavior Matching The evidence concerning the imitation of movements and gestures is less abundant than the evidence on imitation of facial expressions. Although theorists have always treated the automatic imitation of postures, gestures, and movements as a given (e.g., Allport, 1968; K¨ohler, 1927), early “evidence” was almost entirely anecdotal (see Bavelas et al., 1986; for reviews, see Capella, 1981; Chartrand & Bargh, 1999; Dijksterhuis, 2001; LaFrance, 1979). Later reports, in which posture imitation (or posture mirroring, as it is more often called) was investigated experimentally, suffered from methodological weaknesses (Charney, 1966; Kendon, 1970). Finally, research in the 1970s and early 1980s was concerned not so much with the occurrence of posture imitation per se, but with the relation between imitation and rapport. These studies (e.g., Bernieri, 1988; LaFrance, 1979; LaFrance & Ickes, 1981) speak to the possible function of posture and gesture mirroring in that some experiments clearly show a strong correlation between posture imitation and rapport. However, they do not shed light on how often people spontaneously engage in posture imitation. The only early investigation we could identify that exceeds the level of mere anecdotal evidence was reported by Eidelberg (1929). In his experiment, participants were instructed to point at their nose upon hearing the word “nose” and to point at a lamp upon hearing the word “lamp.” The experimenter, who was clearly visible to the participants, also pointed at his or her nose or at the lamp upon hearing the corresponding instruction. After a while, the experimenter started to make “mistakes” in that he or she pointed at the lamp upon hearing the word nose and vice versa. Interestingly, participants started to make the same mistakes as well. They spontaneously imitated the gestures made by the experimenter, despite the instruction to follow the verbal cues (i.e., the words “nose” and “lamp”) and not the behavior of the experimenter. Bernieri (1988; see also Bernieri, Reznick, & Rosenthal, 1988) was the first to provide truly solid evidence for posture imitation. In his studies, a somewhat complicated but nonetheless ingenious paradigm was used. First, two participants (A and B) were asked to interact. While they interacted, they were videotaped. A
THE PERCEPTION–BEHAVIOR LINK
13
little later, both participants A and B were asked to engage in another interaction with a different participant, such that A interacted with C and B would interact with D. Again, both interactions were videotaped. Subsequently, two tapes were constructed on which the gestures and postures of both participants A and B were displayed. One concerned the actual interaction between A and B. The other tape pictured A while interacting with C and B while interacting with D. Subsequently, judges—who were unaware of which tape displayed the actual interaction between A and B—estimated the degree of posture similarity. If the degree of matching is greater on the first tape (the actual interaction) than on the second, there is evidence for posture matching. Bernieri (1988) indeed obtained this evidence. People do spontaneously mirror the postures of individuals they interact with. Chartrand and Bargh (1999) replicated and extended these effects. Instead of investigating posture mirroring, they focused on actions such as foot shaking or nose rubbing. In their first experiment, a confederate was instructed to either rub her nose or shake her foot while working with a participant on a task. Importantly, the two were strangers and had only a minimal interaction, greatly reducing the probability that any imitation was motivational in nature—such as part of an attempt to ingratiate the other person. Their hypothesis, that participants would mimic the behavior of the confederate, was confirmed. Under conditions where the confederate rubbed her nose, participants engaged more in nose rubbing than in foot shaking, whereas the opposite was true when participants interacted with the confederate who shook her foot. Chartrand and Bargh (1999) replicated and extended this finding in a second study, in which the confederate purposefully mimicked the body posture of the participant. This study obtained clear evidence that mimicry leads to increased liking of interaction partners. The lack of a motivational basis for these findings supports our thesis of an automatic link between social perception and one’s own behavior in a naturalistic interaction context. 3. Speech-Related Variables Finally, there is evidence of automatic imitation of various speech-related variables. One phenomenon that is investigated by several researchers is syntactic persistence, that is, the tendency to use a certain syntax when this syntax is made cognitively accessible. This phenomenon supports the common-coding approach to language comprehension and language production postulated by Prinz (1990). Prinz argues that we use the same mental representations for both comprehension and production of speech. According to Prinz (see also Studdert-Kennedy, 1987). language comprehension and production develop at the same time during ontogeny: “. . .the ability to produce language is of no use when there is no one to listen, and the ability to understand language is of no use when there is no one to produce it” (p. 177).
14
AP DIJKSTERHUIS AND JOHN A. BARGH
Bock (1986, 1989) reported evidence of syntactic persistence. In one experiment, participants would hear and repeat a sentence such as “The corrupt inspector offered a deal to the bar owner.” Later, participants would see a picture of, for instance, a boy handing a valentine to a girl. This picture can be described as “The boy is handing a valentine to a girl” or as “The boy is handing the girl a valentine.” As the first sentence has a syntactic form similar to that of the priming sentence, this is the description participants most often gave. Syntactic structures appear to carry over from one sentence to another. Whereas in the studies conducted by Bock (1986) participants activated a particular syntax themselve, Levelt and Kelter (1982; see also Schenkein, 1980) investigated syntactic persistence in a social context. In one of their experiments, the experimenter called various shops and asked either “What time does your shop close?” or “At what time does your shop close?” If the former question was asked, shopkeepers more often answered with “Five o’clock,” whereas the answer to the latter question was “At five o’clock” in the majority of cases. Importantly, Levelt and Kelter as well as Schenkein obtained such effects of speech imitation for single words and clauses as well as for the structural format of entire sentences. Finally, Levelt and Kelter showed that cognitive load did not increase these speech imitation effects (which were already very substantial under normal conditions), suggesting that these effects were automatic in nature. Recently, Neumann and Strack (2000) obtained evidence for imitation of tone of voice between interaction partners. In one of their experiments, participants listened to an audiotaped speech given by a stranger. While they were listening, participants were asked to repeat what they heard and were audiotaped themselves. It was found that participants adopted the tone of voice of the person on the tape they listened to. A sad tone of voice on the tape elicited a sad tone of voice in the participant, whereas a happy voice led to a happy voice in the participant. These findings are particularly important, as they rule out the possibility that participants imitated tone of voice for strategic reasons (e.g., to increase cohesion). They did not see the person who delivered the speech, they did not even know who this person was, and no participant was aware of the actual goal of the experiment. Instead, they were successfully led to believe that the experimenters were interested in the reproduction of speech content. 4. Are Emotion and Behavior Matching Strategic? Bavelas and colleagues (Bavelas et al., 1986, 1987) accounted for their findings with a motivational communicative perspective. They argue that participants imitate in order to show the confederates that they are empathizing with them, that they are “feeling their pain.” And if there is more eye contact between the confederates and the participants, the participants imitate more because they know
THE PERCEPTION–BEHAVIOR LINK
15
that the confederates are better able to see their expression. In other words, they interpret the imitation as a motivated, strategic behavior to create an empathic bond with the other person. This model of imitation (which is, according to the division we made earlier between “facilitator option” and “inhibitor option, an example of a facilitator option) is the standard account in the field not only of facial mimicry, but of the related phenomena reviewed above of “behavior matching” (La France, 1979, 1982) and “rhythmic synchrony” (Bernieri, 1988; Condon & Ogston, 1966; Condon & Sander, 1974). Most of this research has sought to link behavioral coordination effects with the establishment of rapport and liking between the parties involved, with some researchers viewing empathy as the cause of mimicry and others considering mimicry to be the cause of empathy (see Bernieri & Rosenthal, 1991; Chartrand & Bargh, 1999, for reviews). Although it is true that there tends to be greater mimicry when the two individuals like each other than when they do not (e.g., Charney, 1966; LaFrance & Broadbent, 1976), so that rapport between the parties is an important moderator of the effect (see Moderators of the Perception–Behavior Link below), this does not mean that the perception–behavior effect requires for its occurrence a motivation or strategy or even positive affect toward the other person as a necessary condition. After all, the evidence reviewed above shows that the only real precondition of imitation of observable behavior is the perception of the behavior. We emphasize that our explanation of an innate express route between perception and action is supported by this evidence, as our explanation would lead one to predict all the reviewed effects to be automatic and nonstrategic as opposed to other explanations that claim these effects to be strategic and intentional. There is no evidence at all for the strategic nature of the imitation effects reviewed above, whereas the support for the automatic and unintentional nature of imitation is evident. Meltzoff and Moore (1977) demonstrated a tendency to imitate among newborns. O’Toole and Dubin (1968) showed that mothers tend to imitate their children and there really is no strategic reason to do so. Although Bernieri (1988) showed imitation among people who engaged in an extended interaction (potentially allowing the interactants to engage in motivated imitation), Chartrand and Bargh (1999) showed that even minimal interaction with a complete stranger led to imitation. Finally, Neumann and Strack (2000) obtained evidence for imitation of tone of voice when the person being imitated was not even present. In sum, there is considerable evidence showing that people automatically imitate observed behavior—ranging from facial expression and postures to speech patterns. There is no evidence for the strategic nature of the imitation effects, whereas the support for the automatic and unintentional nature of imitation is evident. That is, in the experiments reviewed above, people did not imitate because they wanted to imitate. Instead, they imitated for no other reason than that they are designed to do so.
16
AP DIJKSTERHUIS AND JOHN A. BARGH
B. TRAIT INFERENCES As alluded to earlier, social perception entails much more than the encoding of observable behavior. We tend to automatically encode a person’s social behavior in terms of the trait concepts relevant to it (e.g., Bargh, 1994; Gilbert, 1989; Higgins, 1989; Winter & Uleman, 1984; Uleman, Newman, & Moskowitz, 1996). In this section, we review the evidence demonstrating that the automatic activation of personality trait constructs in the course of social perception leads to behavior corresponding to these constructs. It leads, in other words, to imitation. If we see a person walk very slowly, we automatically infer the trait “slow,” and we automatically tend to become slow. The evidence for trait-induced social or interpersonal behavior is abundant. In a seminal study, Carver, Ganellen, Froming, and Chambers (1983) primed the concept of hostility among half of their participants by incidentally exposing them to words related to this concept. The remaining half of the participants were not primed with hostility. Subsequently, participants played the role of a teacher in a learning task based on the classic experiment of Milgram (1963). Participants had to administer electrical shocks to a second participant (actually a confederate) whenever this second participant gave an incorrect answer to a question. The participants, however, were free to choose the intensity of the shocks. The results showed that participants primed with hostility delivered more intense shocks than did control participants. In other words, priming hostility indeed led to more hostile behavior. Various other social behaviors have been shown to be affected by activated traits and stereotypes as well. Bargh, Chen, and Burrows (1996, Experiment 1) presented their participants with a scrambled sentence task in which they were to construct grammatically correct sentences out of a random ordering of words (see Srull & Wyer, 1979) as a purported test of language ability. In one condition, the scrambled sentences contained some words related to rudeness (e.g., aggressively, bold, rude), whereas in a second condition the scrambled sentences contained some words related to politeness (e.g., respect, patiently, polite). In a third condition, the scrambled sentence task did not contain words related to either rudeness or politeness. The experimenter left the room after the participants had been given the instruction necessary to complete the scrambled sentence task. Participants were requested to meet the experimenter in a different office upon finishing the scrambled sentence task. When participants approached the experimenter, the experimenter was talking to a confederate. The confederate surreptitiously measured the time it took for participants to interrupt the conversation. Participants who were primed with rudeness were more likely (63%) than control participants (38%) to interrupt whereas participants primed with politeness were least likely to interrupt (17%). In experiments reported by Macrae and Johnson (1998), consequences of activation of the trait helpful were investigated. In their experiments, half of the
THE PERCEPTION–BEHAVIOR LINK
17
participants were primed with the concept of helpfulness through the use of a scrambled sentence task, whereas the remaining participants were not primed. Upon finishing the task, the experimenter picked up her possessions from a desk (books, a paper, a bag, pens) and asked the participants to follow her to another experimenter. As she approached the door, she “accidently” dropped some of the items she was carrying. As expected, participants primed with helpfulness picked up more items from the floor (i.e., behaved in a more helpful way) than did control participants. Epley and Gilovich (1999) primed participants with stimuli related to either conformity or nonconformity. A third group of participants was not primed. Later, participants were asked to evaluate various aspects of the experiment in the presence of a number of confederates, who expressed their favorable evaluations before the participants were given the opportunity to do so. Participants primed with conformity conformed more to the confederates (i.e., evaluated the experiment more positively) than no-prime controls and than participants who were primed with nonconformity. Participants primed with nonconformity, however, did not conform less than no-prime controls. There are various explanations for this asymmetric finding (see Epley & Gilovich, 1999); in our view the most likely being that the social pressure on participants to conform in the experimental situation was rather strong, leaving less room for the nonconformity prime to be effective. To summarize, activation of trait concepts elicits corresponding behavior. Activation of the trait rude makes us rude and activation of the trait helpful makes us helpful. It is also evident that the effects are not restricted to a particular behavioral domain. Our tendency to imitate or to match behavior of our social environment seems to affect many forms of overt social behavior.
C. SOCIAL STEREOTYPES The automatic activation of social stereotypes in the course of perceiving another person produces the same effects on behavior as does the activation of single-trait concepts, because stereotypes are to some extent schematic knowledge structures composed of several different trait concepts, ostensibly descriptive of the stereotyped group. However, the trait concept becomes activated in perception either because of trait-relevant behavior by the other person or because it participates in a cultural stereotype relevant to the perceived person; it will have the same effect on one’s own behavior. For example, if we meet an elderly person, the category elderly becomes activated as well as associated traits such as slow. In both cases, the activation of the trait construct slow will guide one’s behavior irrespective of why or how the trait was activated. In what follows, we review evidence of stereotype activation on motor behavior, on various forms of interpersonal behavior, and on intellectual performance.
18
AP DIJKSTERHUIS AND JOHN A. BARGH
Bargh et al. (1996, Experiment 2) were the first to report effect of stereotype activation on motor behavior. In their experiment, some participants were primed with the stereotype of the elderly, whereas others were not. The participants in the experimental condition were primed by exposing them to words related to the elderly (i.e., gray, bingo, Florida) in the context of a scrambled sentence language task, whereas participants in the control condition were not exposed to these words. After participants finished the priming task, they were told that the experiment was over. A confederate, however, recorded the time it took participants to walk from the experimental room to the nearest elevator. The data clearly showed that participants primed with the elderly stereotype walked significantly slower than control participants. In other words, people displayed behavior corresponding to the activated stereotype. Elderly are associated with slowness, and activating the stereotype of the elderly indeed led to slowness among the participants. A conceptual replication of these findings was reported by Kawakami, Young, and Dovidio (2000). In their experiments, some participants were presented with various photographs of elderly people, whereas others were presented with photographs of university students. The photographs were primes in a lexical decision task. Each photograph was accompanied by a personality trait and the task of the participant was to decide whether the presented traits were descriptive of the social category displayed on the photograph (elderly vs student). As would be expected from the present thesis, reaction latencies on the words were longer when the words were preceded by a photograph of an elderly person than when the words were preceded by photographs of younger people. Dijksterhuis, Spears, and Lepinasse (2001) obtained comparable results in a different paradigm. In their study, some participants were instructed to form an impression of various elderly individuals while looking at the photographs of these individuals. The second task, which was ostensibly unrelated to the first task, was a lexical decision task in which participants were asked to decide as fast as possible whether words presented on the screen were existing words (car, shop) or random letter strings (ikn, geru). Participants primed with the elderly stereotype showed reaction times that were considerably slower than participants who were not primed. In sum, activation of the elderly stereotype makes one slow, whether it pertains to one’s walking speed or one’s reaction time. As noted above, crucial in the onset of behavioral changes are trait constructs. We can infer the trait slow from seeing someone walking slowly or we can activate the trait slow because it is part of an activated stereotype. But of course, there are other ways. We can, for instance, activate the trait slow by presenting participants with very slow animals. Theoretically, this should lead to slowness as the relevant concept is activated. There is no reason to assume that our brain makes a difference between whether slowness is activated because of exposure to animals or to members of a stereotyped group.
THE PERCEPTION–BEHAVIOR LINK
19
Aarts and Dijksterhuis (2001) investigated this possibility. They obtained evidence demonstrating that priming participants with names of animals also affects motor behavior. In their study, participants were either primed with animals associated with speed (cheetah, antelope) or with animals (snail, turtle). Subsequently, participants were asked to pick up a questionnaire in an adjacent room. The time it took participants to collect the questionnaire was assessed. In line with predictions, participants primed with fast animals were considerably faster than participants primed with slow animals. This study shows that we can also automatically imitate animals and not just fellow human beings. Bargh et al. (1996, Experiment 3) also demonstrated effects of stereotype activation on interpersonal behavior. In their experiment, participants were seated behind a computer and were asked to engage in a very laborious task. While engaging in this task, some participants were subliminally primed with photographs of male African Americans, whereas others were subliminally presented with male Caucasian faces. After participants had been performing the laborious task for a while, the computer program beeped and displayed an error message stating “F11 error: Failure saving data.” Subsequently, the experimenter pressed a button upon which the message “You must start the program over again” appeared. The participants were videotaped during these moments and the dependent variable was the level of hostility participants displayed upon hearing that they had to start all over again. As expected, both the experimenter (who was blind to conditions) and several independent coders rated the reaction of the participants primed with the stereotype of African Americans as more hostile than the reaction of the participants primed with Caucasian faces. This finding was replicated and extended to the domain of self-fulfilling prophecy effects by Chen and Bargh (1997). Dijksterhuis and van Knippenberg (2000) demonstrated behavioral effects of activation of the stereotype of politicians. In earlier work, they had established that politicians are associated with longwindedness. That is, people believe that politicians talk a lot without saying much. In an experiment, Dijksterhuis and van Knippenberg activated the stereotype of politicians with the use of a scrambled sentence procedure for half of their participants. Subsequently, participants were asked to write an essay in which they argued against the French nuclear testing program in the Pacific (this experiment was carried out in 1996). As expected, participants primed with politician-related stimuli wrote essays that were considerably longer those of control participants. A third domain in which it has been demonstrated that stereotypes and traits lead to corresponding behavior concerns the domain of intellectual (or mental) performance. Dijksterhuis and van Knippenberg (1998) improved people’s intellectual performance in a series of experiments. In their first experiment, they requested half of their participants to think about college professors and to write down everything that came to mind regarding the typical attributes of professors. The remaining half of the participants were not given this task. In an ostensibly
20
AP DIJKSTERHUIS AND JOHN A. BARGH
unrelated second experiment, participants were asked to answer 42 general knowledge questions that were taken from the game “Trivial Pursuit” [such as “What is the capital of Bangla Desh?” (a) Dhaka, (b) Bangkok, (c) Hanoi, or (d) Delhi]. In line with the prevailing stereotype of professors as being intelligent, primed participants answered more questions correctly than did no-prime control participants. In their set of studies, it was also shown that the magnitude of the change in intellectual performance was a linear function of the strength of the priming manipulation. Participants primed for longer durations outperformed participants primed for shorter durations, who in turn outperformed participants who were not primed. In another experiment conducted by Dijksterhuis and van Knippenberg (1998), it was shown that participants could also be led to perform worse on a general knowledge task by having them think previously about soccer hooligans, a social group that is associated with stupidity. It has also been shown that activation of the stereotype of the elderly affects memory performance (Dijksterhuis, Aarts, Bargh, & van Knippenberg, 2000; Dijksterhuis, Bargh, & Miedema, 2000; Levy, 1996). In an experiment conducted by Levy, elderly participants were primed with either positive (e.g., wise, experienced) or negative (e.g., senile, dementia) terms associated with the elderly. Subsequently, participants were asked to perform various memory tasks. As she predicted, priming positive words led to improved memory performance, whereas priming negative words led to deteriorated performance. Indeed, Dijksterhuis, Bargh, and Miedema (2000) did obtain evidence showing that activation of the “elderly” stereotype affects memory performance among college students (i.e., participants for whom the stereotype is not self-relevant). In their experiment, participants were seated behind a desk on which 15 objects were placed (a book, a pencil, a bag, etc.). Some participants were asked to answer questions about elderly people (“How often do you meet elderly people?” “Do you think elderly people are conservative?”), and others were asked to answer questions about college students. After answering questions for 3 min, participants were placed in a different experimental room and asked to recall as many objects present in the previous room as possible. As expected, participants primed with the elderly stereotype recalled fewer objects than other participants. The deteriorating effects of activation of the elderly stereotype on memory have been replicated and extended by Dijksterhuis, Aarts, Bargh, and van Knippenberg (2000), who used subliminal priming procedures and different memory paradigms. 1. Relation to Stereotype Threat Stereotype priming effects on behavior bear a close relation to the well-known phenomenon of “stereotype threat” (e.g., Aronson, Quinn, & Spencer, 1998; Levy, 1996; Shih, Pittinsky, & Ambady, 1999; Steele, 1997; Steele & Aronson, 1995). When aspects of one’s identity related to task performance are made salient,
THE PERCEPTION–BEHAVIOR LINK
21
performance is affected in the direction of that aspect of identity. Asian American children as young as 5 years old do better on math tests if their Asian aspect of identity has just been made salient; if their female identity is made salient, girls do worse on the same math test; African Americans do worse on academic tests in general if their racial group membership is subtly made salient (through a standard first questionnaire on which one checks off one’s racial and ethnic group identity); women do worse on math tests if their gender identity is made salient; and elderly participants do worse on memory tests if their age was just made salient. All of these effects are attributable to the trait concepts activated or primed by the identity salience manipulation—of being poor in math, having poor memory, and so on. But the conclusion generally drawn from these studies is that activation of stereotypes only affects the behavior of members of those stereotyped groups. For example, Levy (1996) reported that elderly participants performed better on a memory task when primed with positive aspects of the elderly stereotype but worse when primed with negative aspects, and she reported no such effects of the elderly stereotype priming on college-age participants. She concluded that activated stereotypes only exert behavioral effects when these stereotypes are selfrelevant. However, as reviewed above, there are now many studies showing stereotype priming effects among nongroup members. In the case of the elderly stereotype in particular, Bargh et al. (1996, Experiment 2) found college students to walk more slowly leaving the experimental session after priming with the elderly stereotype, and Dijksterhuis et al. (2000) found poorer incidental memory performance for elderly-primed college students. In a study related to a different case of stereotype threat, Dijksterhuis and Corneille (2001) examined the cultural sterotype of women as being poor at mathematics. They demonstrated that subliminally priming participants with the female stereotype by exposing them to words related to this sterotype reduced performance (of both male and female participants) on a calculus task relative to participants who were not primed. This finding was replicated in several experiments. Thus, when considered in the context of the now-abundant research showing stereotype activation effects on behavior of randomly selected participants (reviewed above)—that is, for those who are not members of the stereotyped group— the following synthesis between stereotype threat research and perception–behavior research can be made. Stereotype priming effects on the behavior of a member of the stereotyped group (i.e., stereotype threat) are likely to be larger (and easier to obtain) because for group members there are two routes, not just one, to the representation of that kind of behavior. The first is the activated stereotype, but the second is the person’s self-representation or social identity, which constitutes a second and strong source of activation of the particular trait concept representation. Nongroup members have just the one route, through the activated stereotype (or perceived behavior). In other words, as argued above, stereotype threat and
22
AP DIJKSTERHUIS AND JOHN A. BARGH
perception–behavior effects share the same common mechanism (the perceptually activated stereotype or stereotype-relevant trait representation) but for stereotyped group members this representation is, in effect, activated twice, producing still stronger effects on behavior. Wheeler, Jarvis, and Petty (2001) received some support for this idea. They primed their participants with the stereotype of African Americans, after which these participants performed worse on a math test than control participants. The participants were all non-African Americans, but Wheeler et al. (2001) reported an interesting moderator. Their priming procedure consisted of the instruction to write about a day in the life of a certain individual (a person with either a typical African American name or a typical Caucasian American name). Some participants wrote their short stories in the first person (thereby adopting the perspective of the target), while others wrote their stories in the third person. The participants who wrote their stories in the first person showed stronger priming effects than participants who wrote in third person. One way to explain the findings is to assume that participants who wrote in the first person identified more with the target person and, because of this identification, activated their stereotype to a greater extent. On the other hand, Dijksterhuis and Corneille (2001) did not obtain stronger effects of stereotype activation for group members than for nongroup members. In their research, performance on a calculus task of male and female participants deteriorated to the same extent after activation of the female stereotype. Even a manipulation of task diagnosticity (known to increase the salience of task-related identity and hence to enhance stereotype threat) did not lead to further deterioration of performance of female participants when this manipulation was combined with a priming manipulation. That is, performance on a subsequent calculus task showed that the effects of the two manipulations were not additive. However, it is possible that the priming procedure used by Dijksterhuis and Corneille (2001) activated the stereotype to such an extreme degree that additional activation, and therefore further deterioration of performance, was prevented.
VI. Mediators of the Perception–Behavior Link A. THE ROLE OF BEHAVIOR REPRESENTATIONS In the next section, mediational evidence is presented that sheds light on the process by which perception affects overt behavior. How, in other words, can perception of a smile or activation of an abstract construct such as a stereotype lead to behavioral changes such as decreased walking speed or altered intellectual performance? Let us first discuss the relation between behavior representations and actual behavior.
THE PERCEPTION–BEHAVIOR LINK
23
We discussed the idea of so-called ideomotor action under “Neurophysiological Evidence”. The principle of ideomotor action, introduced long ago (Carpenter, 1874; James, 1890; Jastrow, 1908), states that merely thinking about an action leads automatically to the tendency to engage in this action. As James (1890, p. 522) defined it, “Whenever movement follows unhesitatingly and immediately the notion of it in mind, we have ideo-motor action.” If we translate this line of thinking in the psychological language we use today, ideomotor action implies that the activation of a behavioral representation elicits the tendency to engage in this same behavior. In concrete terms, the activation of the mental representation of “walking” should lead to the tendency to walk. This in itself is not surprising. It is hard to see how we can be able to walk without first activating some neural correlate of this behavior in the brain. What was important for the theorists cited earlier was not only that such behavior representations were activated after conscious decisions (“let’s walk”) but also that a fleeting notion of the behavior was enough to evoke the behavior itself. As alluded to earlier, recent techniques developed in the neuropsychological domain allow a test of these old ideas. Paus, Petrides, Evans, and Meyer (1993), for instance, have shown that thinking about a word or a gesture leads to the same activation in the anterior cingulate cortex as actually uttering the word or making the gesture. Jeannerod and colleagues have conducted a series of experiments in which they demonstrated that imagining somewhat more complex actions (such as running, rowing, or weightlifting) has neurophysiological consequences that are largely comparable to the neurological consequences of actually engaging in an action (Decety, Jeannerod, Germain, & Pastene, 1991; Jeannerod, 1994, 1997). Crucial in their research program are so-called motor programs, as these programs are ultimately reponsible for overt behavior. As Jeannerod and others have shown, imagining an action leads to activation of exactly the same motor programs as does performing the action. To conclude, what is needed for behavioral changes are activated motor programs and what is needed for activated motor programs are activated behavior representations. And this, we argue, is what happened in the experiments reviewed in this chapter. In all these studies behavior representations were activated. The important difference between the various experiments is the process leading to activation of behavior representations. Imitations of facial expressions, speechrelated variables, and gestures and postures are the consequences of the mere perception of these behaviors in others. It is easy to see how relevant behavior representations are activated in these cases. One perceives a smile, and this is enough to activate the representation of a smile, which in turn is enough to activate the programs controlling facial muscles. In the same vein, perceiving a gesture (as, e.g., in Chartrand & Bargh, 1999) activates the representation of this gesture, presumably in the same way as thinking about a gesture does (as in Paus et al., 1993).
24
AP DIJKSTERHUIS AND JOHN A. BARGH
B. FROM STEREOTYPES TO IMITATION However, the research on imitation mediated by activated stereotypes or trait concepts is different in that activation of behavior representations is mediated by activation of intermediate representations. Research shows that upon activation of a social category (e.g., elderly), associated stereotypic traits (e.g., slow, forgetful) are also activated (Blair & Banaji, 1996; Devine, 1989; Dijksterhuis & van Knippenberg, 1996; Dovidio, Evans, & Tyler, 1986; Macrae, Stangor, & Milne, 1994). As one would expect, effects of stereotype activation on behavior are mediated by trait activation. Dijksterhuis, Aarts, Bargh, and van Knippenberg (2000) showed that activation of the stereotype of the elderly led to forgetfulness, but only among participants who indeed associated elderly with forgetfulness. That is, only participants who indeed activated the trait forgetfulness after being primed with the category elderly display actual forgetfulness. Hence, the effects of stereotype activation on behavior are mediated by activation of traits. These effects were replicated by Dijksterhuis and Corneille (2001), who showed that only participants who associated women with poor math performance suffered from poor performance after being primed with the female stereotype. Trait concepts, in turn, can activate behavior representations. Activating the trait slow leads to activation of more concrete behavior representations such as linger or dawdle, whereas the trait intelligent leads to activation of behavior representations such as concentrate and think. In a recent study, Dijksterhuis and Marchand (2000) showed that activation of the stereotype of professors leads to activation of such concrete behavior representations as think and concentrate. Furthermore, these effects were mediated by activation of the trait intelligent. Behavioral representations on the level of abstractness of think or dawdle activate motor programs, as Jeannerod and others have shown for comparable behavior representations such as run or row. In sum, stereotypes can automatically affect behavior because they activate—via the activation of traits and behavior representations—motor programs. To recapitulate, the effects of stereotype activation on changes in overt behavior can be explained by a series of steps. First, stereotypes activate associated traits. These traits, in turn, activate more concrete behavior representations. Finally, these behavior representations activate the motor programs responsible for actual behavior.
VII. Behavioral Contrast The findings reviewed in the previous section demonstrate that primed traits and stereotypes elicit corresponding behavior in the perceiver. These behavioral
THE PERCEPTION–BEHAVIOR LINK
25
effects can be characterized as manifestations of behavioral assimilation and are reminiscent of findings from the social judgment domain in which it has been demonstrated that primed constructs—such as traits—lead to judgmental assimilation (e.g., Higgins, Rholes, & Jones, 1977; Srull & Wyer, 1979, 1980). In Higgins et al.’s (1977) seminal demonstration, participants were surreptitiously primed with either the positive or the negative version of traits that could both be used to describe the same type of behavior (e.g., adventurous vs reckless, independent vs aloof ). Later, they were asked to form an impression of a person named Donald who performed ambiguous behaviors that could each be interpreted in either a positive or a negative way. These impressions showed that the primed traits indeed led to assimilation: Participants primed with the positive set of traits formed more positive impressions of Donald, whereas those primed with negative traits formed more negative impressions. Importantly, priming with positive and negative traits that were unrelated to the later behaviors had no such effects on impressions. Trait primes, it has often been argued, lead to assimilation because they work as interpretation frames, causing perceptual input to be interpreted in line with this trait construct (Higgins, 1996; Schwarz & Bless, 1992; Stapel, Koomen, & van der Pligt, 1996). The effects of primed traits on judgments are comparable to those on behavior. In one case traits guide perceptual input, while in the other case traits guide behavioral output. However, in both cases accessible traits constructs direct ongoing processes in an assimilative fashion. The social judgment literature has, however, demonstrated a second effect. Under some conditions, primes do not elicit judgmental assimilation; instead, they elicit jugdmental contrast (e.g., Stapel, Koomen, & van der Pligt, 1996, 1997). That is, the prime led to a biasing effect in the direction opposite to what was implied by the prime. Crucial in causing such contrast effects are comparisons. Priming Adolf Hitler still activates the concept of hostility but it also renders a more likely comparison between Hitler and the stimulus person to be judged. These comparisons, in turn, elicit contrast effects. If one is primed with Hitler and asked to judge a somewhat hostile person named Donald, a comparison between Donald and Hitler will lead to a less hostile assessment of Donald (“Well, Donald isn’t that hostile”). In recent years, judgmental contrast effects after exemplar priming have been documented extensively. It is now known that comparisons—and hence contrast effects—are more likely to occur if the exemplar is extreme rather than moderate and sufficiently concrete and when the comparison is relevant under the circumstances at hand (e.g., Herr, 1986: Herr, Sherman, & Fazio, 1983; Manis, Nelson, & Shedler, 1988; Schwarz & Bless, 1992; Stapel, Koomen, & van der Pligt, 1996, 1997). The notion of judgmental contrast prompted the question of whether it would also be possible to demonstrate behavioral contrast. Is it possible, in other words, that exemplar priming leads to behavioral contrast by evoking a comparison between the primed exemplar and the self? Dijksterhuis, Spears, et al. (1998)
26
AP DIJKSTERHUIS AND JOHN A. BARGH
tackled this question. In their first experiment, they primed participants with either stereotypes or exemplars. Both the stereotypes and the exemplars could designate intelligence or lack of intelligence. Concretely, participants were primed with stimuli related to professors, to supermodels, or to specific exemplars such as Albert Einstein or Claudia Schiffer. After the priming procedure, participants were asked to answer a number of general knowledge questions. As expected, priming stereotypes led to behavioral assimilation. Participants primed with professors outperformed those primed with supermodels (as in Dijksterhuis & van Knippenberg, 1998). However, priming exemplars led to behavioral contrast. Participants primed with Einstein performed worse than participants primed with Claudia Schiffer. These effects of behavioral contrast were also demonstrated in the paradigm first used by Bargh, Chen, and Burrows (1996). Whereas they had shown that priming the elderly stereotype led participants to walk slower, Dijksterhuis, Spears, et al. (1998) showed that priming an elderly exemplar (the 88-year-old Dutch Queen Mother) prompted participants to walk faster. In a third experiment, evidence was obtained indicating that a comparison between the primed exemplar and the self is indeed crucial for contrast to occur. It was shown that priming professor only led to a heightened accessibility of the construct of intelligence, whereas priming Einstein led to the formation of an association between the self-concept and the construct of stupidity. In other words, after priming Einstein—but not after priming professor—participants draw the conclusion “I am stupid,” reflecting the comparison they made between Einstein and themselves. In a second series of studies (Dijksterhuis, Spears, & Lepinasse, 2001), behavioral assimilation and behavioral contrast were related to regular impression formation processes. The fact that exemplars lead to contrast whereas traits (and also stereotypes) lead to assimilation led to the more general assumption that concrete stimuli lead to contrast whereas more abstract stimuli lead to assimilation. From the impression formation literature it has been known that person perceivers usually from rather abstract, stereotypical impressions of people, whereas on some occasions they form more concrete, individuated impressions (see, e.g., Bodenhausen, Macrae, & Sherman, 1999; Brewer, 1988; Fiske & Neuberg, 1990). Dijksterhuis, Spears, and Lepinasse (2001) applied this knowledge to the domain of behavioral contrast. In various experiments, they asked participants to form impressions of an elderly person or of elderly people, while manipulating various moderators that are known to affect the stereotypicality of impressions people form. In the first study, participants formed an impression of either one elderly person or five elderly people. In a subsequent reaction time task, participants who formed an impression of five elderly people showed assimilation (they became slower), whereas participants who formed an impression of a single elderly person showed contrast (they became faster). In a second study, it was demonstrated that whereas an impression of a single elderly person under normal circumstances led to contrast,
THE PERCEPTION–BEHAVIOR LINK
27
an impression made under conditions of cognitive load—known to lead to more stereotypical impressions (Bodenhausen & Lichtenstein, 1987; Dijksterhuis & van Knippenberg, 1995; Macrae, Hewstone, & Griffiths, 1993)—led to assimilation. Finally, although an impression of five elderly people was demonstrated to lead to assimilation, an instruction to form an impression as accurately as possible— known to lead to less stereotypical impressions (Erber & Fiske, 1984; Tetlock, 1992)—led to behavioral contrast. In sum, more stereotypical impressions led to behavioral assimilation, whereas individuated impressions led to behavioral contrast. In general, more concrete stimuli can lead to behavioral contrast, whereas more abstract stimuli lead to behavioral assimilation. There are exceptions to this rule, however. It has already been mentioned that for an exemplar to lead to behavioral contrast, this exemplar has to evoke a comparison. Such comparisons only lead to contrast if the comparison is made on a dimension relevant for the behavior at hand (see, e.g., Biernat, Manis, & Nelson, 1991; Stapel, Koomen, & van der Pligt, 1997). In concrete terms, a comparison with Einstein will lead me to conclude that I am not that intelligent after all, and this leads to contrast on a task measuring intelligence. However, it will not lead to effects on tasks that are unrelated to intelligence. Macrae et al. (1998) demonstrated that such comparisons have to be rather specific in order to elicit contrast. They primed participants with the former Formula One world champion Michael Schumacher. Later, participants were requested to perform a counting task during which their speed was measured. One might have predicted that Schumacher would lead to a comparison and that this comparison would lead to the conclusion among participants that they are slow. Such a comparison should in turn lead to contrast; that is, participants should become slow. However, Macrae et al. (1998) found an assimilation effect. Participants primed with Schumacher became faster. It is probably the case that Schumacher made the construct of speed accessible, thereby causing assimilation, while the comparison was not relevant enough for the task to cause contrast. It is likely that a comparison with Schumacher does not lead to the conclusion “I am slow,” but that it leads to the more narrow conclusion “I am a slow driver.” Given that the task was a counting task that had nothing to do with driving, the comparison was not relevant enough to cause contrast (see also Aarts & Dijksterhuis, 2001, for a more elaborate discussion of this explanation). Aarts and Dijksterhuis (2001) also obtained evidence for the important role of comparison processes in causing behavioral contrast. They primed participants with either slow (snail, turtle) or fast (cheetah, antelope) animals. Later, participants were asked to pick up a questionnaire in a different experimental room. Unbeknown to the participants, the time it took the participants to collect this questionnaire was measured. Importantly, Aarts and Dijksterhuis (2001) manipulated the perceived comparability of the animals. In one condition, they emphasized the similarities between humans and other animals, whereas in a second condition they
28
AP DIJKSTERHUIS AND JOHN A. BARGH
emphasized the difference between humans and animals. This perceived comparability proved to be an important moderator. Participants who were led to believe humans and other animals are comparable showed behavioral contrast (that is, participants primed with fast animals became slow and people primed with slow animals became fast), whereas participants who were led to believe humans and animals to be completely different showed assimilation. To summarize, social perception does not always lead to assimilation or imitation. The behavioral contrast findings parallel the judgmental findings: Whereas abstract constructs lead to assimilation, concrete stimuli such as exemplars may led to contrast provided they are extreme enough and the comparison being made is relevant for the behavior under consideration.
VIII. Moderation of the Perception–Behavior Link Earlier, we argued that unconscious imitation is a consequence of the way we have been “built.” Perception is linked to behavior and the activation of a perceptual representation evokes the corresponding action. However, we also argued that these effects could sometimes be inhibited or moderated. Without the possibility to moderate direct effects of perception on behavior, we would indeed behave like the fish or frogs discussed earlier. Our action would always directly follow from our perception without any flexibility whatsoever. We know that this is not the case, of course. Humans are flexible and they can override direct effects of perception on behavior. We do possess a set of moderating modules clearly separating us from fish and frogs. In the next section, we review the findings on moderators of the perception–behavior link.
A. DISINCENTIVES There may be clear costs associated with the perceived behavior based on one’s prior experience that prevent one from engaging in the perceptually suggested behavior. Unlike lemmings, who follow their mates right over the cliff, humans have some knowledge and experience with the painful consequences of falling substantial distances. Disincentive values of stimuli can produce counterforces on behavior that override the perceptual effect on behavior (Bargh & Ferguson, 2001), as shown in a study by Macrae and Johnston (1998). In their first experiment, participants were primed with stimuli either related to helpfulness or not, after which a confederate accidently dropped a number of pens. Under usual circumstances, primed participants indeed displayed more helpfulness—they picked up more items. However, when the pens were leaky—with a clear disincentive
THE PERCEPTION–BEHAVIOR LINK
29
or cost to the act of picking them up—participants were hesitant to help both under priming conditions and under no-prime control conditions. (This finding is reminiscent of an earlier one by Langer, Blank, & Chanowitz, 1978, in which participants were likely to allow a person to cut in front of them in line for clearly bogus reasons except when the cost of doing so was high—as when the requestor was going to make a lot of copies.)
B. CONFLICT WITH CURRENT GOALS AND PURPOSES Macrae and Johnston (1998) used a similar experimental setup to show the moderating role of goals. Again, some participants were primed with helpfulness, whereas other participants were not. And again, after participants were primed a confederate dropped a number of objects. In this experiment, however, some participants were told that they were running late and they had to hurry to the next experimental session. As it turned out, the goal to hurry overruled the effects of priming. Primed participants were more helpful than their no-prime counterparts only under normal circumstances. They were not more helpful when the conflicting goal to hurry was active. These findings suggest that passive effects of perception on behavior are dominated by currently operating goals, when the behaviors required for goal attainment are in conflict with those suggested perceptually. Such a model of action control, with operating goal pursuits inhibiting or overruling automatic access to the motorium, has been proposed by Shallice and his colleagues for many years (1988; Norman & Shallice, 1986). This model is also in harmony with the substantial literature on flexible working memory processes in which task goals can override automatically suggested responses if given enough time and attention (Fazio, Sanbonmatsu, Powell, & Kardes, 1986; Neely, 1977). A wellknown example of this ability is the Stroop color–word task, in which people are generally able to make the correct response (e.g., name the color in which the word is presented) even though the word itself (e.g., “red”) may automatically suggest a different response (e.g., Cohen, Dunbar, & McClelland, 1990; Logan, 1980).
C. SELF-FOCUSED ATTENTION Dijksterhuis and colleagues (Dijksterhuis, Bargh & Miedema, 2000; Dijksterhuis & van Knippenberg, 2000) investigated the potential moderating role of self-focus. Their analysis was based on the literature on action control (Norman & Shallice, 1986; Shallice, 1988) and on a vast body of research on selffocus. The literature on action control demonstrates that sometimes multiple action
30
AP DIJKSTERHUIS AND JOHN A. BARGH
tendencies are active. Under these circumstances, these various action tendencies strive for mental dominance. The one that eventually gains dominance inhibits the other action tendencies and guides overt behavior. In the experiment carried out by Macrae and Johnston (1998), for instance, one can say that the goal to hurry and the helpfulness prime both strived for dominance but that the goal eventually won, thereby inhibiting the prime. Increased self-focus, that is, increased attention to the self, is known to activate action tendencies (Carver & Scheier, 1981; Duval & Wicklund, 1972; Gibbons, 1990). Self-focus makes norms, behavioral standards, and important goals more salient and more accessible. This means that under conditions of self-focus, effects of perception on behavior may be overruled. After all, as the literature on action control suggests, activated norms or goals can inhibit other action tendencies, such as primed constructs. This hypothesis was tested in various experiments. In one study, participants were primed with the stereotype of politicians or they were not primed. Also, they were seated in front of a mirror (a manipulation known to enhance self-focus, see, e.g., Duval & Wicklund, 1972) or not. Later participants were asked to write a short essay about the French nuclear testing program. Based on the stereotype of politicians as longwinded, we hypothesized that primed participants would write longer essays. This was indeed the case. Importantly, and in line with the second hypothesis, this only happened among participants who were not seated in front of a mirror. Participants with heightened self-focus did not show an effect of the prime. This finding was replicated in a second experiment. In this experiment, participants were primed with either the stereotype of professors or the stereotype of soccer hooligans. Again, they were either seated in front of a mirror or not. After being primed, participants received a general knowledge test. As expected, participants primed with professors under no-self-focus conditions outperformed participants primed with soccer hooligans, while no priming effects were apparent under self-focus conditions. Recently, Van Baaren, de Bouter, and van Knippenberg (2001) obtained evidence showing that self-focused attention also inhibits behavioral matching of observables. In their experiment, they closely followed the procedure used by Chartrand and Bargh (1999). A participant and a confederate worked together on a task, while the confederate either engaged in foot shaking or nose rubbing. When participants worked on a task that did not alter their self-focus, the participants indeed mimicked this behavior, thereby replicating the results of Chartrand and Bargh (1999). In a different condition, however, the task the participant and the confederate engaged in was specifically designed to enhance self-focus. They were presented with a text in a foreign language (which both the confederate and the participants did not master) with omissions. The task was to guess which words were omitted and the participants could choose between I, me, or mine. This manipulation enhanced self-focus and, as predicted, no sign of behavior matching was obtained under these conditions.
THE PERCEPTION–BEHAVIOR LINK
31
In a different set of studies, additional evidence was obtained for the moderating role of self-focused attention. Dijksterhuis, Bargh, and Miedema (2000) investigated what happens when participants are told that they are primed and that the prime may influence their behavior. In their experiments, some participants were primed with the stereotype of the elderly, whereas others were not primed. Subsequently, participants were presented with a memory task. Prior to the memory task, however, some participants were told that they were primed with the elderly stereotype and that this may affect their memory performance. As may be expected on the basis of the moderating of self-focus, awareness of the potential influence of the prime eliminated the influence of the prime. That is, making people aware of the fact that their memory performance may be manipulated increases self-focus and thus overrides effects of priming. Spencer, Steele, and Quinn (1999) obtained results that may be explained by the same mechanism, that is, enhanced self-focus. They showed, in agreement with other findings in the stereotype threat domain, that female participants underperformed on a highly diagnostic math test. It is known that diagnostic tests can lead to self-stereotyping among people for whom task-related stereotypes exist. Hence, women confronted with a math test activate the stereotype of women as being poor at math, which undermines their performance. Spencer, Steele, and Quinn (1999), however, observed that women who were explicitly told that the test at hand showed no gender differences did not underperform. In other words, the effects of stereotype activation were overridden. One way to explain this finding is to assume that focusing participants on the potential relevance of gender (or stating that gender is irrelevant for this particular test, thereby implicitly stating that on other occasions it is relevant) increases self-focus and eliminates effects of stereotype activation on performance.
D. LIKING One important moderator, however, serves to increase perception–behavior effects. As noted above, when people like each other, they imitate and behavior match even more than usual (e.g., Bernieri & Rosenthal, 1991; Charney, 1966; LaFrance & Broadbent, 1976). As many studies have shown, feelings of empathy and liking are correlated with the amount of mimicry and imitation; Chartrand and Bargh (1999, Experiment 2) showed that postural mimicry causes greater liking, and their Experiment 3 showed that the more empathic an individual is, the more likely he or she is to mimic the interaction partner’s behavior. Thus the causal effect is bidirectional; greater imitation produces greater liking and rapport, and a greater degree of liking for the other person causes one to imitate and mimic more than usual. It should be noted that although the relation between liking and imitation has often been regarded as a strategic one—people want to be liked and therefore mimic more—this does not have to be the case. It is possible that the
32
AP DIJKSTERHUIS AND JOHN A. BARGH
more people like each other, the more they pay attention to each other, or, in other words, the more they look at each other. It is possible, therefore, that liking simply leads to stronger perceptual effects and to a higher activation level of the perceptual representation and therefore to more pronounced behavioral effects. This explanation is in line with findings obtained by Dijksterhuis and van Knippenberg (1998; see also Wheeler, Jarvis, & Petty, 2000), who showed that stronger priming manipulations (defined by duration of the priming manipulation) lead to stronger behavioral effects than weaker priming manipulations.
IX. Conclusions In this chapter, we reviewed findings showing that social perception automatically results in corresponding social behavior. When we see someone yawn, we start to yawn as well. When we see someone scratch his head, we do so too. When we see elderly people, we start to walk more slowly and we become a bit forgetful. These automatic forms of imitation are the consequence of the way we are wired. Perceptual representations automatically activate corresponding behavior representations. Like other species, such as fish, we automatically imitate others. As imitation is the consequence of “mere” perception, we do not need additional mechanisms to engage in imitation. No motivation is required, nor a conscious decision. We just do it. We start doing it soon after we are born (Meltzoff & Moore, 1977) and we apply our entire perceptual repertoire, ranging from simple gestures to abstract social stereotypes (Bargh, Chen, & Burrows, 1996; Dijksterhuis & van Knippenberg, 1998). On the other hand, unlike that of fish, human automatic imitation is not obligatory. We do it, yes, but the tendency to imitate can be inhibited, for instance, by important goals or by heightened self-focus. In a way, we can conceive of automatic imitation as “default social behavior.” We naturally imitate as long as some other processes do not have a reason to intervene. One question remains: Why do we do it? The fact that automatic imitation is the consequence of the way we are wired is an answer, of course, but then the question becomes why we are wired the way we are. In the Introduction of this chapter, it was already argued that natural selection works on behavior, not on perception. Selection cares not about how we perceive, but about how we behave. So somewhere along the line of our evolutionary history, imitation likely proved to be advantageous over an absence of imitation. With species such as fish and gnus, we can easily see that this is indeed the case. A fish that follows other fish or a gnu that runs away when it sees other gnus do so reduces the probability that it will be eaten by a shark or a lion. So imitation is safe as a basic, default behavioral tendency. Although this was still true for recent ancestors of human beings as well, it is harder to defend that it is still of paramount importance for human beings today (although escaping a
THE PERCEPTION–BEHAVIOR LINK
33
building merely upon seeing others do so is still better than waiting for someone to tell you there is a fire). So are there other benefits of automatic imitation that caused human beings (and maybe other higher animals) that helped the capacity for automatic imitation to stay intact or even to develop more? This is very well possible. Human beings have a fundamental “need to belong” (Baumeister & Leary, 1995; Caporael, Dawes, Orbell, & van der Kragt, 1989; Leary & Baumeister, 2000). They do not want to be the odd one out; instead, they need to be accepted and they need to be liked. As some of the evidence reviewed before shows, imitation certainly leads to greater cohesion and to greater liking. It makes our social interactions simply go more smoothly and without as many conflicts. So a system that allows automatic imitation, that is, a system that translates perception into corresponding behavior, helps us to fulfill an enormously important social need. One may object to such a functional perspective by claiming that not all the individual consequences of the perception–behavior link are functional. Becoming more stupid in the presence of soccer hooligans may be helpful, but not necessarily. Driving very fast after watching a Formula One Grand Prix is certainly not functional. However, for a mechanism to be functional, all that is needed is that the vast majority of its consequences are beneficial. Or, more precisely, for a mechanism to be functional what is needed is that the consequences in general are beneficial compared to the consequences in general of not having this mechanism. It does not imply that all individual consequences are functional. Toes have a function, but everyone can recall an unfortunate encounter with a cupboard or a stone that prompted the wish to not have toes at all. To conclude, automatic imitation is safe and it leads to social acceptance and belonging. Strange as it may sound, the author of this chapter who received a fine after driving too fast essentially did this because of a basic mechanism of mind that developed to increase safety and social acceptance. He just wanted to survive and to be liked.
Acknowledgments Preparation of this chapter was supported by a fellowship of the Royal Netherlands Academy of Sciences and by Grant R01-60767 from the U.S. Public Health Service.
References Aarts, H., & Dijksterhuis, A. (2001). Comparability is in the eye of the beholder: Contrast and assimilation effects of primed animal exemplars on person judgments. British Journal of Social Psychology (in press).
34
AP DIJKSTERHUIS AND JOHN A. BARGH
Allport, G. W. (1937). Personality: A psychological interpretation. New York: Holt. Anisfield, M. (1979). Response to Meltzoff and Moore. (1977). Science, 205, 214. Aronson, J., Lustina, M. J., Good, C., Keough, K., Steele, C. M., & Brown, J. (1999). When white men can’t do math: Necessary and sufficient factors in stereotype threat. Journal of Experimental Social Psychology, 35, 29–46. Bargh, J. A. (1994). The four horsemen of automaticity: Awareness, intention, efficiency and control in social cognition. In R. S. Wyer Jr. & T. K. Srull (Eds.), The handbook of social cognition (Vol. 2, pp. 1–40). Hillsdale, NJ: Erlbaum. Bargh, J. A. (1999). The cognitive monster. In S. Chaiken & Y. Trope (Eds.), Dual process theories in social psychology (pp. 361–382). New York: Guilford. Bargh, J. A. (2000). The psychology of the mere. In J. A. Bargh & D. Apsley (Eds.), Unraveling the complexities of social life: A festschrift in honor of Robert B. Zajonc. Washington, DC: American Psychological Association. Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of being. American Psychologist, 54, 462–476. Bargh, J. A., Chen, M., & Burrows, L. (1996). The automaticity of social behavior: Direct effects of trait concept and stereotype activation on action. Journal of Personality and Social Psychology, 71, 230–244. Bargh, J. A., & Ferguson, M. J. (2001). Beyond behaviorism: On the automaticity of higher mental processes. Psychological Bulletin (in press). Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental huam motivation, Psychological Bulletin, 117, 497–529. Bavelas, J. B., Black, A., Lemery, C. R., & Mullett, J. (1986). “I show how you feel”: Motor mimicry as a communicative act. Journal of Personality and Social Psychology, 50, 322–329. Bavelas, J. B., Black, A., Lemery, C. R., & Mullett, J. (1987). Motor mimicry as primitive empathy. In N. Eisenberg & J. Strayer (Eds.), Empathy and its development (pp. 317–338). Cambridge, UK: Cambridge Univ. Press. Berkowitz, L. (1984). Some effects of thoughts on anti- and prosocial influences of media events: A cognitive-neoassociation analysis. Psychological Bulletin, 95, 410–427. Bernieri, F. (1988). Coordinated movement and rapport in teacher–student interactions. Journal of Nonverbal Behavior, 12, 120–138. Bernieri, F., Reznick, J. S., & Rosenthal, R. (1988). Synchrony, pseudo synchrony, and dissynchrony: Measuring the entrainment process in mother–infant interactions. Journal of Personality and Social Psychology, 54, 243–253. Bernieri, F., & Rosenthal, R. (1991). Interpersonal coordination: Behavioral matching and interactional synchrony.In R. S. Feldman & B. Rime (Eds.), Fundamentals of nonverbal behavior (pp. 401–432). Cambridge, UK: Cambridge Univ. Press. Biernat, M., Manis, M., & Nelson, T. E. (1991). Stereotypes and standards of judgment. Journal of Personaility and Social Psychology, 60, 485–499. Blair, I. V., & Banaji, M. R. (1996). Automatic and controlled processes in gender stereotyping. Journal of Personality and Social Psychology, 70, 1142–1163. Bodenhausen, G. V., & Lichtenstein, M. (1987). Social stereotypes and information processing strategies: The impact of task complexity. Journal of Personality and Social Psychology, 52, 871–880. Bodenhausen, G. V., Macrae, C. N., & Sherman, J. W. (1999). On the dialectics of discrimination: Dual processes in social stereotyping. In S. Chaiken & Y. Trope (Eds.), Dual process theories in social psychology (pp. 271–292). New York: Guilford. Bock, J. K. (1986). Syntactic persistence in sentence production. Cognitive Psychology, 18, 355–387. Bock, K. (1989). Closed-class immanence in sentence production. Cognition, 31, 163–186. Breder, C. M. (1976). Fish schools as operational structures. Fishery Bulletin, 74, 471–502.
THE PERCEPTION–BEHAVIOR LINK
35
Brewer, M. B. (1988). A dual process model of impression formation. In R. S. Wyer, Jr. & T. K. Srull (Eds.), Advances in social cognition (Vol. 1, pp. 1–36). Hillsdale, NJ Erlbaum. Buytendijk, F. J. J. (1922). De wijsheid der mieren [The intelligence of ants]. Amsterdam, Meulenhoff. Capella, J. N. (1981). Mutual influence in expressive behavior: Adult–adult and infant–adult dyadic interaction. Psychological Bulletin, 89, 101–132. Caporael, L. R., Dawes, R. M., Orbell, J. M., & van der Kragt, A. (1989). Selfishness examined: Cooperation in the absence of egoistic incentives. Behavioral and Brain Sciences, 12, 683– 739. Carpenter, W. B. (1874). Principles of mental physiology. New York, Appleton. Carver, C. S., Ganellen, R. J., Froming, W. J., & Chambers, W. (1983). Modeling: An analysis in terms of category accessibility. Journal of Experimental Social Psychology, 19, 403–421. Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory approach to human behavior. New York, Springer-Verlag. Charney, E. J. (1966). Psychosomatic manifestations of rapport in psychotherapy. Psychosomatic Medicine, 28, 305–315. Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: The perception–behavior link and social interaction. Journal of Personality and Social Psychology, 76, 893–910. Chen, M., & Bargh, J. A. (1997). Nonconscious behavioral confirmation processes: The self-fulfilling nature of automatically activated stereotypes. Journal of Experimental Social Psychology, 33, 541–560. Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97, 332–361. Condon, W. S., & Ogston, W. D. (1966). Sound film analysis of normal and pathological behavior patterns. Journal of Nervous and Mental Disease, 143, 338–347. Condon, W. S., & Sander, L. W. (1974). Synchrony demonstrated between movements of the neonate and adult speech. Child Development, 45, 456–462. Decety, J., Jeannerod, M., Germain, M., & Pastene, J. (1991). Vegetative response during imagined movement is proportional to mental effort. Behavioural Brain Research, 42, 1–5. Dennett, D. C. (1995). Darwin’s dangerous idea. New York, Simon & Schuster. Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5–18. Duval, S., & Wicklund, R. A. (1972). A theory of objective self-awareness. San Diego, Academic Press. Dijksterhuis, A. (2001). Automatic social influence: The perception–behavior link as an explanatory mechanism for behavior matching. In J. Forgas (Ed.), Social Influence: The Sydney symposium of social psychology (in press). Dijksterhuis, A., Aarts, H., Bargh, J. A., & van Knippenberg, A. (2000). On the relation between associative strength and automatic behavior. Journal of Experimental Social Psychology., 36, 531–544. Dijksterhuis, A., Bargh, J. A., & Miedema, J. (2000). Of men and mackerels: Attention and automatic behavior. In H. Bless & J. P. Forgas (Eds.), Subjective experience in social cognition and behavior, (pp. 36–51), Philadelphia, Psychology Press. Dijksterhuis, A., & Corneille, O. (2001). On the relation between stereotype activation and intellectual performance. Unpublished manuscript. University of Nijmegen. Dijksterhuis, A., & Marchand, M. (2000). The route from stereotype activation to overt behavior. Manuscript in preparation, University of Nijmegen. Dijksterhuis, A., Spears, R., & Lepinasse, V. (2001). Reflecting and deflecting stereotypes: Assimilation and contrast in automatic behavior. Journal of Experimental Social Psychology (in press). Dijksterhuis, A., Spears, R., Postmes, T., Stapel, D. A., Koomen, W., van Knippenberg, A., & Scheepers, D. (1998). Seeing one thing and doing another: Contrast effects in automatic behavior. Journal of Personality and Social Psychology, 5, 862–871.
36
AP DIJKSTERHUIS AND JOHN A. BARGH
Dijksterhuis, A., & van Knippenberg, A. (1995). Memory for stereotype-consistent and stereotypeinconsistent information as a function of processing pace. European Journal of Social Psychology, 25, 689–694. Dijksterhuis, A., & van Knippenberg, A. (1996). The knife that cuts both ways: Facilitated and inhbibited access to traits as a result of stereotype activation. Journal of Experimental Social Psychology, 32, 271–288. Dijksterhuis, A., & van Knippenberg, A. (1998). The relation between perception and behavior or how to win a game of Trivial Pursuit. Journal of Personality and Social Psychology, 74, 865– 877. Dijksterhuis, A., & van Knippenberg, A. (2000). Behavioral indecision: Effects of self-focus on automatic behavior. Social Cognition, 18, 55–74. Dimberg, U. (1982). Facial reactions to facial expressions. Psychophysiology, 19, 643–647. Di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V., & Rizzolatti, G. (1992). Understanding motor events: A Neurophysiological study. Experimental Brain Research, 91, 176–180. Dovidio, J. F., Evans, N., & Tyler, R. B. (1986). Racial stereotypes: The contents of their cognitive representations. Journal of Experimental Social Psychology, 22, 22–37. Eidelberg, L. (1929). Experimenteller Beitrag zum Mechanismus der Imitationsbewegung. Jahresb¨ucher f¨ur Psychiatrie und Neurologie, 46, 170–173. Epley, N., & Gilovich, T. (1999). Just going along: Nonconscious priming and conformity to social pressure. Journal of Experimental Social Psychology, 35, 578–589. Erber, R., & Fiske, S. T. (1984). Outcome dependency and attention to inconsistent information. Journal of Personality and Social Psychology, 47, 709–726. Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation of attitudes. Journal of Personality and Social Psychology, 50, 229–238. Field, T., Woodson, R., Greenberg, R., & Cohen, D. (1982). Discrimination and imitation of facial expression by neonates. Science, 218, 179–181. 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 (Vol. 3, pp. 1–74). San Diego, CA: Academic Press. Gibbons, F. X. (1990). Self-attention and behavior: A review and theoretical update. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology, 23, 249–303. Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton–Mifflin. Gilbert, D. T. (1989). Thinking lightly about others: Automatic components of the social inference process. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 189–211). New York: Guilford. Greenwald, A. G. (1970). Sensory feedback mechanisms in performance control: With special reference to the ideomotor mechanism. Psychological Review, 77, 73–99. Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem and stereotypes. Psychological Review, 102, 4–27. Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. Cambridge, UK: Cambridge Univ. Press. Herr, P. M. (1986). Consequences of priming: Judgment and behavior. Journal of Personality and Social Psychology, 51, 1106–1115. Herr, P. M., Sherman, S. J., & Fazio, R. H. (1983). On the consequences of priming: Assimilation and contrast effects. Journal of Experimental Social Psychology, 19, 323–340. Higgins, E. T. (1989). Knowledge accessibility and activation: Subjectivity and suffereing from unconscious sources. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 75–123). New York: Guilford. Higgins, E. T. (1996). Knowledge activation: Accessibility, applicability, and salience. In E. T.
THE PERCEPTION–BEHAVIOR LINK
37
Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 133– 168). New York: Guilford. Higgins, E. T., & Bargh, J. A. (1987). Social congnition and social perception. In M. R. Rosenzweig & L. W. Porter (Eds.), Annual review of psychology (Vol. 38, pp. 369–425). Palo Alto, CA: Annual Reviews. Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility and impression formation. Journal of Experimental Social Psychology, 13, 141–154. Ingle, D. (1973). Two visual systems in a frog. Science, 181, 1053–1055. Jacobson, S. W., & Kagan, J. (1979, July 13). Response to Meltzoff and Moore (1977). Science, 205, 215. James, W. (1890). Principles of psychology. New York: Holt. Jastrow, J. (1908). The subconscious. Boston: Houghton—Mifflin. Jeannerod, M. (1994). The representing brain: Neural correlates of motor intention and imagery. Behavioral and Brain Sciences, 17, 187–245. Jeannerod, M. (1997). The cognitive neuroscience of action. Oxford: Blackwell. Kawakami, K., Young, H., & Dovidio, J. F. (2000). Automatic Stereotyping: Category, trait and behavior activations. Unpublished Manuscript. University of Nyuegen. Kendon, A. (1970). Movement coordination in social interaction: Some examples described. Acta Psychologica, 32, 1–25. K¨ohler, W. (1927). The mentality of apes (2nd ed.). New York: Harcourt. LaFrance, M. (1979). Nonverbal synchrony and rapport: Analysis by the cross-lag panel technique. Social Psychology Quarterly, 42, 66–70. LaFrance, M. (1982). Posture mirroring and rapport. In M. Davis (Ed.), Interaction rhythms: Periodicity in communicative behavior (pp. 279–298). New York: Human Sciences Press. LaFrance, M., & Broadbent, M. (1976). Group rapport: Posture sharing as a nonverbal indicator. Group and Organization studies, 1, 328–333. LaFrance, M., & Ickes, W. (1981). Posture mirroring and interactional involvement: Sex and sex typing effects. Journal of Nonverbal Behavior, 5, 139–154. Langer, E., Blank, A., & Chanowitz, B. (1978). The mindlessness of ostensibly thoughtful action: The role of ‘placebic’ information in interpersonal interaction. Journal of Personality and Social Psychology, 36, 635–642. 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: Academic Press. Lepore, L., & Brown, R. (1997). Category and stereotype activation: Is prejudice inevitable? Journal of Personality and Social Psychology, 72, 275–287. Levelt, W. J. M., & Kelter, S. (1982). Surface form and memory in question answering. Cognitive Psychology, 14, 78–106. Levy, B. (1996). Improving memory in old age through implicit self-stereotyping. Journal of Personality and Social Psychology, 71, 1092–1107. Lhermitte, F. (1983). “Utilization behaviour” and its relation to lesions of the frontal lobes. Brain, 106, 237–255. Logan, G. D. (1980). Attention and automaticity in Stroop and priming tasks: Theory and data. Cognitive Psychology, 12, 523–553. Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit thought and action: A theory of an act of control. Psychological Review, 91, 295–327. MacKay, (1981). The problem of rehearsal or mental practice. Journal of Motor behavior, 13, 274– 285. Macrae, C. N., Bodenhausen, G. V., Milne, A. B., Castelli, L., Schloerscheidt, A. M., & Greco, S. (1998). On activating exemplars. Journal of Experimental Social Psychology, 34, 330–354.
38
AP DIJKSTERHUIS AND JOHN A. BARGH
Macrae, C. N., Hewstone, M., & Griffiths, R. J. (1993). Processing load and memory for stereotypebased information. European Journal of Social Psychology, 23, 77–87. Macrae, C. N., & Johnston, L. (1998). Help, I need somebody: Automatic action and inaction. Social Cognition, 16, 400–417. Macrae, C. N., Stangor, C., & Milne, A. B. (1994). Activating social stereotypes: A functional analysis. Journal of Experimental Social Psychology, 30, 370–389. Manis, M., Nelson, T. E., & Shedler, J. (1988). Stereotypes and social judgment: Extremity, assimilation, and contrast. Journal of Personality and Social Psychology, 55, 28–36. McArthur, L. Z., & Baron, R. M. (1983). Toward an ecological theory of social perception. Psychological Review, 90, 215–238. Meltzoff, A. N., & Moore, M. K. (1977). Imitation of facial and manual gestures by human neonates. Science, 198, 75–78. Meltzoff, A. N., & Moore, M. K. (1979). Note responding to Anisfeld, Masters, and Jacobson and Kagan’s comments on Meltzoff and Moore (1977). Science, 205, 217–219. Meltzoff, A. N., & Moore, M. K. (1983). Newborn infants imitate adult facial gestures. Child development, 54, 702–709. Metcalfe, J., & Mischel, W. (1999). A Hot/Cool-System analysis of delay of gratification: Dynamics of Willpower. Psychological Review, 106, 3–19. Milgram, S. (1963). Behavioral study of obedience. Journal of Abnormal and Social Psychology, 67, 371–378. Milner, A. D., & Goodale, M. A. (1995). The visual brain in action. Oxford, UK: Oxford Univ. Press. M¨usseler, J., & Hommel, B. (1997). Blindness to response-compatible stimuli. Journal of Experimental Psychology: Human Perception and Performance, 23, 861–872. Neely, J. H. (1977). Semantic priming and retrieval from lexical memory: Roles of inhibitionless spreading activation and limited capacity attention. Journal of Experimental Psychology: General, 1, 226–254. Norman, D. A., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R. J. Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and self-regulation: Advances in research and theory (Vol. 4, pp. 1–18). New York: Plenum. Neumann, R., & Strack, F. (2000). “Mood contagion”: The automatic transfer of mood between persons. Journal of Personality and Social Psychology, 79, 211–223. O’Toole, R., & Dubin, R. (1968). Baby feeding and body sway: An experiment in George Herberts Mead’s “Taking the role of the other.” Journal of Personality and Social Psychology, 10, 59–65. Passingham, R. (1993). Oxford Psychology Series: The frontal lobes and voluntary action (No. 21), Oxford, UK: Oxford Univ. Press. Paus, T., Petrides, M., Evans, A. C., & Meyer, E. (1993). Role of human anterior cingaluate cortex in the control of oculomotor, manual and speech responses: A positron emission tomography study. Journal of Neurophysiology, 70, 453–469. Pitcher, T. J. (1979). Sensory information and the organization of behavior in a shoaling cyprinid fish. Animal Behavior, 27, 126–149. Prinz, W. (1990). A common coding approach to perception and action. In O. Neumann & W. Prinz (Eds.), Relationships between perception and action (pp. 167–201). Berlin: Springer-Verlag. Provine, R. R. (1986). Yawning as a stereotypical action pattern and releasing stimulus. Ethology, 71, 109–122. Rizzolatti, G., & Arbib, M. A. (1998). Language within our grasp. Trends in Neuroscience, 21, 188–194. Scheflen, A. E. (1964). The significance of posture in communication systems. Psychiatry, 27, 316–331. Schenkein, J. (1980). A taxonomy of repeating action sequences in natural conversation. In B. Butteworth (Ed.), Language production: Vol. 1, Speech and talk, New York, London: Academic Press.
THE PERCEPTION–BEHAVIOR LINK
39
Schwarz, N., & Bless, H. (1992). Constructing reality and its alternatives: An inclusion/exclusion model of assimilation and contrast in social judgment. In L. L. Martin & A. Tesser (Eds.), The construction of social judgments (pp. 217–245). Hillsdale, NJ: Erlbaum. Shallice, T. (1988). From neuroscience to mental structure. New York: Cambridge Univ. Press. Shih, M., Pittinsky, T. L., & Ambady, N. (1999). Stereotype susceptibility: Identity salience and shifts in quantitative performance. Psychological Science, 10, 80–83. Skinner, B. F. (1938). The behavior of organisms: An experimental analysis. New York: Appleton. Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science, 283, 1657–1661. Spencer, S., Steele, C. M., & Quinn, D. (1999). Under suspicion of inability: Stereotype threat and women’s math performance. Journal of Experimental Social Psychology, 35, 1–29. Srull, T. K., & Wyer, R. S., Jr. (1979). The role of category accessibility in the interpretation of information about persons: Some determinants and implications. Journal of Personality and Social Psychology, 37, 1660–1672. Srull, T. K., & Wyer, R. S., Jr. (1980). Category accessibility and social perception: Some implications for the study of person memory and interpersonal judgments. Journal of Personality and Social Psychology, 38, 841–856. Stapel, D. A., Koomen, W., & van der Pligt, J. (1996). The referents of traits inferences: The impact of trait concepts versus actor-trait links on subsequent judgments. Journal of Personality and Social Psychology, 70, 437–450. Stapel, D. A., Koomen, W., & van der Pligt, J. (1997). Categories of category accessibility: The impact of trait versus exemplar priming on person judgments. Journal of Experimental Social Psychology, 33, 44–76. Steele, C. M. (1997). A threat in the air. American Psychologist, 52, 613–629. Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 69, Stengel, E., Vienna, M. D., & Edin, L. R. C. P. (1947). A clininal and psychological study of echoreactions. Journal of Mental Science, 93, 598–612. Studdert-Kennedy, M. (1987). The phoneme as a perceptuomotor structure. In A. Allport, D. G. MacKay, W. Prinz & E. Scheerer (Eds.), Language perception and production, London: Academic. Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of human smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social Psychology, 54, 768–777. Tetlock, P. E. (1992). The impact of accountability on judgment and choice: Toward a social contingency model. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 331–376). San Diego, CA: Academic Press. Uleman, J. S., Newman, L. S., & Moskowitz, G. B. (1996). People as flexible interpreters: Evidence and issues from spontaneous trait inference. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 28, pp. 211–179). San Diego, CA: Academic Press. Van Baaren, R., de Bouter, C., & van Knippenberg, A. (2001). Self-focus and the chameleon effect. Manuscript in preparation. Vaughan, K. B., & Lanzetta, J. T. (1980). Vicarious instigation and conditioning of facial expressive and autonomic responses to a model’s expressive display of pain. Journal of Personality and Social Psychology, 38, 909–923. Watson, J. B. (1913). Psychology as the behaviorist views it. Psychological Review, 20, 158–177. Wheeler, L. (1966). Toward a theory of behavioral contagion. Psychological Review, 73, 179– 192. Wheeler, S. C., Jarvis, W. B. G., & Petty, R. E. (2001). Think unto others...The self-destructive impact of negative racial stereotypes. Journal of Experimental Social Psychology (in press).
40
AP DIJKSTERHUIS AND JOHN A. BARGH
Winter, L., & Uleman, J. S. (1984). When are social judgments made? Evidence for the spontaneousness of trait inferences. Journal of Personality and Social Psychology, 47, 237–252. Zajonc, R. B., Adelmann, K. A., Murphy, S. T., & Niedenthal, P. M. (1987). Convergence in the physical appearance of spouses. Motivation and Emotion, 11, 335–346. Zajonc, R. B., Pietromonaco, P., & Bargh, J. A. (1982). Independence and interaction of affect and cognition. In M. S. Clark & S. T. Fiske (Eds.), Affect and cognition: The seventeenth annual Carnegie symposium on cognition (pp. 211–227). Hillsdale, NJ: Erlbaum.
A DUAL-PROCESS COGNITIVE-MOTIVATIONAL THEORY OF IDEOLOGY AND PREJUDICE
John Duckitt
I. Introduction Over the past 4 decades prejudice has been primarily studied as a group or socially shared phenomenon. Important perspectives have been Pettigrew’s (1958, 1960) sociocultural approach; realistic conflict theory (LeVine & Campbell, 1972); and, more recently, social identity and self-categorization theories (Tajfel & Turner, 1979; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). However, prejudice can also be viewed as an individual phenomenon in the sense that individuals seem to differ in their propensity to adopt prejudiced and ethnocentric attitudes. The view of prejudice as an individual phenomenon has an important empirical basis: Prejudice tends to be generalized over targets. Persons who are less favorable to one out-group or minority tend to be less favorable to other out-groups or minorities. This has been documented empirically by strong positive correlations between attitudes to different out-groups. Such findings have been reported for diverse samples and target groups (Adorno, Frenkel-Brunswick, Levinson, & Sanford, 1950; Altemeyer, 1981, 1988; Bierly, 1985; Glock, Wuthnow, Piliavin, & Spencer, 1975; Prothro & Jensen, 1950; Ray & Lovejoy, 1986). The effects have typically been quite powerful, with correlations usually around .50 (Duckitt, 1992a), and have even been apparent for social distance to completely fictitious groups (Fink, 1971; Hartley, 1946).
ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 33
41
C 2001 by Academic Press, Inc. Copyright All rights of reproduction in any form reserved. 0065-2601/01 $35.00
42
JOHN DUCKITT
II. Prejudice and Personality A. THE AUTHORITARIAN PERSONALITY The finding of the generality of prejudice has been interpreted as suggesting that some stable attribute of individuals, such as personality or enduring beliefs, causes them to hold prejudiced and ethnocentric attitudes in general. This idea was the basis of Adorno et al. ’s (1950) classic theory of the authoritarian personality. Their findings indicated that generalized prejudice, ethnocentric in-group glorification, politicoeconomic conservatism, and profascist attitudes covaried powerfully to form a general attitudinal syndrome. They viewed this attitudinal syndrome as an expression of a basic personality dimension, consisting of nine covarying traits, which they termed the authoritarian personality and which they attempted to measure with their F scale. Initially their theory of the authoritarian personality and the F scale attracted enormous interest. However, by the early 1960s interest in this perspective had largely collapsed because of its numerous weaknesses. Particularly critical in this respect were the psychometric flaws of the F scale, which largely derived from its lack of reliability and unidimensionality when acquiescent bias due to the all positive formulation of its items was controlled. As Altemeyer (1981) later showed, it appeared to be measuring several poorly related factors. A second weakness of the F scale attracted much less attention. Whereas the F scale was widely viewed as measuring an authoritarian personality, its items did not pertain to behavioral consistencies or personality trait terms as those of personality scales typically do. Instead, as Stone, Lederer, and Christie (1993, p. 232) point out in their review of the theory, and as other commentators have periodically noted (Feldman & Stenner, 1997; Goertzel, 1987; Verkuyten & Hagendoorn, 1998), the items of the F scale consist of statements of social attitude and belief of a broadly ideological nature. Adorno et al. (1950) seem to have assumed that these statements of ideological beliefs were expressive of underlying personality traits or dispositions, but this was never directly demonstrated. After 2 decades of obscurity, the idea of an authoritarian personality was revived by Bob Altemeyer in 1981. His research suggested that only three of the original nine facets of authoritarianism described by Adorno et al. (1950)—conventionalism, authoritarian aggression, and authoritarian submission—covaried strongly to form a unitary social attitude dimension, and he developed his Right-wing Authoritarianism (RWA) Scale to measure this dimension. Subsequent research by Altemeyer (1981, 1988, 1998) and others (cf. Stone, Lederer, & Christie, 1993) has shown that the RWA scale is a unidimensional and reliable psychometric measure of authoritarianism. It powerfully predicts a wide range of political, social, ideological, and intergroup phenomena as well as generalized prejudice to out-groups and minorities and chauvinistic ethnocentrism.
IDEOLOGY AND PREJUDICE
43
Altemeyer’s work has been important in reestablishing the idea of an authoritarian personality as a central explanatory construct in social and political psychology. However, like the earlier F scale, the RWA scale consisted of items that did not refer to behavior or personality traits, but were statements of social attitude or belief of a broadly ideological nature (see also, e.g., Duckitt, 1989; Feldman & Stenner, 1997; Goertzel, 1987; Saucier, 2000; Stone, Lederer, & Christie, 1993, p. 232). Once again, it was assumed that these relatively stable and enduring social attitudes and beliefs reflected an underlying dimension of personality, but this personality dimension was never directly measured and was merely inferred. B. SOCIAL DOMINANCE ORIENTATION During the 1990s an important new perspective on group conflict and ethnocentrism, social dominance theory, was developed (Pratto, Sidanius, Stallworth, & Malle, 1994; Sidanius & Pratto, 1993). Sidanius and Pratto suggested that societies minimize group conflict by promoting consensual ideologies that legitimize social and intergroup inequality and discrimination. Their theory centered on an individual difference dimension, social dominance orientation (SDO), measured by their SDO scale. They described SDO as a “general attitudinal orientation toward intergroup relations, reflecting whether one generally prefers such relations to be equal, versus hierarchical” and the “extent to which one desires that one’s ingroup dominate and be superior to outgroups” (Pratto et al., 1994, p. 742). Research with the SDO scale has shown that it powerfully predicts a range of sociopolitical and intergroup phenomena very similar to that predicted by the RWA scale. Both scales have been particularly powerful predictors of generalized prejudice and ethnocentrism (Altemeyer, 1998; Pratto et al., 1994; Sidanius, Pratto, & Bobo, 1994). Despite this, however, evidence from several sources indicates that the two scales are largely independent of each other. First, the two scales have shown different patterns of correlation with other social and personal characteristics. Thus, Altemeyer (1998) has pointed out the following: Most notably, high SDOs are not particularly religious, but high RWAs usually are. Similarly, high scorers on the SDO scale do not claim to be benevolent, but high RWAs do. In contrast, social dominators have a wisp of hedonism about them, but authoritarians disavow such. The former do not need structure nor value conformity and traditions, but the latter do. Social dominators tend to be men; right-wing authoritarians do not. And quite strikingly, high SDOs do not see the world being nearly as dangerous as authoritarians do, nor do they appear to be nearly as self-righteous. . . . (Altemeyer, 1998, p. 61)
44
JOHN DUCKITT TABLE I CORRELATIONS BETWEEN THE RIGHT-WING AUTHORITARIANISM AND SOCIAL DOMINANCE SCALES FROM FOUR STUDIES Study Altemeyer (1998) September (1996) October (1996) November (1996) October (1995) October (1996) January (1997)
Sample
Correlation
354 students 116 students 362 students 501 parents 482 parents 331 parents
.22∗∗ .08 .11∗ .28∗∗ .18∗∗ .24∗∗
McFarland and Adelson (1996)
478 students 297 parents
.07 .21∗∗
McFarland (1998)
175 students
.13
Pratto et al. (1994)
96 students
.14
∗ p < .05. ∗∗ p < .01.
Second, a number of studies have shown that the RWA and SDO scales predict generalized prejudice independently of each other and together account for a very substantial proportion of the variance in prejudice (e.g., this ranged from 48 to 58% for Altemeyer’s six studies) (Altemeyer, 1998; McFarland, 1998; McFarland & Adelson, 1996). A large number of other measures of personality, social attitudes, or values were included in these studies, but added only trivial amounts of explained variance in generalized prejudice to that accounted for by RWA and SDO. And third, the correlations reported between the RWA and SDO scales from a series of studies using large student and parent samples, though invariably positive, have often been nonsignificant and, when significant, relatively weak (Altemeyer, 1998; McFarland, 1998; McFarland & Adelson, 1996; Pratto et al., 1994). The correlations between RWA and SDO reported in these studies are summarized in Table I and suggest an interesting qualification to their relative independence. Although the correlations for student samples were very weak and mostly nonsignificant, those for parents were somewhat stronger and all significant. A meta-analytic aggregation of these correlations produced an average correlation of .23 (n = 1611) for the parent samples and .125 (n = 1581) for the student samples, with the difference between the two significant ( p = .002). A possible explanation for this apparent age effect could be that the social attitudes expressed in the items of the RWA and SDO scales may be initially acquired relatively independently during socialization, but later come to influence each other over time. This seems plausible because in most western societies politics tends to be organized around a single broad left–right dichotomy with high
IDEOLOGY AND PREJUDICE
45
RWA and SDO associated with the political right and low RWA and SDO with the left. Thus, with political involvement and socialization in late adolescence, the prior acquisition of either authoritarian or social dominance beliefs should generate increasing psychological consistency pressures to adopt the other. This reasoning predicts reciprocal causal effects of RWA and SDO on each other, at least from mid to late adolescence onward. This reasoning would also suggest that RWA and SDO might be more highly correlated in more strongly ideologized sociopolitical systems. For example, this should be the case in societies where party politics is more sharply polarized along a single explicitly ideologically articulated left–right dimension, as seems historically to have been the case in Western Europe, with its powerful socialist labor movements, as opposed to the United States. Consistent with this, Pratto (personal communication, 2/16/00) has found much stronger correlations between SDO and RWA in Germany than have been obtained in the United States. To sum up, therefore, it seems that two relatively orthogonal to moderately correlated social attitude dimensions independently predict a substantial proportion of the variance in generalized prejudice and ethnocentrism. Altemeyer (1998) has noted that the RWA and SDO scales both seem to relate to different sets of the original nine “trait” clusters listed by Adorno et al. (1950) as descriptive of the authoritarian personality. He therefore suggested that the RWA and SDO scales were measuring two distinct kinds of authoritarian personality: the “submissive” and the “dominant.” Altemeyer’s suggestion, however, overlooks the point noted here, and by others (Feldman & Stenner, 1997; Goertzel, 1987; Saucier, 2000; Stone, Lederer, & Christie, 1993, p. 232), that the items of both the RWA and the F scale do not pertain to personality traits or behavior, but to social attitudes and beliefs of a broadly ideological nature. The items of the SDO scale also consist of statements of social attitude and belief, and Pratto et al. (1994) have usually described the SDO scale as a measure of enduring beliefs. This is clearly illustrated by four items randomly selected from each of the scales, which are shown in Table II. Both the RWA and SDO scales therefore would seem more appropriately viewed as measuring social attitude or ideological belief dimensions rather than personality. Whereas these social attitude or ideological belief dimensions may well reflect distinct underlying personality dimensions, this has never been demonstrated and remains merely a plausible inference. This means that predicting prejudice from the RWA and SDO scales is essentially predicting one set of attitudes (intergroup) from another (social and ideological). It also means that the question of what, if any, personality dimensions may underlie prejudice has not yet been answered. Indeed it suggests that the questions could be profitably broadened to that of what underlying psychological dynamics in individuals, such as personality, might underlie both prejudice and the two ideological belief dimensions of authoritarianism and social dominance.
46
JOHN DUCKITT TABLE II ILLUSTRATIVE ITEMS FROM THE RWA AND SDO SCALES Four SDO items All groups should be given an equal chance in life. It’s OK if some groups have more of a chance in life than others. Inferior groups should stay in their place. We should strive to make incomes as equal as possible. Four RWA items A “woman’s place” should be wherever she wants to be. The days when women are submissive to their husbands and social conventions belongs strictly to the past. What our country really needs is a strong determined leader who will crush evil, and take us back to our true path. We should treat protestors and radicals with open arms and open minds, since new ideas are the lifeblood of progressive change. The facts on crime, sexual immorality, and the recent public disorders all show we have to crack down harder on deviant groups and troublemakers, if we are going to save our moral standards and preserve law and order.
C. SOCIOPOLITICAL ATTITUDES AND THE TWO DIMENSIONS The view of SDO and RWA as ideological belief or social attitude dimensions suggests that similar dimensions should have emerged in the research literatures investigating sociopolitical attitudes and sociocultural values. This has been so in both cases. Many investigations, summarized in Section A of Table III, have reported that sociopolitical attitudes and values appear to be organized around two relatively orthogonal dimensions. One dimension has been labeled authoritarianism, social conservatism, or traditionalism, at one pole, versus openness, autonomy, liberalism, or personal freedom at the other pole. The second dimension has been labeled economic conservatism, power, or belief in hierarchy or inequality at its one pole versus egalitarianism, humanitarianism, social welfare, or concern at its other pole. Thus, Eysenck (1954) found two orthogonal social attitude factors of radicalism versus conservatism and toughness versus tenderness, with the latter more appropriately relabeled humane versus inhumane attitudes by Brown (1965). A number of investigations have found orthogonal dimensions of social conservatism and economic conservatism (with support for inequality the major component of the latter) (e.g., Hughes, 1975; Middendorp, 1991; Vollebergh, Iedema, & Meeus, 1999). Saucier’s (2000) recent comprehensive study of the structure of social attitudes found that the two most powerful factors, which were orthogonal to each other, were defined by authoritarianism–conservatism (termed “alphaisms”) and social dominance–Machiavellianism (termed “betaisms”). Not surprisingly, therefore, measures of social conservatism have typically been highly correlated with
47
IDEOLOGY AND PREJUDICE TABLE III RELATIVELY ORTHOGONAL DIMENSIONS EMERGING FROM INVESTIGATIONS OF SOCIAL, POLITICAL, AND CULTURAL VALUES AND ATTITUDES Study A. Sociopolitical attitudes Hughes (1975) Middendorp (1991) Katz and Hass (1988)
Forsyth (1980) Eysenck (1954)
Kerlinger (1984)
Tomkins (1964) Rokeach (1973) Braithwaite (1994) Saucier (2000) B. Cultural values Schwartz (1996) Hofstede (1980) Triandis and Gelfand (1998) Trompenaars (1993)
RWA-like
Social Conservatism vs Liberalism Cultural Conservatism vs Openness High vs low Protestant Ethic Low vs high Relativism (tolerance) Conservatism vs Liberalism Conservatism (mainly high vs low Protestant ethic/traditionalism) Normativism (high vs low conservatism) Low vs high Freedom High vs low National Strength and Order Alphaisms (conservatism and authoritarianism) Conservatism vs Openness Collectivism vs Individualism Collectivism vs Individualism Group loyalty vs Individualism
SDO-like
Economic Conservatism vs social welfare Economic Conservatism vs equality Low versus high Humanitarianism– Egalitarianism Low vs high Idealism (altruism/ social concern) Tough vs tender (humane vs inhumane social attitudes: Brown, 1965) Liberalism (mainly low vs high egalitarian/humanitarianism) Humanism (low vs high humanitarianism) Low vs high Equality Low vs high International Harmony and Equality Betaisms (SDO and Machiavellianism) Power vs Egalitarianism High vs low Power Distance Vertical vs Horizontal Values Hierarchy vs Egalitarianism
the RWA or F scales (Duckitt, 1993; McHosky, 1996; Ray, 1985b) and scaled together with authoritarianism measures to form a single factor (Saucier, 2000; Raden, 1999), suggesting that the two constructs may be essentially isomorphic. For example, the correlation between conservatism and authoritarianism was .77 in Saucier’s (2000) study. Many other investigations of sociopolitical attitudes and values have obtained similar findings. Thus, Kerlinger’s (1984) study of sociopolitical critical referents obtained orthogonal factors of conservatism (mainly Protestant ethic and traditionalism) and liberalism (mainly egalitarian and humanitarian referents). Tomkins (1964) similarly found orthogonal ideological belief dimensions of normativism and humanism. Katz and Hass (1988) found that socially conservative Protestant
48
JOHN DUCKITT
ethic beliefs were orthogonal to a humanitarianism–egalitarianism belief dimension. Forsyth (1980), studying ethical ideologies, found two orthogonal dimensions of ethical ideology. One dimension was labeled relativism (endorsement of a relativistic as opposed to an absolutist moral code) and the second was labeled idealism (concern for the well-being of others and avoiding harm to them). RWA and conservatism correlated very strongly with the relativism dimension but not at all with idealism (McHosky, 1996). Rokeach’s (1973) classic investigation of values obtained two orthogonal terminal value dimensions, freedom and equality. Braithwaite (1994) found two independent sociopolitical value dimensions of national strength and order on the one hand and international harmony and equality on the other hand. Finally, Ray (1985a) found that the ideological dimension of radical egalitarian–humanism was orthogonal to authoritarianism. There have also been investigations of the structure of sociopolitical attitudes that on secondary analysis of their primary dimensions have found a single broadly bipolar left–right dimension rather than orthogonality (e.g., Sidanius & Ekehammar, 1976; Wilson, 1970). However, this is not entirely inconsistent with the findings of orthogonality because the orthogonality would tend to be relative and vary according to samples and contexts. Thus, evidence has already been noted suggesting that RWA and SDO seem to be virtually uncorrelated in younger student samples (at least in North America) but are significantly more strongly correlated in older adult North American samples and may also be quite strongly correlated in more highly ideologized sociopolitical systems. Although these studies of the structure of sociopolitical attitudes have used a variety of measures of the two basic dimensions that have appeared, RWA and SDO seem to have been the most powerful and consistent predictors of political– ideological and intergroup phenomena such as prejudice and ethnocentrism. This has been clearly demonstrated by several recent studies which compared the predictive power of these scales with a large number of other social attitude and value measures over a number of samples (Altemeyer, 1998, pp. 55–60; McFarland, 1998; McFarland & Adelson, 1996). This may be because both are psychometrically highly reliable and unidimensional measures. In contrast, most of the other factors or measures listed in Table III have not been particularly reliable and have frequently been contaminated by response biases (with direction of wording influencing the factors as well as content; see, e.g., Katz & Hass’s, 1988, dimensions).
D. SOCIOCULTURAL VALUES AND THE TWO DIMENSIONS Research investigating the structure of sociocultural values has also obtained two central value dimensions that directly parallel the two social attitude or ideological
IDEOLOGY AND PREJUDICE
49
belief dimensions of RWA and SDO, with remarkably convergent findings from a number of different investigations (Fiske, Kitayama, Markus, & Nisbett, 1998). These are summarized in Section B of Table III. The most methodologically sophisticated investigation has been that by Schwartz (1992, 1996), whose data from 54 countries and 44,000 participants indicated a circular ordering of value types around two basic dimensions: conservatism (conformity and tradition) versus openness (hedonism and self-direction) and self-enhancement (hierarchy or power) versus self-transcendence (egalitarianism or social concern). These two pivotal dimensions were found to describe both individual differences within cultures or societies and country differences—though the actual structuring of value types was simpler at the societal or global level (with 7 value types for societies as opposed to 10 for individuals). There is a clear conceptual similarity between Schwartz’s two pivotal dimensions and the two “authoritarian” dimensions of RWA and SDO. This has been empirically confirmed by Altemeyer’s (1998) findings showing that Schwartz’s two conservatism value types of traditionalism and conformity correlated .51 and .40 with RWA, but nonsignificantly with SDO, whereas the two opposing power and egalitarianism values correlated .43 and −.31 respectively with SDO, but nonsignificantly with RWA. Similar dimensions have emerged from other investigations of cultural values. Hofstede’s (1980) earlier investigation using only country-level data identified dimensions with close similarities to Schwartz’s. The first two factors emerging in his research were individualism–collectivism (corresponding to Schwartz’s conservatism versus openness and RWA) and power distance (the degree to which social inequality and hierarchy are accepted), which corresponds with Schwartz’s power versus egalitarianism and SDO. Trompenaars and his coworkers (Smith, Dugan, & Trompenaars, 1996; Trompenaars, 1993) used data from organizational employees from 43 countries. Their primary dimensions were quite similar to Schwartz and Hofstede’s two primary dimensions, with the first being egalitarian commitment versus power and hierarchy (cf. power distance and SDO) and the second comprising a utilitarian orientation to others versus loyalty to one’s ingroups (cf. individualism versus collectivism or conservatism versus openness and RWA). And finally, Triandis and Gelfand (1998) have reported evidence indicating that cultural values seem to be organized in terms of the two orthogonal dimensions of individualist versus collectivist and vertical (hierarchy) versus horizontal (equality) values. Collectively these findings suggest that the two dimensions measured by the RWA and SDO scales seem to represent two basic sociopolitical and sociocultural attitude–value–belief dimensions. These dimensions seem to powerfully structure social, political, and intergroup behavior and to have broad cross-cultural validity. This leads to the question, “What could be the psychological bases of such culturally general value–attitude ideological dimensions?”
50
JOHN DUCKITT
III. A Theoretical Model of Personality, Ideology, and Prejudice A. THE TWO DIMENSIONS AS MOTIVATIONAL GOAL-SCHEMAS Schwartz (1996) has argued that his value dimensions are universal because they directly express basic motivational goals that relate to universal requirements of human existence. This suggests, for example, that the authoritarian–conservatism dimension measured by the RWA scale can be seen as the expression in values and social attitudes of the opposing motivational goals of social control, security, and conformity versus autonomy and individual freedom. The power–inequality– social dominance dimension measured by the SDO scale can be seen as expressing the opposing motivational goals of superiority, dominance, or power over others versus egalitarian and altruistic social concern for others. The intriguing causal question this raises is why one such motivational goal, rather than its opposite, becomes salient for people. What makes the motivational goal of social control and security salient rather than that of personal autonomy? Two psychological anthropologists, Roy D’Andrade (1992) and Claudia Strauss (1992), have proposed a cultural analysis of motives that suggests an answer in terms of the construct of “motivational goal-schema.” Their analysis proposes that motivational goals are activated or made salient for individuals by schema-driven perceptions or interpretations of reality. Such schemas will typically have been made highly accessible for individuals through cultural socialization. Such chronically accessible schemas about the social world will lead to stable interpretations of, or beliefs about, social reality, which Ross (1993) has termed worldviews and which have the capacity to activate motivational goals and instigate action. This analysis can be applied to the two dimensions measured by the RWA and SDO scales. It suggests that the two opposing sets of motivational goals expressed in these two ideological belief or social attitude and value dimensions may be activated by the chronic accessibility of corresponding sociocultural schemas that generate particular social worldviews or stable beliefs about the nature of the social world. There are clear indications from the empirical literature of what kinds of opposing schemas or social worldviews seem to underlie these two dimensions. Thus, strong relationships have been found between indicators of social threat and authoritarianism (e.g., Altemeyer, 1988; Doty, Peterson, & Winter, 1991; Peterson, Doty, & Winter, 1993; Sales, 1973; Sales & Friend, 1973). Most notably, Altemeyer (1988) has reported powerful positive correlations between his RWA scale and the perception of the social world as dangerous and threatening (measured by his Belief in a Dangerous World Scale). A view of the world as dangerous, unpredictable, and threatening, arising from a threat schema made highly accessible through socialization, would activate the motivational goal of social control and security. This motivational goal would be expressed in the collectivist sociocultural values of conformity and traditionalism and in the authoritarian social
IDEOLOGY AND PREJUDICE
51
attitudes of the RWA scale (e.g., the F and RWA scale item: “Obedience and respect for authority are the most important virtues children should learn”). The opposite view of the social world as safe, stable, and secure, on the other hand, would derive from a social schema of security. This security schema would have been made highly accessible through socialization and would activate the opposing nonauthoritarian motivational goal of personal freedom or autonomy, which would be expressed in individualistic sociocultural values and nonauthoritarian social beliefs and attitudes (e.g., the contrait RWA item: “Everyone should have their own lifestyle, religious beliefs, and sexual preferences, even if it makes them different from everyone else”). In terms of D’Andrade’s (1992) goal-schema construct, these could be termed the opposing motivational goal-schemas of threat– control versus security–autonomy. What kind of opposing worldviews would trigger the motivational goals of superiority–power–dominance versus altruistic social concern? A valuable clue comes from two highly correlated scales (Personal Power, Meanness, and Dominance, or PP-MAD, and Exploitive Manipulative Amoral Dishonesty, or EE-MAD) constructed by Altemeyer (1998), which correlated powerfully with SDO but not with RWA. These two scales, like the Machiavellianism scale from which many of these items were drawn, seem conceptually heterogeneous because some items seem to relate to personality and behavior, whereas most express social attitudes and beliefs. A number of the social belief items seem to directly express a belief that the social world is a competitive jungle characterized by a ruthless and amoral Darwinian struggle for survival (e.g., “It’s a dog eat dog world where you have to be ruthless at times”). Sidanius, Pratto, and Bobo (1994), in using a short alternative ad hoc version of the SDO scale, also found that items expressing a ruthlessly competitive view of the world (“Winning is more important than how the game is played”) scaled together with more typical SDO scale items expressing a belief in social inequality. It is interesting that such a competitive-jungle worldview also permeated Hitler’s speeches and writings. For example, in a speech at Kulmbach in 1928: The idea of struggle is as old as life itself, for life is only preserved because other living things perish through struggle . . . In this struggle, the stronger, the more able, win, while the less able, the weak, lose . . . It is not by the principles of humanity that man lives or is able to preserve himself above the animal world, but solely by the means of the most brutal struggle. (cited in Bullock, 1962, p. 36) This schema-based view of the world as a competitive jungle characterized by a ruthless, amoral struggle for resources and power in which might is right, and winning everything, seems likely to unleash the motivational goal of seeking power, superiority, and dominance over others. This goal would be expressed in the
52
JOHN DUCKITT
sociocultural values of power, dominance, and hierarchy and in the social attitudes expressed in high scores on the SDO scale: a belief in social inequality and a desire that one’s group be superior and dominate others (Pratto et al., 1994). The direct opposite would be a schema-based view of the world as one of cooperative harmony in which people care for, help, and share with each other. This would trigger the altruistic motivational goals of helping and valuing others and sharing with them as equals, which would be expressed in the social attitudes and values of egalitarianism and humanitarianism and reflected in low scores on the SDO scale.
B. SOCIALIZATION, PERSONALITY, AND WORLDVIEW D’Andrade (1992) and Strauss (1992) suggested that it is cultural socialization that makes particular schemas highly accessible and so generate stable social worldviews. However, they do not address the question of exactly how cultural socialization does this. One possibility is that people’s social worldviews are simply learned directly from others or from their experiences with the social world. This possibility is considered in more detail later. A second possibility, which would presumably operate simultaneously, is that certain culturally favored socialization practices might produce personality-based dispositional tendencies in individuals to interpret and respond to the social world in particular ways. Some important cross-cultural evidence does suggest that childhood socialization practices may produce personality dispositions that could lead to prejudice and ethnocentrism. Ross (1993) investigated the correlates of internal conflict and external warfare in 90 preindustrial societies. Internal and external conflict were highly associated with each other, and both were significantly and independently predicted by the degree to which two factorially distinct dimensions of childhood socialization practices were characteristic of those societies. One dimension was characterized by punitive, strict, harsh, versus permissive, tolerant, indulgent socialization. Interestingly, this socialization practice dimension is very similar to that which Adorno et al. (1950) suggested would produce authoritarian personalities. Ross’s second socialization practice dimension was characterized by unaffectionate versus affectionate socialization practices (i.e., the degree to which socialization practices lack or involve affection, generosity, and valuing others and children and emphasize honesty and trust toward others). Ross (1993) suggested that these two socialization practice dimensions create dispositions in people that lead them to interpret their social world in ways that are conducive to ethnocentrism and social conflict. This finding and reasoning suggest a hypothetical causal sequence. First, childhood socialization practices produce or favor particular personality dispositions. Second, these personality dispositions facilitate the adoption of particular social worldviews. Third, both these personality dispositions and their associated social worldviews activate particular
IDEOLOGY AND PREJUDICE
53
TABLE IV TWO SETS OF PSYCHOLOGICAL DIMENSIONS UNDERLYING PREJUDICE AND ETHNOCENTRISM, WITH CAUSALITY FROM LEFT TO RIGHT Socialization Punitive vs Tolerant
Personality Conforming vs Autonomous
Worldview
Motivational goal
Threatening/ dangerous vs Safe/secure
Social control/ security vs Personal freedom
Ideological beliefs
Authoritarian/ conservative vs Autonomy/ openness Unaffectionate vs Tough-minded vs Competitive Superiority and Social Affectionate Tender-minded jungle vs dominance vs dominance vs Cooperative– Altruistic concern Egalitarian– harmony humanism
motivational goals. And fourth, these motivational goals are expressed in social value–attitude–belief dimensions such as RWA and SDO and impact through them on generalized prejudice and ethnocentrism. Ross (1993) did not address the issue of what kind of personality trait dimensions might be causally conditioned by the two socialization practice dimensions identified in his research. However, it is not difficult to hypothetically extrapolate what kind of personalities these socialization dimensions should produce and how these personalities might causally impact on social worldviews and ideological beliefs. The analysis is summarized in Table IV. Thus, punitive and strict as opposed to permissive and tolerant socialization seems likely to produce personalities high on social conformity as opposed to individual autonomy. Social conformity would tend to be associated with a strong attachment to conventional society and intolerance of deviance or nonconformity, which should create heightened sensitivity to threats to established society and therefore a greater tendency to see the social world as dangerous and threatening. Both the personality disposition of social conformity and the view of the social world as dangerous and threatening would activate the motivational goal of social control and security, which would be expressed in authoritarian and conservative social attitudes and values. At the other pole of this socialization practice dimension, permissive and tolerant socialization practices should produce personalities high on autonomy, who would be likely to see the social world as safe and secure. This social worldview would activate the motivational goal of seeking personal freedom and would be expressed in nonauthoritarian and individualistic social attitudes and values. Although no prior research has directly investigated social conformity as a behavioral or personality dimension and RWA, the link seems plausible because the items of the RWA scale express proconformity attitudes and beliefs. A link between punitive childhood socialization and authoritarian attitudes was originally proposed by Adorno et al. (1950) and has been quite extensively investigated.
54
JOHN DUCKITT
The correlations obtained have generally been positive but weak and quite often nonsignificant (e.g., Altemeyer, 1981; Milburn, Conrad, Sala, & Carberry, 1995; see also the review by Duckitt, 1992a, pp. 200–202). This would not be inconsistent with the hypothesis here because the link proposed here is one between childhood socialization and social conformity as the personality substrate of authoritarian attitudes. Any effects of socialization on authoritarian attitudes might then be indirect and mediated through personality and could therefore be relatively weak. The second socialization dimension, unaffectionate versus affectionate socialization practices, seems likely to produce personalities who are interpersonally hard, tough, ruthless, unfeeling, and cynical; that is, tough- as opposed to tendermindedness. Such personalities seem likely to be dispositionally inclined to view the social world as a competitive jungle. Both tough-mindedness and a competitivejungle worldview seem likely to make salient the motivational goal of power and dominance and to be expressed in social dominance attitudes and values of hierarchy and inequality. At the other pole of this socialization practice dimension, affectionate socialization practices should produce personalities high on empathy, compassion, generosity, and caring for others (tender-mindedness), who would be inclined to a view of the social world as a place of cooperative harmony. Viewing the social world as a place of cooperative harmony should make salient the motivational goal of social concern and altruism to others, which in turn would be expressed in low social dominance attitudes and values (egalitarian and humanitarian). There is some evidence from previous research for a link between a personality dimension of tough-mindedness and the ideological attitude dimension of social dominance. Both Goertzel (1987) and Eysenck (1954) have previously provided evidence that the personality dimension of tough- versus tender-mindedness seems to be an important predictor of sociopolitical attitudes. The personality-relevant items contained in Altemeyer’s (1998) PP-MAD scale, which correlated strongly with SDO but not with RWA, tend to express an extreme tough-mindedness and rejection of tender-mindedness and scale together with items expressing a view of the world as a competitive jungle. In addition, a lack of empathy and concern for others has been found to be related to both prejudice (McFarland, 1998) and the SDO scale (Pratto et al., 1994). In order to more directly test the links hypothesized here between these two personality and ideological attitude dimensions, three small pilot investigations were conducted during 1997. These are described in the next section.
C. STUDY 1: PERSONALITY AND IDEOLOGICAL ATTITUDES 1. Samples and Measures of RWA and SDO Three samples of undergraduate students at McGill University in Montreal, Canada (n = 56); University College London in England (n = 53); and the
IDEOLOGY AND PREJUDICE
55
University of Auckland in New Zealand (n = 51) completed the 14-item version of the SDO scale and shortened 14-item versions of the RWA scale (Duckitt, 1993). Unfortunately, a page was omitted during the printing of the questionnaire used for the Auckland sample so that this sample completed only 4 RWA items. However, this 4-item RWA scale proved reliable enough (␣ = .54) to be usable. The reliabilities for the 14-item RWA and SDO scales were much higher, ranging from .78 to .87. 2. Measuring Social Conformity as a Personality Trait Dimension In order to measure social conformity, a personality trait rating scale developed by Saucier (1994) was adapted. Saucier developed a set of personality measures in a series of large-sample multimethod multireplication studies using a large representative set of familiar personality-trait-descriptive adjectives which produced a set of factorially clean and nonevaluative personality trait scales balanced against acquiescence and with high levels of internal consistency. One of his scales, which he named “norm orientation,” provided a direct measure of social conformity as a personality trait dimension. Testing Saucier’s items for the present research, however, revealed two problems. First, 2 items tended to be unfamiliar to many students and had high nonresponse rates (“reverent” and “unmoralizing”) and were therefore discarded. Second, 5 other items seemed able to be construed as attitudinal rather than personality trait descriptors (conservative, liberal, traditional, religious, and nonreligious). Because this could build in content overlap between personality and attitudinal measures, these items were also discarded and not used in any of the analyses reported in this chapter. This left 13 of Saucier’s personality trait adjectives, with 6 protrait and 7 contrait. To balance the scale, one further protrait item was added that had performed well in item testing showing high item–total correlations with the other items (“respectful”). These 14 items were used in the first pilot study (Montreal) and are listed in Table V (items 1–14). Six further items (items 15–20 in Table V) were not used by Saucier but showed satisfactory item–total correlations with his scale during preliminary item testing and were included with 10 of Saucier’s items (items 1–10) in the next two pilot studies (London and Auckland). These 14- and 16-item social conformity scales showed high levels of internal consistency, with ␣ coefficients ranging from .84 to .87 in the three pilot studies (see Table VI). The instruction for both this scale and the Tough- versus Tender-mindedness scale, which follows, asked participants to rate “the extent to which you feel each of the following descriptive adjectives is characteristic or uncharacteristic of YOUR PERSONALITY AND BEHAVIOUR” (underlining and capitals in the original). Nine-point response scales ranging from +4 (most characteristic/very strongly agree) to −4 (most uncharacteristic/very strongly disagree) were used for both personality and attitude items.
56
JOHN DUCKITT TABLE V ITEMS MEASURING THE TWO PERSONALITY DIMENSIONS Social conformity
Tough-mindedness
1. Rebellious 2. Unorthodox 3. Conforming 4. Conventional 5. Old-fashioned 6. Free-living 7. Non-conforming 8. Moralisitc 9. Obedient 10. Unconventional 11. Unpredictable 12. Erratic 13. Respectful 14. Predictable 15. Experimenting 16. Innovative 17. Individualistic 18. Respectable 19. Law-abiding 20. Compliant
1. Kind 2. Compassionate 3. Ruthless 4. Cynical 5. Tough-minded 6. Tender-minded 7. Forgiving 8. Hard 9. Caring 10. Giving 11. Merciless 12. Gentle 13. Hard-hearted 14. Unfeeling 15. Soft-hearted 16. Brutal 17. Humane 18. Sympathetic 19. Uncaring 20. Harsh
Note. The instruction read “Rate the extent to which you feel each of the following descriptive adjectives is characteristic or uncharacteristic of YOUR PERSONALITY AND BEHAVIOUR.”
3. Measuring Tough-mindedness as a Personality Trait Dimension Goertzel (1987), in his study of personality and sociopolitical attitudes, used a set of bipolar forced-choice trait adjective items to measure the personality construct of tough-mindedness versus tender-mindedness. Ten of Goertzel’s trait adjectives seemed to fit the construct of tough- versus tender-mindedness conceptualized here as an “interpersonal affective orientation of emotional concern and caring towards others as opposed to one of hard, uncaring, indifference to others’ fate and feelings.” These 10 trait adjectives, with 5 protrait and 5 contrait, were used in a unipolar rating format (i.e., the same format as for the social conformity items) in the first pilot study (Montreal). The ␣ coefficient obtained was .75, indicating an adequate level of internal consistency for a scale of this length. However, four of the items showed quite weak item–total correlations and were therefore discarded for the next two studies. Fourteen new trait adjective items (7 protrait and 7 contrait) were selected on the basis of their fidelity to the construct definition above (rated by two independent judges) and for showing high item–total correlations with the 6 core items from Goertzel’s scale. The resulting 20-item fully
57
IDEOLOGY AND PREJUDICE TABLE VI CORRELATIONS BETWEEN PERSONALITY AND IDEOLOGICAL ATTITUDES IN THREE SAMPLES (STUDY 1) Sample Auckland (n = 51) Montreal (n = 56) London (n = 53) ∗p
Personality scale
␣
RWA r
SDO r
Social conformity Tough-mindedness
.84 .90
.54∗∗ −.26
−.07 .42∗∗
Social conformity Tough-mindedness
.87 .75
.55∗∗ .09
Social conformity Tough-mindedness
.87 .85
.61∗∗ −.04
.18 .34∗∗ .20 .31∗
< .05. < .01.
∗∗ p
balanced tough- versus tender-mindedness scale is shown in Table V and was used in the next two studies (London and Auckland), with ␣ coefficients of .85 and .90, respectively. 4. Findings Table VI shows the ␣ coefficients for the social conformity and tough-mindedness scales and their correlations with the RWA and SDO scales. The findings are clear and completely consistent across the three samples. In each case, social conformity correlated significantly with RWA, but not with SDO. Moreover, these correlations exceeded .50 in each case, indicating very powerful effects (according to Cohen’s, 1988, conventions for effect size). Tough- versus tender-mindedness, on the other hand, correlated significantly with SDO in each case but not with RWA. In this case the effects were not quite as strong, being in the moderate range of effect size according to Cohen’s conventions. Later findings suggest that this difference is because the association of tough-mindedness with SDO is wholly and powerfully mediated by its corresponding social worldview. RWA and SDO were positively correlated in all three samples, with the correlation being nonsignificant in the Montreal sample (r = .19) and significant in the Auckland (r = .43) and London (r = .54) samples. The stronger correlations in the latter two samples seems consistent with the earlier speculation that RWA and SDO would be more strongly associated in more highly ideologized societies. Both England and New Zealand have been historically characterized by strong labor and socialist movements, with politics powerfully polarized along a single left–right polarity.
58
JOHN DUCKITT
Finally, whereas the correlations between social conformity and toughmindedness were nonsignificant in all three samples (−.10, Auckland; −.14, London; −.24, Montreal), they were consistently negative. A weak negative relationship seemed plausible because the traits characteristic of high tough-mindedness generally tend to be somewhat socially undesirable and deviant and thus seem less likely to be characteristic of socially conforming persons. Summing up, therefore, these three preliminary pilot studies showed that two theoretically derived personality dimensions of social conformity and toughmindedness could be reliably measured and correlated as expected with the two ideological attitude dimensions of authoritarianism and social dominance. This supported the idea that social conformity and tough-mindedness might be the personality bases of these two ideological attitude dimensions. However, the theoretical framework, which was broadly summarized in Table IV, goes well beyond this to propose a systematic set of causal links between socialization, personality, social worldviews, ideological attitudes, and prejudice. The causalities proposed are elaborated in the next section.
D. THE THEORETICAL MODEL The causal links between socialization, personality, social worldview, ideological attitudes, and ethnocentric attitudes proposed or implied by the theoretical framework summarized in Table IV are diagrammed as a hypothetical causal model in Fig. 1. The core model predictions are that the two socialization practice dimensions, punitive and unaffectionate socialization, impact on the two personality dimensions, social conformity and tough-mindedness respectively, which impact
Fig. 1. A causal model of the impact of socialization, personality, and worldview on the two ideological belief or social attitude dimensions of right-wing authoritarianism (RWA) and social dominance orientation (SDO) and their impact on intergroup attitudes.
IDEOLOGY AND PREJUDICE
59
on the two social worldviews, belief in a dangerous and competitive-jungle world respectively. Both personality and worldview then impact on RWA and SDO. This core model thus comprises a theory of the dual psychological bases of the two ideological attitude dimensions of authoritarianism and social dominance. In addition, following the findings of Altemeyer (1998) and McFarland (1998); (McFarland & Adelson, 1996) that RWA and SDO independently predict prejudice and ethnocentrism, it is proposed that these two ideological attitude dimensions would have independent causal impacts on both pro-in-group ethnocentric attitudes and antiout-group prejudice. The model, therefore, sees RWA and SDO as being determined by two quite different socialization, personality, and social worldview systems. Tentative links between these two systems can also be proposed, but consistent with the empirically demonstrated relative independence between RWA and SDO, these should mostly be relatively weak. First, the two socialization dimensions, punitive and unaffectionate socialization, seem likely to be positively correlated, with no causal direction indicated. Second, the three pilot studies reported earlier suggested a weak negative relationship between social conformity and tough-mindedness, with causality more likely seeming to be from social conformity to tough-mindedness than the reverse. Third, it seemed likely that viewing the world as dangerous and threatening and as a competitive jungle might well be positively associated. If so, the more plausible causal direction seems likely to be from seeing the world as a competitive jungle to seeing it as dangerous and threatening rather than the reverse. Fourth, RWA and SDO, as discussed above (cf. Table I), would be expected to impact reciprocally on each other due to pressures toward cognitive consistency, at least in more highly ideologized sociopolitical systems such as New Zealand. In societies such as North America, where politics has not been as strongly organized around a single left– right ideological polarity, these effects might be weak and, particularly in the case of younger, student samples, nonsignificant. The degree of reciprocal causality between RWA and SDO should therefore be sample and context dependent. Finally, following social identity theory it has been frequently assumed that in-group attitudes should causally influence out-group attitudes (Tajfel & Turner, 1979). Research, however, has suggested that this effect may be variable (Hinkle & Brown, 1990) and hold only when the relationship between groups involves some degree of threat and competition (Duckitt & Mphuthing, 1998). Because this does appear to be the case for majority–minority relations in New Zealand and Afrikaner–African relations in South Africa (cf., e.g., Duckitt & Mphuthing’s, 1998, findings), a causal impact of in-group attitudes on out-group attitudes was tentatively assumed for the studies reported in this chapter. However, reverse and reciprocal effects between in-group and out-group attitudes were also viewed as possible and tested.
60
JOHN DUCKITT
IV. Testing the Model A. INTRODUCTION A complex causal model relating individual difference variables such as socialization, personality, social worldview, and ideological attitudes is not easily tested experimentally. Such models can, however, be tested using structural equation modeling (SEM) with nonexperimental data in order to assess how well hypothesized causal models fit obtained data. The following sections report such findings from three different samples; two large student samples from Auckland University in New Zealand and a third consisting of White Afrikaans students from a historically White Afrikaans South African university (the University of Pretoria). The first study was only a partial test of the model as measures of the two socialization dimensions and one of the social worldview dimensions, the competitive-jungle social worldview, were not yet available when that study was done. The next two studies, however, tested the entire model diagrammed in Fig. 1.
B. STUDY 2: THE 1998 NEW ZEALAND STUDY 1. Scope and Objectives of the Study Survey data was obtained from large-enough samples of Auckland University undergraduate students to use SEM with latent variables to test certain core aspects of the proposed causal model. Thus, data from a very large sample (total sample) could be used to test the interrelationships proposed by the model between the two personality dimensions, social conformity and tough-mindedness; one worldview dimension, viewing the world as dangerous and threatening; and the two ideological attitude dimensions, RWA and SDO. Second, a smaller subsample of these participants also completed measures of interethnic attitudes, enabling a test of the same model expanded to include antiminority and promajority ethnic attitudes. 2. Total Sample and Measures of Personality and Ideology The total sample consisted of 511 undergraduate students (167 males and 343 females) at Auckland University. They completed a questionnaire that included 18 items from Altemeyer’s (1996) RWA scale and 10 items from Pratto et al.’s (1994) SDO scale. In both cases items were randomly sampled from the full scales to give equal numbers of pro- and contrait items. Both scales proved reliable in this sample with ␣ coefficients of .90 (n = 489) and .78 (n = .504), respectively. Social conformity and tough-mindedness were each measured using 14 of the trait
IDEOLOGY AND PREJUDICE
61
adjectives developed for these scales shown in Table V (items 1–14), again with each 14-item scale consisting of equal numbers of pro- and contrait adjectives. The participant instructions and response scales were the same as in the pilot studies. The ␣ coefficients for social conformity and tough-mindedness in this sample were .86 (n = 478) and .81 (n = 494), respectively. 3. Conceptualizing Social Worldview and Measuring Dangerous World The concept of social worldview was defined as a coherent set of beliefs about the nature of the social world, and specifically about what people are like, how they are likely to behave to one, and how they should be responded to and treated. The worldview of a dangerous and threatening social world was conceptualized as having two opposing poles. At one pole was the belief that the social world is a dangerous and threatening place in which good, decent people’s values and way of life are threatened by bad people versus, at the other pole, the belief that the social world is a safe, secure, and stable place in which almost all people are fundamentally good. The items of Altemeyer’s (1988) Belief in a Dangerous World scale express a belief that the social world is, or is not, a dangerous and threatening place and seemed to fit the theoretical conception of this particular worldview well. The questionnaire contained six items, randomly sampled to give equal numbers of pro- and contrait items, from the 12 items of Altemeyer’s (1988) Dangerous World scale. The ␣ in this study was .69 (n = 504) and the items are shown in Table IX (items 1–6). 4. European/Pakeha Sample and Promajority and Antiminority Ethnic Attitudes The European/Pakeha sample consisted of a subset of 311 of the participants from the total sample who, in addition to the measures described above, also completed measures of attitudes toward the four main ethnic groups in New Zealand. These were the European origin (locally known as “Pakeha”) politically dominant majority (comprising about 80% of the population) and the three minorities: Asians, Maori, and Pacific Islanders (Polynesians). The ethnic composition of this sample was 224 European/Pakeha, 44 Asian, 13 Maori, 9 Pacific Islanders, and 21 who gave ethnic identity as “other.” Because the analyses for this subsample investigated antiminority prejudice and promajority European/Pakeha attitudes, only the data from the 224 European/Pakeha were used in the analysis. Promajority (European/Pakeha) and antiminority (Asian, Maori, Pacific Islanders) ethnic attitudes were measured using four eight-item Likert scales with each consisting of four positive and four negative evaluative statements about each of the four groups. Illustrative items are shown in Table X. Earlier research and
62
JOHN DUCKITT
pilot testing had shown content validity for these scales and high levels of internal consistency reliability and factorial unidimensionality (Duckitt, 1998). The ␣ coefficients for these European/Pakeha participants were .80 for the proEuropean/Pakeha attitude scale (n = 219), .87 for anti-Asian attitudes (n = 220), .85 for anti-Maori attitudes (n = 218), and .86 for anti-Pacific Islander attitudes (n = 220). The three antiminority attitude scales intercorrelated positively and quite substantially (ranging from .48 to .67), suggesting that they could be viewed as tapping a single antiminority latent variable. 5. Analysis The items from each of the two personality and two ideological attitude scales were randomly assigned to create four manifest indicators per scale or latent variable. Thus, the manifest indicators consisted of item sets of two, four, or six items with each item set having been allocated to have equal numbers of pro- and contrait items. Because the Dangerous World scale had just six items, only three indicators were created for it, with each consisting of one pro- and one contrait item. The same manifest indicators were also used for the analysis of the European/ Pakeha sample. In addition, for this sample, the eight items of the pro-European/ Pakeha scales were randomly allocated to three manifest indicators (consisting of four, two, and two balanced item sets). Finally, the three antiminority scales (anti-Maori, anti-Pacific Islander, and anti-Asian) were each used as a manifest indicator for the single latent variable of antiminority attitudes. In the analysis, all manifest indicators were allowed to relate only to their specific latent variable. Maximum likelihood estimation was used for both analyses. The theoretical model proposed a reciprocal causal relationship between RWA and SDO, with each influencing the other. It has been recommended that prediction errors between reciprocally related dependent variables should be allowed to correlate in order to remove the influence of any unmeasured external variables affecting their covariation, which might otherwise distort the two reciprocal parameters (Rigdon, 1994; E. Rigdon, personal communication, July, 19, 1999; Shumacker & Lomax, 1996). A correlated error was therefore permitted between these two variables, with no other correlated errors allowed. All analyses were also repeated without this correlated error, with only minor differences being observed in the strength of the two parameters, and none on model fit or any other parameters. Overall model fit was assessed using Hu and Bentler’s (1999) recent recommendations. Their investigation of the optimal cutoff values of standard maximum likelihood LISREL fit indices suggested that good fit would be best indicated by values close to or better than 0.06 for RMSEA, 0.08 for SRMR, and 0.95 for CFI and GFI. In the case of GFI, Steiger (1989) and others (Maiti & Mukerjee, 1990) have noted that Joresk¨og’s widely used index of sample GFI will be biased downward when df are large relative to sample size and proposed a correction to
63
IDEOLOGY AND PREJUDICE
GFI that provides a more robust population estimate. Because the SEM analyses in this study and the next two studies tested mostly very large models relative to sample size, Steiger’s (1989) corrected population GFI was used rather than Joresk¨og’s sample GFI. Finally, the widely used rule of thumb for very large models, where chi-square values tend to be large, of viewing a 2/df ratio of less than 2 as a criterion of good fit was also used (Shumacker & Lomax, 1996). 6. Results for the Basic Model: Total Sample The correlations between the social conformity, tough-mindedness, dangerous world, RWA, and SDO scales for the total sample are shown in Table VII. The correlation between RWA and SDO was .37 (n = 484), which seemed higher than that typically found for student samples in North America (cf. Table I). Despite a truncated age range (over 80% of the sample was between ages 18 and 24 years), the correlation of RWA and SDO was stronger for older students (r = .46 with n = 224 for age >19 years) than for younger students (r = .28 with n = 224 for age ≤19 years). This difference was statistically significant (p = .02). The covariance matrix for the latent variable analysis for the total sample was based on an N of 444 (with listwise deletion of missing data). Figure 2 shows the standardized path coefficients for the model tested. The overall model fit was good, 2(144) = 249.8, 2/df ratio = 1.74, RMSEA = 0.041, SRMR = 0.044, corrected GFI = 0.98, CFI = 0.98, and was clearly superior to the null model in which all parameters were assumed to be uncorrelated, 2(171) = 4427.7. All predicted paths were significant, and no nonpredicted paths emerged as significant. Thus, social conformity had significant positive paths to dangerous world and RWA and TABLE VII CORRELATIONS AMONG THE VARIABLES FOR THE TOTAL SAMPLE OF STUDY 2 (n’S VARY FROM 459 TO 497) Variable 1. RWA 2. SDO 3. Soc Conf 4. Tough M 5. Dang W
1 .37∗∗∗ .56∗∗∗ .03 .45∗∗∗
2
3
.18∗∗ .30∗∗∗ .15∗∗
−.15∗∗∗ .26∗∗∗
4
−.09∗
Note. RWA = right-wing authoritarianism; SDO = social dominance orientation; Soc Conf = social conformity; Tough M = tough-mindedness; Dang W = dangerous world. ∗ p < .05. ∗∗ p < .01. ∗∗∗ p < .001.
64
JOHN DUCKITT
Fig. 2. Standardized maximum likelihood coefficients from Study 2 (total sample; n = 444) for the structural equation model of two personality, one social worldview, and two ideological attitude latent variables. Manifest variables are indicated with boxes and latent variables with ovals. A correlated error was permitted between RWA and SDO. All coefficients shown are statistically significant (p < .05).
dangerous world a significant path to RWA. Tough-mindedness had a significant positive path to SDO, and the reciprocal paths between SDO and RWA were both positive and significant. In addition, the more tentative expectation that social conformity would have a weak but significant negative path to tough-mindedness was supported. All the manifest indicators, which had been constrained to load only on their latent variables, loaded powerfully on those latent variables. The model accounted for 57% of the variance in RWA and 32% of the variance in SDO. 7. Results for the Extended Model: European/Pakeha Sample The covariance matrix for the SEM LISREL analysis for the European/Pakeha sample was based on an n of 213 (pairwise deletion of missing data was used because of the relatively low sample size). The correlations between the two personality, one worldview, two ideological attitude, and four ethnic attitude scales are shown in Table VIII. Figure 3 shows the statistically significant standardized path coefficients that were obtained. To simplify the figure the manifest indicators are not shown; however, the paths from latent variables to manifest indicators were
65
IDEOLOGY AND PREJUDICE
TABLE VIII CORRELATIONS AMONG THE VARIABLES FOR THE EUROPEAN/PAKEHA SUBSAMPLE OF STUDY 2 (n’S VARY FROM 201 TO 221) Variable 1. RWA 2. SDO 3. Soc Conf 4. Tough M 5. Dang W 6. Pro-Euro 7. Anti-Min
1
2
3
4
5
.36∗∗∗ .51∗∗∗ −.01 .41∗∗∗ .32∗∗∗ .46∗∗∗
.18∗∗ .31∗∗∗ .03 .46∗∗∗ .52∗∗∗
−.16∗ .20∗∗ .16∗ .10
−.11 .13 .27∗∗∗
.07 .30∗∗∗
6
.64∗∗∗
Note. RWA = right-wing authoritarianism; SDO = social dominance orientation; Soc Conf = social conformity; Tough M = tough-mindedness; Dang W = dangerous world; pro-Euro = proEuropean/Pakeha attitudes; Anti-Min = antiminority attitudes. ∗ p < .05. ∗∗ p < .01. ∗∗∗ p < .001.
all highly significant and generally powerful, with the weakest coefficient being .52 and the remainder all greater than .60. Overall model fit was good, 2(260) = 459.5, 2/df ratio = 1.77, RMSEA = 0.060; SRMR = 0.060, corrected GFI = 0.94, CFI = .94, and clearly superior to that of the null model, 2(300) = 2987.5. Once again all paths predicted by the theoretical model were significant (though one path, from RWA to pro-in-group attitudes, was marginally significant). First, the significant paths obtained for the total sample were replicated in this subsample: Social conformity had positive paths to dangerous world and RWA and dangerous world had a positive path to RWA. Tough-mindedness had a positive path to SDO, and social conformity again had the predicted weak negative path to tough-mindedness. RWA and SDO had positive reciprocal effects on each other. Second, for the extended model, both RWA and SDO had the expected significant positive paths to promajority (European/Pakeha) attitudes and antiminority attitudes. Promajority in-group attitudes had a powerful significant positive path to antiminority attitudes. A model with a reciprocal causal impact of antiminority attitudes on in-group attitudes was also tested but this path was nonsignificant. Thus, the basic pattern of interrelationships that were theoretically expected was supported. In addition to the relationships predicted by the model, two nonpredicted paths to antiminority prejudice were statistically significant. Social conformity had a significant negative path to antiminority prejudice (i.e., reducing prejudice), and dangerous world beliefs had a significant positive path to antiminority prejudice (i.e., increasing prejudice). Social conformity thus seemed to have two opposing effects on antiminority prejudice: first, the predicted indirect effect mediated
66
JOHN DUCKITT
Fig. 3. Standardized maximum likelihood coefficients from study 2 (European/Pakeha sample; n = 213) for the structural equation model of two personality, one worldview, two ideological attitudes, and pro-European/Pakeha in-group and antiminority out-group attitude latent variables. To simplify, manifest variables and the paths from latent to manifest variables are not shown. Latent variables are shown with ovals. A correlated error was permitted between RWA and SDO. The dashed-line arrows show nonpredicted paths. A superscript plus sign denotes a marginally significant (p < .10) path coefficient; all other coefficients shown are significant at the 5% level.
through RWA increasing prejudice (indirect effect = +.34) and, second, an unexpected direct effect of reducing prejudice (−.26). The model accounted for a substantial proportion of the variance in antiminority prejudice (R2 = .88) and in RWA (R2 = .54). In the case of SDO, the proportion of variance accounted for was lower (R2 = .35), but this could well have been due to the absence of a measure of the competitive-jungle worldview. 8. Overview and Implications of the Findings Overall, both SEM analyses indicated a good fit of data to the theoretical model. The findings thus supported the factorial independence of the latent variable constructs as well as being consistent with the pattern of causal relationships hypothesized between them. In particular, the core predictions about the relationships between the personality, worldview, and ideological attitude variables were clearly supported. Thus, social conformity and dangerous world beliefs had substantial effects on RWA but no effects on SDO. Tough-mindedness, on the other hand, had a reasonably substantial effect on SDO but not on RWA. In addition, the analysis of the extended model supported the expected independent effects for RWA and SDO on both pro-European/Pakeha majority group
IDEOLOGY AND PREJUDICE
67
ethnocentrism and antiminority prejudice; a finding that is consistent with the prior correlational findings by Altemeyer (1998) and McFarland (1998); (McFarland & Adelson, 1996). The two personality dimensions, social conformity and toughmindedness, did not increase antiminority prejudice or promajority attitudes directly. They did, however, have significant indirect effects mediated through RWA and SDO of increasing both antiminority prejudice (indirect effects, .34 and .31 respectively, with p < .05 for both) and promajority attitudes (indirect effects, .18 and .25 respectively, with p < .05 for both). Although two nonpredicted effects did emerge as significant, both seemed reasonably plausible and did not appear to invalidate the logic underlying the model. Thus, besides its significant indirect effects of increasing prejudice, social conformity also had a significant direct effect of reducing prejudice. Although unexpected, such a direct effect for social conformity would seem plausible in liberal social contexts, such as the one in which this study was done, in which antiminority prejudice was distinctly counternormative. Second, dangerous world beliefs had a significant direct effect of increasing antiminority prejudice. In many societies dominant, majority groups tend to view culturally dissimilar or militant minorities as sources of threat. In such situations, and New Zealand would appear to be one, it seems plausible that viewing the social world as dangerous and threatening would be associated with antiminority attitudes. Finally, it should be noted that three of the constructs in the theoretical model were not used in the analyses: that is, the competitive-jungle worldview and both childhood socialization measures. Thus, these analyses could only provide a partial test of the model. Nevertheless, the findings provided a useful basis for crossvalidation in the next two studies, which tested the complete model.
C. STUDY 3: THE 1999 NEW ZEALAND STUDY 1. Scope and Objectives Study 3 used data from a new 1999 survey of introductory psychology students at Auckland University to test the complete model diagrammed in Fig. 1. Once again two SEM analyses were done, with the first using all participants to test the core model of socialization, personality, worldview, and ideological attitudes and the second using only European/Pakeha participants to test an extended model including pro-European/Pakeha majority and antiminority ethnic attitudes.
2. Participants and Measures Survey questionnaires were completed by 381 introductory psychology students, 240 female and 139 male (with 2 not divulging gender). The ethnic
68
JOHN DUCKITT
breakdown was 59 Asian, 12 Maori, 16 Pacific Islanders (Polynesian), 237 Pakeha or European-origin New Zealanders, and 57 who gave ethnic identity as “other.” The mean age was 20.3 years (SD = 5.19). The personality, worldview, and ideological attitude measures that had been used in Study 2 were used again, though in some cases with additional items as follows: (1) RWA was measured by the same balanced 18-item version used in Study 2, with an ␣ of .88 (n = 372) in this sample. (2) SDO was a balanced 12item scale with 2 further items from Pratto et al. (1994) added to the 10 used in Study 2. The ␣ was .82 (n = 376). (3) Belief in a Dangerous World was a balanced 10-item scale with 2 new specially written items plus 2 additional items from Altemeyer’s (1988) scale added to the 6 items used in Study 2 (see Table IX). The ␣ was .81 (n = 376). (4) Social conformity was measured by the same 14 items used in Study 2 with an ␣ of .85 (n = 367). (5) Tough-mindedness was measured by the same 14 items used in Study 2 plus 2 additional items from the item pool in Table V (items 1–16), giving a balanced 16-item scale with an ␣ of .86 (n = 371). The four balanced 8-item ethnic attitude scales, which had been used in the previous study (Study 2), were used again to measure pro-European majority attitudes and antiminority attitudes using the European/Pakeha participants only. The ␣ were as follows: anti-Asian attitudes, .86 (n = 236); anti-Pacific Islander attitudes, .86 (n = 235); anti-Maori attitudes, .85 (n = 236); and pro-European/Pakeha attitudes, .75 (n = 235). Once again the correlations between attitudes to the three minorities were positive and substantial (ranging from .45 to .68), justifying using them as manifest indicators for a single antiminority latent variable (Table X). 3. Measuring a Competitive-Jungle Social Worldview The broad concept of social worldview was described as a coherent set of beliefs about the nature of the social world, and specifically about what people are like, how they are likely to behave to one, and how they should be responded to and treated. A competitive-jungle social worldview was defined as the belief that the social world is a competitive jungle characterized by a ruthless, amoral struggle for resources and power in which might is right and winning everything (protrait pole) versus a view of the world as one of cooperative harmony in which people care for, help, and share with each other (contrait pole). A number of the items in Altemeyer’s (1998) PP-MAD and EE-MAD scales (many in turn derived from various Machiavellianism scales) seemed to express this social worldview and were thus used with other specially written items as an item pool. These items were rated by three independent judges for fidelity to the
69
IDEOLOGY AND PREJUDICE TABLE IX THE CONSTRUCT OF SOCIAL WORLDVIEW AND THE TWO WORLDVIEW DIMENSIONS
The construct of social worldview A coherent set of beliefs about the nature of the social world, specifically, about what people are like, how they are likely to behave to one, and how they should be responded to and treated. Dangerous/threatening social worldview Belief that the social world is a dangerous and threatening place in which good, decent people’s values and way of life are threatened by bad people versus Belief that the social world is a safe, secure and stable place in which almost all people are fundamentally good. Competitive-jungle social worldview Belief that the social world is a competitive jungle characterized by a ruthless, amoral struggle for resources and power in which might is right and winning everything versus Belief that the social world is a place of cooperative harmony in which people care for, help, and share with each other. Dangerous world scale items 1. Although it may appear that things are constantly getting more dangerous and chaotic, it really isn’t so. Every era has its problems, and a persons chances of living a safe, untroubled life are better today than ever before. 2. Any day now chaos and anarchy could erupt around us. All the signs are pointing to it. 3. There are many dangerous people in our society who will attack someone out of pure meanness, for no reason at all. 4. Despite what one hears about “crime in the street,” there probably isn’t any more now than there ever has been. 5. If a person takes a few sensible precautons, nothing bad is likely to happen to him or her; we do not live in a dangerous world. 6. Every day as society become more lawless and bestial, a person’s chances of being robbed, assaulted, and even murdered go up and up. 7. My knowledge and experience tells me that the social world we live in is basically a safe, stable and secure place in which most people are fundamentally good. 8. It seems that every year there are fewer and fewer truly respectable people, and more and more persons with no morals at all who threaten everyone else. 9. The “end” is not near. People who think that earthquakes, wars, and famines mean God might be about to destroy the world are being foolish. 10. My knowledge and experience tells me that the social world we live in is basically a dangerous and unpredictable place, in which good, decent and moral people’s values and way of life are threatened and disrupted by bad people. Competitive-jungle worldview scale items 1. Winning is not the first thing; it’s the only thing. 2. The best way to lead a group under one’s supervision is to show them kindness, consideration, and treat them as fellow workers, not as inferiors. 3. If it’s necessary to be cold blooded and vengeful to reach one’s goals, then one should do it. 4. Life is not governed by the “survival of the fittest.” We should let compassion and moral laws be our guide. 5. Money, wealthy and luxury are what really count in life. 6. It is much more important in life to have integrity in your dealings with others than to have money and power. (continues)
70
JOHN DUCKITT TABLE IX (continued) 7. It’s a dog-eat-dog world where you have to be ruthless at times. 8. Charity (i. e., giving somebody something for nothing) is admirable, not stupid. 9. You know that most people are out to “screw” you, so you have to get them first when you get the chance. 10. My knowledge and experience tells me that the social world we live in is basically a competitive “jungle” in which the fittest survive and succeed; power, wealth, and winning are everything; and might is right. 11. One should give others the benefit of the doubt. Most people are trustworthy if you have faith in them. 12. We can make a society based on unselfish cooperation, sharing, and people generously helping each other, and not on competition and acquisitiveness. 13. If you have power in a situation, you should use it however you have to to get your way. 14. It is better to he loved than to be feared. Note. All items for the dangerous world scale except 7 and 10 are from Altemeyer’s (1988) scale. All items for the compatitive-jungle worldview scale except items 10 and 12 were taken or adapted from Altemeyer’s (1998) PP-MAD and EE-MAD scales.
definition above, and 12 items (6 protrait and 6 contrait) that had been unanimously rated as conforming to the definition were used in the survey. These 12 items, which are shown in Table IX (items 1–12), had an ␣ in this sample of .85 (n = 379), with highly significant item–total correlations for all 12 items. 4. Measuring the Childhood Socialization Dimensions Two sets of self-report items were written to assess the two socialization dimensions described in Ross’s research (1993) (where they had been studied as societal/cultural level indicators). Seven statements were selected on the basis of judges’ ratings for each scale, with one set of statements expressing childhood experience of punitive, strict, and harsh versus permissive, nonpunitive, and tolerant socialization by parents or caregivers and the other unaffectionate versus affectionate childhood socialization by parents/caregivers. One of the seven items initially selected for the punitive, strict socializaiton scale was discarded due to it having weak correlations with the total scale, leaving six items (five protrait and one contrait), which produced an ␣ of .86 (n = 375) in this sample. Five of the seven unaffectionate socialization items were included in the current survey (two protrait and three contrait), producing an ␣ of .81 (n = 379). The items of both these scales are shown in Table XI. 5. Analyses As in Study 2, two SEM analyses with latent variables were conducted. The first used the total sample to test the basic model of socialization, personality,
IDEOLOGY AND PREJUDICE
71
TABLE X ILLUSTRATIVE ITEMS FROM THE ETHNIC ATTITUDE SCALES USED IN NEW ZEALAND (ATTITUDES TOWARD ASIANS, MAORI, PACIFIC ISLANDERS, AND PAKEHA/EUROPEAN) AND SOUTH AFRICA (ATTITUDES TOWARD WHITE AFRIKANERS AND BLACKS) Attitudes toward Asians Asian immigrants are benefiting New Zealand and more should be encouraged to settle here. Young Asian students, through their hard work and perseverance, provide positive role models for the youth of New Zealand. If many more Asians were allowed into New Zealand they would take over the country. The prejudice against Asian immigrants in this country is really sick. Attitudes toward Pakeha (European-origin New Zealanders) The arrogant and conceited attitude of many Pakeha New Zealanders is really offensive to other ethnic groups in this country. Pakeha New Zealanders have ruthlessly expoited the natural resources of New Zealand and through their greed are ruining the enviorment. In general, Pakeha New Zealanders are treating other ethnic groups very fairly. Pakeha, with their energy and hard work, are an example that other groups in New Zealand should follow to get ahead. Attitudes toward Pacific Islanders More Pacific Islanders should be allowed to immigrate to New Zealand. Too many Pacific Islanders are taking advantage of the welfare system in New Zealand. An important reason why Pacific Islanders struggle to get ahead in New Zealand is the prejudice and discrimination against them. Pacific Islanders just don’t try hard enough to get ahead in this country. Attitudes toward Maori The main reason why the Maori standard of living is so low is the injustices done to them not only in the past but in the present as well. Much more needs to be done to redress the wrongs that have been done to Maori in this country. Maori in New Zealand have a privileged position today that is unfair to other ethnic groups here. Too much is being done for Maori in New Zealand today. Attitudes to White Afrikaners/Blacks (Study 4) It really upsets me to hear someone say something negative about Afrikaans-speaking people/Black people. White Afrikaans-speaking people/Black people have some very bad characteristics. I have a very positive attitude to White Afrikaans-speaking people/Black people. There is very little to admire about White Afrikaans-speaking people/Black people.
worldview, and ideological attitudes. The second extended this model to incorporate pro-European majority attitudes and antiminority attitudes and therefore used only the European/Pakeha participants. Three manifest indicators were used for each hypothesized latent variable. In the case of antiminority attitudes, the three manifest indicators were the full antiMaori, anti-Asian, and anti-Pacific Islander scales. All the other manifest indicators were constructed by randomly assigning the items of each scale to three subsets, with, where possible, equal numbers of pro- and contrait items in each subset. All
72
JOHN DUCKITT TABLE XI ITEMS OF THE TWO SOCIALIZATION SCALES
Punitive socialization scale items Harsh socialization subscale 1. I was severely scolded and reprimanded while growing up. 2. I was often physically punished in a painful manner while growing up. 3. I was often physically beaten while growing up. Strict socialization subscale 1. I would describe my upbringing as very strict. 2. I was strictly disciplined while I was growing up. 3. I grew up in an environment where I was rarely punished. Unaffectionate socialization scale items 1. I grew up in an unaffectionate environment. 2. I grew up in a hard and brutal environment. 3. I was emotionally close to my father (or my primary male caregiver) during my upbringing. 4. I received a great deal of affection from my parents/caregivers during my childhood. 5. I grew up in a caring an loving environment. 6. I did not receive love or affection from my father while I was growing up. 7. I did not receive much loving attention from others during my childhood.
the manifest indicators were thus balanced (consisting of two, four, or six items each) except for the two socialization scales where the unequal numbers of proand contrait items meant this was not possible for all their indicators. Maximum likelihood estimation was used for both analyses. Once again correlated errors were permitted only between the two reciprocally related variables of RWA and SDO (cf. Rigdon, 1994; Rigdon, personal communication, July 19, 1999; Shumacker & Lomax, 1996). 6. Results for the Basic Model Using the Total Sample The correlations between all the variables for the total sample are shown in Table XII. Once again the correlation between RWA and SDO of .40 (n = 369) was higher than has been reported for North American student samples. Again, despite a truncated age range, this correlation tended to be higher for older than for younger students (r = .28 with n = 175 for age<19 years versus r = .49 for n = 194 for age ≥ 19 years) with the difference significant ( p = .02). The covariance matrix for the LISREL analysis for the total sample was based on an n of 347 (listwise deletion of missing data). Figure 4 shows the standardized path coefficients for the model tested. The overall fit indices for the model indicated good fit for a model of this size, 2(238) = 411.5, 2/df ratio = 1.73, RMSEA = 0.046, SRMR = 0.050, corrected GFI = 0.96, CFI = 0.96, which was clearly better than that for the null model, 2(276) = 5131.9. All the latent variables loaded powerfully on their manifest indicators.
73
IDEOLOGY AND PREJUDICE TABLE XII CORRELATIONS AMONG THE VARIABLES FOR STUDY 3 (n’S VARY FROM 360 TO 377) Variable
1
1. RAW 2. SDO .40∗∗∗ 3. Soc Conf .52∗∗∗ 4. Tough M .05 5. Dang W .54∗∗∗ 6. Comp W .11∗ 7. Pun Soc .19∗∗∗ 8. Unaff Soc −.03 9. Pro-Euro .25∗∗∗ 10. Anti-Min .31∗∗∗
2
3
.03 .38∗∗∗ −.25∗∗∗ .21∗∗∗ . .26∗∗∗ .60∗∗∗ −.14∗∗ .12∗ .05 .08 −.12∗ .42∗∗∗ .12∗ .51∗∗∗ .08
4
5
.05 .55∗∗∗ .11∗ .17∗∗∗ .09 .33∗∗∗ −.05 .08 .15∗∗ .20∗∗∗ .25∗∗∗
6
7
8
9
.15∗∗∗ .16∗∗ .44∗∗ .21∗∗∗ −.01 .02 .36∗∗∗ −.04 .00 .56∗∗∗
Note. RWA = right-wing authoritarianism; SDO = social dominance orientation; Soc Conf = social conformity; Tough M = tough-mindedness; Dang W = dangerous world; Comp W = competitivejungle world; Pun Soc = punitive socialization; Unaff Soc = unaffectionate socialization; Pro-Euro = pro-European/Pakeha attitudes; Anti-Min = antiminority attitudes. ∗ p < .05. ∗∗ p < .01. ∗∗∗ p < .001.
Fig. 4. Standardized maximum likelihood coefficients from Study 3 (total sample: n = 347) for the structural equation model of two socialization, two personality, two worldview, and two ideological attitude latent variables. Manifest variables are indicated with boxes and latent variables with ovals. A correlated error was permitted between RWA and SDO. All coefficients shown are statistically significant ( p < .05). The dashed-line arrow shows a nonpredicted path.
74
JOHN DUCKITT
All the paths predicted by the model, except one, were statistically significant. Moreover, both the two socialization, personality, and worldview variable sets powerfully predicted their corresponding ideological attitude dimension, almost entirely as expected. Thus, in predicting RWA, punitive, strict socialization had the expected significant paths to social conformity. Social conformity had significant paths to belief in a dangerous world and RWA and dangerous world belief a significant path to RWA, with these effects mostly in the moderate to powerful range of effect size (Cohen, 1988). The model accounted for 64% of the variance in RWA. Second, in predicting SDO, unaffectionate socialization had the predicted significant path to tough-mindedness, which had a powerful path to belief in a competitive-jungle worldview, which in turn had a powerful path to SDO. The model accounted for 60% of the variance in SDO. One predicted effect here, a direct effect of tough-mindedness on SDO, was clearly nonsignificant (when its parameter was estimated it was only .01, t = 0.11) and was therefore dropped from the model. There was, however, a highly significant indirect effect of toughmindedness on SDO (.45, p > .0001) mediated through belief in a competitivejungle worldview. Finally, the correlational or causal links hypothesized between these two socialization, personality, worldview, and ideological attitude complexes were confirmed. Punitive socialization and unaffectionate socialization were positively correlated, social conformity had a negative path to (therefore decreasing) toughmindedness, belief in a competitive-jungle world had a positive path to (therefore increasing) belief in a dangerous world, and RWA and SDO had reciprocal positive effects on each other. Only one nonpredicted effect emerged as statistically significant. There was a highly significant though not very strong negative effect of unaffectionate socialization on social conformity (t = −2.95, p = .003). Thus, more unaffectionate socialization seemed to lower social conformity. 7. Results for the Extended Model: European/Pakeha Sample The covariance matrix for the latent variable analysis for the European/Pakeha sample was based on an n of 216 (listwise deletion of missing data). Figure 5 shows the statistically significant standardized path coefficients that were obtained for this model. To simplify the diagram, the manifest indicators, which all had powerful and highly significant paths from their corresponding latent variables (the weakest being .69), are omitted. The fit indices indicated good overall fit for the model: 2(385) = 604.2, 2/df ratio = 1.57, RMSEA = 0.051; SRMR = 0.058, corrected GFI = 0.94, CFI = 0.95, which was clearly better than that for the null model, being 2(435) = 4569.7. The model accounted for a substantial 54% of the variance in antiminority prejudice and for 29% of the variance in pro-European/Pakeha majority attitudes.
IDEOLOGY AND PREJUDICE
75
Fig. 5. Standardized maximum likelihood coefficients from Study 3 (European/Pakeha subsample: n = 216) for the structural equation model of two socialization, two personality, two worldview, two ideological attitude, and pro-European/Pakeha in-group and antiminority out-group attitude latent variables. To simplify, manifest variables and the paths from latent to manifest variables are not shown. A correlated error was permitted between RWA and SDO. The dashed-line arrows show nonpredicted paths. A superscript plus sign denotes a marginally significant (p < .10) path coefficient; all other coefficients shown are significant at the 5% level.
The significant paths for this subsample between the two socialization, personality, worldview, and ideological attitude latent variables were exactly the same as that obtained in the preceding analysis for the total sample. For the extended model, RWA and SDO, again, as in Study 2, had the expected significant positive paths to both antiminority prejudice and pro-European/Pakeha attitudes. Pro-European/Pakeha attitudes again had a strong positive path to antiminority attitudes. In the previous study (Study 2), two nonpredicted paths to antiminority attitudes had been significant: one a positive path from dangerous world beliefs, and the other a negative path from social conformity. The first of these was clearly replicated in this study. Belief in a dangerous world again had a significant positive path to antiminority attitudes. The second, however, was not replicated. When modeled, the path from social conformity to antiminority attitudes was .08 and nonsignificant and therefore dropped from the analysis. 8. Overview and Implications of the Findings The findings from both SEM analyses for Study 3 provided good support for the theoretical model. Despite the size and complexity of the model, its overall fit
76
JOHN DUCKITT
was good, and all the paths predicted were significant, with only one exception. Tough-mindedness did not have a direct path to SDO; however, it did have a moderate to strong indirect effect on SDO mediated entirely through belief in a competitive-jungle worldview. The analyses did reveal two nonpredicted significant paths. First, the direct effect of dangerous world beliefs increasing antiminority prejudice observed in Study 2 was replicated. A second nonpredicted path was from unaffectionate socialization to social conformity. Whereas it seems plausible that persons whose childhood socialization lacked affection would be less socially conforming, this path was relatively weak and only marginally significant for the European/Pakeha analysis. This suggests that it needs cross-validation in a new sample. Finally, the nonpredicted finding in the previous study (Study 2), that social conformity had a direct negative path to antiminority prejudice, was not replicated in this analysis, suggesting that it might have been due to chance. The significant path from punitive socialization to social conformity is particularly interesting because the zero-order correlation between these two variables (see Table XII) was only .05 and clearly nonsignificant. This illustrates the power of SEM to detect effects that would not be apparent using less sophisticated analyses. In this instance, the power derives from both the use of latent variables with measurement error eliminated and multivariate partialing to release spuriously suppressed effects. The spuriously suppressed effects in this case derive from punitive and unaffectionate socialization having a strong positive association with each other, but opposite associations with social conformity. That is, having more punitive parents increases social conformity, but more punitive parents also tend to be less affectionate, which reduces social conformity. This confound therefore results in univariate analyses underestimating the associations of both punitive and unaffectionate socialization with social conformity. However, although the predicted path from punitive socialization to social conformity was statistically significant, it did seem notably weaker than the model would have expected for both the total and European/Pakeha analyses. This could have been due to weak measurement. However, although the scale to measure punitive socialization was quite short and not fully balanced, this was also the case for the unaffectionate socialization scale, which showed the expected reasonably robust paths to its personality dimension, tough-mindedness. A possible reason for the relatively weak path coefficients from punitive socialization to social conformity was revealed when the correlations between the individual items of this scale and social conformity were examined. Three punitive socialization items correlated significantly with social conformity, whereas the other three did not, being very close to zero. Inspection of the content of these two item sets suggested a clear content difference, with the set correlating with social conformity seemingly dealing with parental strictness and the other with parental harshness (see Table XI).
IDEOLOGY AND PREJUDICE
77
An exploratory factor analysis on the six items had originally seemed consistent with unidimensionality, as only one eigenvalue greater than 1 was obtained. However, close inspection of the eigenvalues suggested that a scree test might have located a break between the second and third eigenvalues, which would be consistent with a correlated two-factor solution. This was confirmed using both oblique and varimax rotated two-factor solutions, with the three strictness items loading powerfully on one factor, and the three harshness items on the other. Each of the two item sets produced adequate alpha coefficients despite having only three items (.75 for strict socialization and .79 for harsh socialization). Using the three strict socialization items (with each as a manifest indicator) as a latent variable, instead of the original six-item punitive socialization latent variable, in the two SEM analyses produced clearly stronger paths to social conformity (i. e., .30 and .26 for total and European samples). A harsh socialization latent variable resulted in nonsignificant paths to social conformity (.00 and −.06 for the total and European/Pakeha samples respectively). However, an important caveat is in order. The distinction between parental harshness and strictness seems intuitively compelling, and it seems theoretically plausible that parental strictness might increase social conformity and parental harshness not. Nevertheless, this distinction was arrived at in post hoc fashion and the analyses separating the two could have been capitalizing on chance variation in the sample. Thus, the same strict and harsh socialization factors were used in the next study to see if their apparently different effects could be replicated.
D. STUDY 4: THE 1999 SOUTH AFRICAN STUDY 1. Objectives, Participants, and Measures In order to test the model in an entirely new and different sample, a survey questionnaire was administered in late 1999 to introductory psychology students at the University of Pretoria, a predominantly White Afrikaans language university in South Africa. The constructs measured were the same as in the previous study, except that the intergroup attitude measures were pro-White Afrikaner and antiBlack attitudes and punitive socialization was differentiated into strict and harsh socialization scales. The questionnaire was translated into Afrikaans and then backtranslated to check equivalence. 233 White Afrikaans students completed the questionnaire, 145 female and 88 male, average age 19.9 years (SD = 1.87). All the scales produced satisfactory ␣ coefficients, which were as follows: 10 SDO items, .79; 10 RWA items, .67; 10 dangerous world items (Table IX), .82; 14 competitive-jungle world items (Table IX, items 1–14), .76; 14 social conformity items (Table V, items 1–14), .72; 20 tough-mindedness items (Table V, items 1–20), .88; 3 strict socialization
78
JOHN DUCKITT
items (Table XI), .74; 3 harsh socialization items (Table XI), .78; 7 unaffectionate socialization items (Table XI), .78; 8 pro-White Afrikaner items (Table X), .72; and 8 anti-Black items (Table X), .84. 2. Analyses, Results, and Conclusions Three manifest indicators were constructed for each hypothesized latent variable by randomly assigning the items of each scale to three subsets, with, where possible, equal numbers of pro- and contrait items. The indicators for the strict and harsh socialization scales each consisted of three single items. The correlations between all the variables are shown in Table XIII. The correlation between RWA and SDO was .21, which was lower than the correlations obtained in the New Zealand studies. The correlation between RWA and SDO was also not moderated by age as it had been in the New Zealand samples, with similar correlations for younger (≤19 years; r = .24, n = 111) and older students (>19 years; r = .18, n = 114). This seems consistent with the idea that strong RWA–SDO correlations would be characteristic of societies where politics is highly ideologized and a tendency for the correlation to increase with age associated with socialization in societies where politics tends to be ideologically TABLE XIII CORRELATIONS AMONG THE VARIABLES FOR STUDY 4 (n’S VARY FROM 215 TO 233) Variable
1
2
1. RWA 2. SDO .21∗∗∗ 3. Soc Conf .44∗∗∗ −.00 4. Tough M −.11 .38∗∗∗ 5. Dang W .45∗∗∗ .29∗∗∗ 6. Comp W .03 .58∗∗∗ 7. Strict Sc .21∗∗∗ .16∗ 8. Harsh Sc .07 .14∗ 9. Unaff Sc −.08 .08 10. Pro-Afrik .49∗∗∗ .43∗∗∗ 11. Anti-Bla .39∗∗∗ .49∗∗∗
3
−.17∗ .17∗ −.18∗∗ .11 −.02 −.08 .27∗∗∗ .04
4
5
.12 .48∗∗∗ .19∗∗ .10 .14∗ .13 .18∗∗ .30∗∗∗ −.15∗ .06 .37∗∗∗ .20∗∗ .49∗∗∗
6
.14∗ .22∗∗∗ .14∗ .12 .29∗∗∗
7
8
9
10
.60∗∗∗ .20∗∗ .21∗∗∗ .11 .02 −.08 .17∗∗ .10 .00 .50∗∗∗
Note. RWA = right-wing authoritarianism; SDO = social dominance orientation; Soc Conf = social conformity; Tough M = tough-mindedness; Dang W = dangerous world; Comp W = competitivejungle world; Strict Sc = strict socialization; Harsh Sc = harsh socialization; Unaff Sc = unaffectionate socialization; Pro-Afrik = pro-White Afrikaner attitudes; Anti-Bla = anti-Black attitudes. ∗ p < .05. ∗∗ p < .01. ∗∗∗ p < .001.
IDEOLOGY AND PREJUDICE
79
Fig. 6. Standardized maximum likelihood coefficients from Study 4 (n = 206) for the structural equation model of two socialization, two personality, two worldview, two ideological attitude, and pro-White Afrikaner ingroup and anti-Black outgroup attitude latent variables. To simplify, manifest variables and the paths from latent to manifest variables are not shown. All coefficients shown are statistically significant (p < .05). The dashed-line arrows show nonpredicted paths.
organized. In South Africa political parties represent ethnic groups and politics is ethnically rather than ideologically organized (see, e. g., Guelke, 1996; Johnson & Schlemmer, 1994). The covariance matrix for the SEM analysis for the total sample was based on an n of 206 (listwise deletion of missing data). Figure 6 shows the standardized path coefficients for the model tested. To simplify the diagram, the manifest variables and paths from latent to manifest indicators are not shown. The overall fit indices for the model indicated good fit for a model of this size, 2(386) = 506.4, 2/df ratio = 1.31, RMSEA = 0.039, SRMR = 0.062, corrected GFI = 0.96, CFI = 0.95, which was clearly better than that for the null model, 2(435) = 3039.2. All the latent variables loaded powerfully on their manifest indicators (the weakest loading being .53 with no other loading below .65). Overall, the model accounted for substantial proportions of the variance in RWA (62%), SDO (63%), pro-Afrikaner attitudes (56%), and anti-Black attitudes (58%). All the paths predicted by the model were significant except one: Toughmindedness once again did not have a direct effect on SDO, but did have a reasonably substantial mediated effect through a competitive-jungle worldview. In this, and indeed almost all, respects, the pattern of significant paths closely replicated those obtained in the previous two New Zealand studies. Thus, the originally nonpredicted effects of unaffectionate socialization in reducing social
80
JOHN DUCKITT
conformity, and of dangerous world beliefs in increasing anti-out-group attitudes, were again obtained. Another originally nonpredicted effect, a direct effect of social conformity in reducing anti-out-group prejudice, which had been obtained in the first New Zealand study (Study 2), but not in the second (Study 3), was again significant. The results of the previous study had suggested that the original six-item punitive socialization measure might have encompassed two strongly related but distinct dimensions, strict and harsh socialization, with only the former relating to social conformity. This was confirmed in this study. Exploratory factor analyses showed the strict and harshness item sets shown in Table XI loading on two separate factors. When modeled as a latent variable (with the three items as manifest indicators) in the SEM, strict socialization had a significant path to social conformity, while harsh socialization, and the full six-item-based punitive socialization latent, did not. However, although both strict and unaffectionate socialization had significant paths to social conformity, both paths were relatively weak. Finally, whereas in the New Zealand samples RWA and SDO had reciprocal effects on each other, in this White Afrikaner sample only RWA had a significant path to SDO, with the reverse path nonsignificant (−.05, nonsignificant). This is not inconsistent with the model, which had expected the relationships between RWA and SDO to be context dependent. It also seems interpretable in terms of a White Afrikaner nationalism characterized by an intense need for social control and cohesion arising out of an historical experience of extreme insecurity and threat from powerful out-groups (initially British imperialism and later African nationalism). Thus, as historical commentators have suggested, in this society the need to dominate rival groups may have been secondary to, and stemmed from, threat-driven authoritarianism and not the reverse (de Kiewiet, 1957; MacCrone, 1937; see also Duckitt, 2000). The very much more powerful path from RWA to pro-Afrikaner attitudes (.57) in this sample compared to the equivalent paths in the New Zealand samples (.14 and .27) also fits this characterization. Further evidence supporting this characterization is discussed later.
E. OVERALL CONCLUSIONS FROM THE MODEL TESTING STUDIES 1. The Core Model: Personality, Worldview, and Ideology The results of the three SEM studies indicated good fit of the empirical data to the theoretical model of socialization, personality, worldview, ideological attitudes, and ethnocentrism. Ideally, the model would have been compared to plausible alternative models. In these studies the alternative model was the null or independence model simply because no reasonably plausible alternative models seemed apparent. However, the difficulty of obtaining good fit indices for very large and complex
IDEOLOGY AND PREJUDICE
81
models has frequently been commented on (Potthast, 1993; Yung & Bentler, 1994), and the model tested here was clearly a very large and complex one. In addition, all but one of the paths predicted by the model were significant over all three analyses. The one consistently nonsignificant path suggested only a minor modification to the model: that the personality dimension of tough-mindedness does not effect social dominance directly, but only indirectly through powerfully facilitating the adoption of a competitive-jungle social worldview. In contrast, the personality dimension of social conformity had, as expected, a direct effect of facilitating dangerous world beliefs, and both dangerous world beliefs and social conformity had moderate to strong effects on authoritatrian attitudes. In essence, therefore, the core of the model, that two ideological attitude dimensions of RWA and SDO will be each determined by two distinct personality and social worldview dimensions, was clearly supported, and the model accounted for very substantial proportions of the variance in both ideological attitude dimensions. 2. Ideology and Intergroup Attitudes As predicted, the two ideological attitude dimensions both had direct and independent effects on pro-in-group and anti-out-group attitudes, and the model accounted for substantial proportions of the variance in both, but particularly in out-group attitudes. These findings are consistent with the earlier correlational findings by Altemeyer (1998) and McFarland (1998); (McFarland & Adelson, 1996). However, two nonpredicted, but theoretically plausible, paths to anti-out-group prejudice were also significant. First, seeing the social world as dangerous and threatening had a direct effect of increasing out-group prejudice, over and above its expected indirect effect through increasing authoritarian attitudes. This effect seems plausible for out-groups that would be viewed as threatening by in-group members, as would tend to be the case for majority–minority relations in New Zealand and Afrikaner–Black relations in South Africa. Second, social conformity, in addition to its expected effect of indirectly increasing prejudice through increasing authoritarianism, also had an unexpected direct effect of reducing prejudice in two of the three studies. Such a direct effect seems plausible in social contexts in which prejudice is non-normative, such as in New Zealand and in the new South African “rainbow nation.” Socially conforming persons should show greater conformity to these general societal norms of nonprejudice. But then, why was the effect not evident in the second New Zealand study? A possible reason is that the kind of tolerance associated with social conformity may be relatively “skin deep.” Ambivalent and relatively superficial attitudes of racial tolerance have been well documented in the United States in research on
82
JOHN DUCKITT
reverse discrimination, symbolic racism (McConahay & Hough, 1976), and “liberal discrimination” (Gaertner, 1973). Rogers and Prentice-Dunn (1981) have also shown that, when angered by a Black, a White’s tolerance may rapidly regress to overt racism and discrimination. In New Zealand there tends to be periodic upsurges of outrage in the media and among the majority general public over “excesses” by Maori activists, “welfare abuse” by minority and immigrant groups, or the “problem” posed by the influx of Asian and Pacific Island migrants. It is possible that at such times social norms of nonprejudice against minorities might weaken and the “tolerance” of more socially conforming majority members evaporate. If so, direct effects of social conformity in reducing prejudice might be situationally variable and fluctuate with shifts in the normative climate. 3. Socialization and the Two Personality Dimensions The more speculative prediction from the model that two parental socialization dimensions might influence the two personality dimensions of social conformity and tough-mindedness was generally supported, but produced somewhat more complex findings. In one instance the findings were clear-cut. Unaffectionate socialization had its expected effect on tough-mindedness, with effect sizes in the moderate to strong range. The findings for social conformity were less clear. First, unaffectionate socialization had a weak but consistent nonpredicted effect of reducing social conformity. This effect seems theoretically quite plausible. Persons who have experienced very little affection from their parents may have generally weaker attachments to others and society in general and thus tend to be lower in social conformity. Second, the original dimension of punitive socialization was relatively weakly associated with social conformity and seemed to encompass two distinct though quite strongly related subdimensions: strict socialization, which was related to social conformity, and harsh socialization, which was not. Although this was replicated in a second study, the relationship of strict socialization to social conformity, though significant, was relatively weak. This might have been due to the weakness of the measure, which used only three items. The development of a more adequate measure that sampled the domain of strict parental socialization practices more comprehensively might produce stronger relationships. 4. Authoritarianism and Social Dominance as Ideological Dimensions The research also produced several findings that seemed to support the view of RWA and SDO as measuring ideological or social attitude dimensions rather than personality trait dimensions. First, the tendency for the correlation between RWA and SDO to be stronger in older than younger persons in prior North American research (summarized in Table I) was replicated in the New Zealand samples but not in the South African one. This age effect had seemed explicable in terms of political
IDEOLOGY AND PREJUDICE
83
socialization and participation creating cognitive pressure for ideological consistency in societies in which politics was ideologically organized; that is, with the political right associated with both high RWA and SDO and the left with both low RWA and SDO. Thus, the absence of an age effect in South Africa could have been due to politics being ethnically rather than ideologically organized in that society. And second, this reasoning also suggested that RWA and SDO should be more strongly associated in more highly ideologized societies. This was supported by their significantly stronger correlation in the New Zealand samples (aggregated r from Studies 1, 2, and 3 was .39, n = 904) and the UK sample (r = .54 in Study 1) than in prior North American research (r’s of .125 and .23 from Table I) and the South African sample (r = .21). 5. Some Methodological Issues Finally, in concluding, two methodological issues pertinent to these studies seem to merit comment. First, a frequently neglected methodological caution for research using psychometric scales to investigate construct interrelationships should be noted. This is that content overlap between subsets of items in different scales can build in spurious relationships between the scales. A number of steps were taken to control for this. In pilot testing, items which did not clearly correlate more powerfully with their own scales than other scales were deleted or rewritten. Independent judges carefully checked all the measures used for item content overlap between scales and discarded any dubious items. For example, Saucier (1994) had included items such as liberal, conservative, traditional, religious, and nonreligious in his original norm orientation personality scale, from which the social conformity scale was derived. These were excluded from the social conformity scale because they could be used to describe attitudes as well as behavior and might have spuriously inflated the relationship between conformity and RWA. Exploratory factor analyses were conducted on the data from the three studies to assess scale unidimensionality and the factorial independence of the scales. As Altemeyer (1981) has recommended, the correlations between individual items and other scales were checked to confirm that scale intercorrelations derived from all the items of the scales and were just not created by particular subsets of items. And finally, good fit for structural equation models with latent variables provides a powerful test for the unidimensionality of the latent variables and their factorial independence from each other. Second, the issue of causality merits comment. SEM can test the fit of hypothesized causal models to empirical data. Good fit will support a causal model, but does not directly demonstrate causality in the unequivocal manner that experimental research can. However, although the evidence from the SEM studies reported here cannot “prove” the model, there are several other considerations that also support the plausibility of its causal propositions. Thus, it seems theoretically reasonable that socialization will influence personality and that personality will
84
JOHN DUCKITT
influence social attitudes and beliefs. Moreover, the relatively few relevant experimental or longitudinal research studies that have been done support the causalities proposed by the model. For example, Katz and Hass (1988) showed experimentally that priming two ideological attitude measures similar to RWA and SDO (Protestant Ethic beliefs and a humanitarianism–egalitarianism scale) influenced racial attitudes, but priming racial attitudes did not influence ideological attitudes. Altemeyer (1988) showed that scenarios that depicted threatening and dangerous social changes markedly increased RWA. Several other longitudinal or experimental studies have shown that social threat seems to increase authoritarianism and prejudice (Doty, Peterson, & Winter, 1991; Meloen 1983; Sales, 1974; Sales & Friend, 1973).
V. Dual Dimensions of Ideology and Prejudice The theoretical model outlined here suggests that there are two different sets of cognitive and motivational dynamics driving prejudice and ethnocentrism: the two motivational goal–schema dimensions of threat–control and competision– dominance, with each expressed in their corresponding ideological belief dimensions of conservative–authoritarianism and social dominance–hierarchy. This has two important implications. First, the model provides a rudimentary theory of the psychological bases of politics and political ideologies. It suggests that different combinations of these two motivational–ideological belief dimensions would underlie support for or belief in particular political ideologies. Thus, as shown in Table XIV, fascism is generated by a combination of competitive dominance motivation (social dominance or inequality beliefs) and threat–control motivation (authoritarian– conservative beliefs), communism is generated by cooperative altruistic motivation (egalitarianism) together with threat–control motivation (authoritarianism), TABLE XIV THE TWO MOTIVATIONAL GOAL-SCHEMA DIMENSIONS EXPRESSED IN DUAL DIMENSIONS OF IDEOLOGICAL BELIEFS THAT DEFINE A FOURFOLD TYPOLOGY OF CONTEMPORARY POLITICAL IDEOLOGIES Opposing motivational goal-schemas and ideological beliefs Competitive–dominance (Social dominance) versus Cooperative altruism (Egalitarianism)
Threat-control versus (Authoritarianism)
Security-autonomy (Libertarianism)
Fascism
Free-market capitalism
Communism
Social democracy
IDEOLOGY AND PREJUDICE
85
free-market capitalism is generated by competitive–dominance motivation (social dominance/inequality beliefs) together with security–autonomy motivation (libertarianism), and social democracy is generated by security–autonomy motivation (libertarianism) and cooperative altruistic motivation (egalitarianism). Second, the model suggests the existence of two qualitatively different kinds of prejudice. One kind of prejudice will be driven by fear and threat generating selfprotective, defensive motivational needs for social control and security. Here outgroups will be disliked because they are seen as threatening and dangerous to social or group cohesion, security, order, values, and stability. This would tend to create a particular kind of readiness to categorize in a prejudiced fashion; that is, categorizing the social world into “us,” who are good, normal, moral, decent people and who are threatened by “them,” who are bad, disruptive, immoral, and deviant. This could also be conceptualized as priming a particular kind of intergroup schema. The second kind of prejudice will be driven by a view of the world as a competitive jungle characterized by a ruthless and amoral struggle for resources and power in which the fittest and most powerful succeed and the unfit and weak fail. This activates self-enhancement motives for power, status, superiority, and dominance. Here out-groups will be derogated because they are seen as inferior, weak, inadequate, and failures. This would create a readiness for a different kind of prejudiced social categorization, with the social world categorized into “us,” who are superior, strong, competent, and dominant (or should rightfully be dominant), versus “them,” who are inferior, incompetent, weak, and worthless. In practice, despite their origin in different motives, these two kinds of categorization or intergroup schemas could be highly correlated. This would be expected because both dimensions would be expressions of negativity, though different kinds of negativity, to the same out-group. This would be particularly the case when the out-group was viewed as being both “bad and threatening” and “incompetent and inferior,” as has often been the case for those seriously stigmatized social groups who have been most studied in research on prejudice. Nevertheless, if two largely independent personality, worldview, motivational, and ideological systems drive two distinct kinds of prejudice proneness, or readiness to categorize in prejudiced fashion, this should generate two correspondingly different kinds of prejudice, which should be identifiable as distinct factors. There is a good deal of empirical evidence for the existence of two such dimensions of prejudice. This evidence has emerged in research on intergroup stereotypes, attitudes, and affect as well as from research on discriminatory behavior and racism. The two dimensions that have emerged in each of these different research areas are summarized in Table XV. A number of studies have identified two clearly distinct dimensions of intergroup stereotypes (Fiske, 1998; Fiske, Xu, Cuddy, & Glick, 1999, Giles & Ryan, 1982; Poppe & Linssen, 1999; Singh, Choo, & Poh, 1998). One dimension, typically
86
JOHN DUCKITT TABLE XV TWO KINDS OF SOCIAL CATEGORIZATION, THEIR MOTIVATIONAL BASES, AND THEIR EXPRESSION IN DUAL DIMENSIONS OF GROUP STEREOTYPES, ATTITUDES, AFFECT, AND DISCRIMINATORY BEHAVIOR
Readiness to categorize motivation
Kind of social categorization schema
Threat–control motivated: authoritarian
Good/decent Beneficence Disliking versus traits versus bad/deviant liking
Competitive– Superior dominance versus motivated: inferior social dominance
Stereotype dimension
Attitudinal dimension
Affective dimension
Racism/ discrimination
Negative: Aversive/ Anger, fear, symbolic/“hot” anxiety discrimination
Competence Disrespect Low positive traits versus affect respect (disdain)
Dominative/ traditional/ “cold” discrimination
labeled competence, loads on trait attributes involving achievement, ability, and prestige, whereas the second, usually labeled beneficence or warmth, loads on trait attributes involving moral goodness versus badness. Fiske (1998) has recently described two distinct dimensions of intergroup attitude that directly parallel and seem to derive from these two stereotype dimensions. Thus, attitudes of liking versus disliking seem to be engendered by beneficence stereotypes, and attitudes of respecting versus disrespecting seem be inspired by competence stereotypes. These two dimensions of liking and respecting also emerged as clearly distinct ethnic attitude dimensions in Brewer and Campbell’s (1976) classic empirical study of East African tribal groups. In the case of intergroup affect, the present model suggests that threat–controldriven prejudice would be characterized by negative affective feelings of fear and anger, and competitive–dominance-driven prejudice by feelings of superiority. Research by Dijker (1987) has revealed two relatively independent dimensions of intergroup affect toward minorities. One dimension was characterized by the three closely related negative affects of anger, worry, and anxiety, which corresponds directly to what would be expected for threat-driven prejudice. Dijker’s (1987) second dimension was characterized by an absence of positive affect. Stangor, Sullivan, and Ford (1991) later showed that both positive and negative emotional responses predicted antiminority prejudice, with the absence of positive affect being a somewhat stronger predictor. It seems plausible that superiority-driven prejudice reflecting a view of out-groups as inferior would be manifest primarily by a lack of positive affects toward those groups. The model would hypothesize as well that at high levels of superiority-driven prejudice a lack of positive affect may be expressed more strongly as disdain or even contempt.
IDEOLOGY AND PREJUDICE
87
Two parallel dimensions have also been identified for discriminatory intergroup behavior and racism. First, Kovel (1970), in his classic comparison of racism in the American North and South, differentiated between two distinct kinds of racism. Aversive racism was characterized by dislike of Blacks and a desire to avoid them, whereas dominative racism was characterized by a belief in their inferiority and support for their social subordination. Second, Fiske (1998) has more recently argued that two kinds of discriminatory behavior can be distinguished that are similarly motivated. She suggests that “hot” discrimination is motivated by negative affects such as anger and hostility and that “people high in right-wing authoritarianism might well enact this kind of affect-laden discrimination” (p. 374). On the other hand, “cold” discrimination is motivated by negative stereotypes about “an outgroup’s interests, knowledge, and motivations”; that is, it would seem, stereotypes of inferiority. A number of researchers have made the important conceptual distinction between symbolic or modern and traditional or old-fashioned racism (Kinder & Sears, 1981; McConahay & Hough, 1976; Sears, 1988), which has a clear similarity to that made here. Symbolic or modern racism has typically been described as blending anti-Black affect with conservative ideological beliefs and the feeling that Blacks threaten and violate traditional American values such as individual responsibility and self-reliance, the Protestant work ethic, obedience, and discipline (see, e.g., Kinder & Sears, 1981, p. 416). Traditional racism, on the other hand, has been seen as consisting primarily of beliefs in Black inferiority, White supremacy, and support for racial segregation. Thus, there seems to be a clear similarity between symbolic racism and threat–control authoritarian prejudice. Traditional racism seems to have a clear similarity to competitive–superiority dominance-driven prejudice, although it probably involves a significant component of threat as well. Although the degree to which the distinction between symbolic and traditional racism has been empirically supported is still controversial (cf. Sniderman & Tetlock, 1986; Weigel & Howes, 1985), there are several important reasons why these findings may have been relatively inconclusive. First, measures of these two kinds of racism have tended to be psychometrically weak, typically being short and not balanced against acquiescence. Second, traditional racism probably blends dominance and threat motives, so involving a view of Blacks as both inferior and bad or threatening, which would create significant overlap between the two kinds of racism. And third, measures or symbolic and traditional racism have invariably been confounded with another conceptual and measurement distinction unrelated to the one being made here: the difference between subtle and blatant expressions of prejudice (cf. Pettigrew & Meertens, 1995). If two distinct forms of prejudice do exist, then an important prediction follows. The two ideological attitude dimensions, authoritarianism and social dominance, should differentially predict these two dimensions of prejudice. There is little evidence bearing on this because of the absence of adequate measures differentiating
88
JOHN DUCKITT
the two prejudice dimensions and because most research has therefore used global intergroup attitude and stereotype measures. However, an important study by Katz and Hass (1988) has reported clearly supportive findings. Their study investigated the relationship between two value orientation dimensions and two factorially distinct measures of racial attitudes. The two values dimensions closely resembled RWA and SDO, being respectively a Protestant Work Ethic (PE) scale (largely social conservatism items) and a Humanitarianism–Egalitarianism (HE) scale (beliefs in equality and concern for others). Their two racial attitude measures were described as Anti-Black and Pro-Black scales. The Anti-Black items expresssed a view of Blacks as deviant, disruptive, and violating important moral values. The Pro-Black scale’s items denied (or asserted) that Black’s low status was due to them being disadvantaged or discriminated against (low scores thus implying that their low status was deserved and due to incapacity). Respectively these two scales strongly suggest the two kinds of categorization schemas identified here of first good, decent, and moral versus bad, immoral, and deviant and second superior and competent versus inferior and incompetent. Katz and Hass’s (1988) findings were clear-cut and completely consistent with the prediction here. PE values most powerfully predicted anti-Black attitudes and HE values most strongly predicted pro-Black attitudes. Katz and Hass (1988) also demonstrated that these relationships appeared to be causal. Thus, priming PE values increased anti-Black but not pro-Black attitudes, whereas priming HE values increased pro-Black but not anti-Black attitudes. Overall, therefore, there is evidence that two distinct dimensions of prejudice can be identified corresponding closely to the two motivationally generated intergroup categorization schemas proposed by the model. Glick and Fiske (this volume) have shown how different combinations of the two stereotype dimensions of competence and warmth provide a fourfold typology of out-group attitudes. These are out-group admiration (out-group competent and warm), envious prejudice (outgroup competent, but not warm), contemptuous prejudice (out-group not warm and incompetent), and paternalistic prejudice (out-group warm, but incompetent). The two categorization schema dimensions proposed here (good versus bad paralleling warmth and superior versus inferior paralleling competence) would generate essentially the same typology, but with one important addition. There should be an intermediate and critically important category of “equal” along the superior versus inferior categorization dimension. This “out-group equal” category, when crossed with the good versus bad categories, adds two further intergroup attitude types. First, the combination of categorizing out-groups as equal and good (supportive) would define an attitudinal zone of tolerance, or genuine nonprejudice. Second, the combination of categorizing out-groups as equal but bad (threatening) would define hostile prejudice, which would be the typical pattern for threat-driven out-group hostility against out-groups relatively equal in status and
IDEOLOGY AND PREJUDICE
89
power with which the in-group is in a directly competitive relationship. The classic exemplar for this type of out-group prejudice was provided by Sherif’s (1967) famous studies of intergroup competition between boys attending summer camps. Two categories (superior and good or superior and bad) concern attitudes of subordinate to dominant groups generating, as Glick and Fiske have proposed, out-group admiration in the former instance and “envious prejudice” in the latter case. The term “resentful dislike” seems preferable to the “envious prejudice” used by Glick and Fiske, since among oppressed groups these attitudes might often be characterized more by a sense of injustice than envy of higher status. Finally, the two categories of inferior and good, and inferior and bad, concern the attitudes of dominant to subordinate groups; that is, following Glick and Fiske, paternalistic and contemptuous prejudice respectively. This sixfold typology would seem to accommodate most of the kinds of prejudice or intergroup attitude that have been distinguished (cf. the six kinds described in Duckitt, 1992b). For example, symbolic racism would seem to correspond to hostile prejudice (out-group viewed as bad and threatening, but not necessarily as inferior), and traditional racism to contemptuous prejudice (out-group viewed as inferior and bad or threatening).
VI. Prejudice as an Individual and a Group Phenomenon A. COGNITIVE-MOTIVATIONAL BASIS OF IDEOLOGICAL ATTITUDES The theoretical core of this model of socialization, personality, worldview, ideology, and ethnocentrism was provided by the concept of motivational goal-schema: a construct developed by psychological anthropologists to explain how sociocultural environments shape and influence the motivational goals of individuals and groups (D’Andrade, 1992; Ross, 1993; Strauss, 1992). As applied in this model, the construct proposes that the socialization experiences and personality dispositions of individuals influence their acquisition of chronically accessible schemas about the social world, which form the basis of reasonably stable and organized social worldviews. These social worldviews activate motivational goals, which, if they are socially shared and collectively held, may be expressed in ideological beliefs and sociocultural values. These ideological attitudes generate a proneness to ethnocentrism and prejudice; that is, a “readiness to categorize” in particular ways by priming particular kinds of intergroup categorizations and identities. The causal chain proposed by the model and supported by the model testing studies can be seen in cognitive terms as linking four basic kinds of internally consistent belief systems or schemas. First, socialization experiences generate self schemas or beliefs about the self (personality) and in particular how the self
90
JOHN DUCKITT
would be likely to respond in given situations. These self schemas influence social schemas (social worldviews); that is, beliefs about social others (what others are like, how they are likely to respond to each other and the self, and how one should respond to them). These social schemas influence ideological beliefs or societal schemas; that is, beliefs about the appropriate structure of society and the relations and roles of individuals and groups toward and within the societal collective. Finally, these ideological beliefs or societal schemas influence people’s readiness to adopt particular kinds of intergroup categorization schemas and social identities (beliefs about how intergroup relations are typically structured and how in-group and out-group are likely to relate to each other).
B. ROLE OF THE INDIVIDUAL: PERSONALITY The model and these findings have important implications for the old controversy in social psychology over the impact of personality on prejudice. The idea that personality directly affects prejudice has been widely held since the publication of The Authoritarian Personality (Adorno et al., 1950). However, the degree and nature of this influence in relation to the influence of group-level processes has been controversial (cf. Brown, 1995; Pettigrew, 1958, 1960; Turner & Giles, 1981). Moreover, most of the evidence cited for a direct impact of personality on prejudice has rested on the flawed assumption that authoritarianism measures such as the F and RWA scales measured personality rather than ideological beliefs (cf. similar critiques by Feldman & Stenner, 1997; Goertzel, 1987; Stone et al., 1993; Verkuyten & Hagendoorn, 1998). The current model and evidence suggest a more complex picture. Personality does not have direct effects on prejudice, but has substantial impacts on ideological beliefs (with the impact being either wholly or partly mediated through social worldviews), and it is these ideological attitudes that influence prejudice and ethnocentrism. Moreover, personality operates by influencing the kinds of motivational goals that are salient for individuals. Particular kinds of motivational goals will have collective relevance for social groups and therefore tend to be socially shared and so be expressed in the form of ideological beliefs or social attitudes such as social dominance or authoritarianism. More specifically, the empirical findings from the model testing studies suggest that this happens somewhat differently for these two ideological attitude dimensions. Tough-minded personalities have a very powerful tendency to see the social world as a ruthlessly competitive jungle. Presumably, the interpersonal styles likely to be adopted toward others by tough- as opposed to tender-minded persons (i.e., hard, cynical, and brutal as opposed to kind, caring, and compassionate) may be particularly likely to elicit behavioral confirmation that the social world really is a competitive jungle. Once adopted, this worldview seems to powerfully activate
IDEOLOGY AND PREJUDICE
91
power and dominance as motivational goals and their expression in social dominance attitudes and values (cf. Winter, 2000). On the other hand, whereas socially conforming personalities do tend to see the world as more dangerous and threatening, the effects in the model testing studies were only weak to moderate. However, having a socially conforming personality and holding dangerous world beliefs both had strong direct effects on authoritarian attitudes, suggesting that both operate to make the motivational goals of social control and security particularly salient for individuals.
C. ROLE OF THE SITUATION: SOCIAL WORLDVIEW An important implication of the model flows from the central role of worldviews in determining ideological beliefs and prejudice. The findings did support Ross’s (1993) hypothesis that social worldviews are influenced by individuals’ personalities; that is, that they would reflect dispositional tendencies to interpret the environment in particular ways. However, it is also plausible that individuals’ direct experience of their social realities will powerfully influence their social worldviews. This suggests that particular social environments will be highly conducive to prejudice and others to tolerance and that these effects will occur primarily through social environments’ impact on individuals’ worldviews. First, prejudice will be generated when the social environment really is dangerous, threatening, and unpredictable as opposed to being safe, stable, and secure. Second, prejudice will be generated when the social world really is a competitive jungle in which people compete ruthlessly for resources, status, and power and in which the “fittest” and most powerful succeed and the weak and “unfit” fail as opposed to a social world of cooperation, sharing, equality, and altruistic concern for others. This also means that when social environments change drastically and seemingly irrevocably, sudden and dramatic changes in social worldviews may occur that would produce equally dramatic changes in motivational goals and their expression in ideological beliefs and prejudiced attitudes. For example, Altemeyer (1988) found that when students were given scenarios in which a future Canadian society had experienced a long decline and was in economic and political crisis with escalating crime, unemployment, and terrorism, their RWA scores increased dramatically. This finding has a direct historical parallel with the catastrophic political and economic crisis triggered by the Great Depression in 1930s Germany that catapulted the Nazis from political obscurity to massive electroal support and state power in only a few years and ultimately culminated in world war and the holocaust. These social worldviews and ideological beliefs are held by individuals but are also socially shared beliefs, or social representations (Moscovici, 1983). Their transformation from individual-level constructs to socially shared ones may
92
JOHN DUCKITT
occur largely because worldviews reflect the social environment, which is inevitably a shared environment (D’Andrade, 1992; Strauss, 1992). Particular social worldviews will therefore tend to characterize particular collectivities, cultures, or social groupings and activate shared motivational goals, which will be expressed in shared ideological beliefs and values. Thus, whereas these worldviews and ideological beliefs may operate as individual difference variables influencing prejudice proneness in individuals, because they are typically socially shared beliefs, they will also influence proneness to prejudice in social groups or cultures.
D. ETHNOCENTRIC CULTURES AND SOCIETIES The model therefore suggests that cultures and societies high on the social worldview and ideological attitude dimensions of threat–control or competitive– dominance should be more prone to ethnocentrism and conflict. Cultures or societies high on both should be particularly so. The relatively little research that has compared cultures and societies on relevant dimensions has produced broadly confirmatory findings. Ross’s (1993) research, which has already been noted, compared 90 preindustrial societies and found that those characterized by punitive and unaffectionate socialization practices were involved in more violent conflicts. Bonta (1997) studied completely peaceful societies in which aggression, violence, and warfare were almost totally absent and found that their most salient characteristic was a powerful normative emphasis on cooperation, with any kind of competitiveness being completely eschewed. Meloen (2000) compared 63 countries and found that Hofstede’s cultural value dimensions of collectivism and power distance were powerfully associated with undemocratic practices and involvement in military conflict. A number of ethnographic studies have described violent and ethnocentric cultures and subcultures that are characterised by a view of the social world as a competitive jungle in which power, toughness, machismo, defence of one’s honor, and dominance become important values and goals. These seem to be central features of cultures of honor and warrior cultures. Such competitive–dominance cultures have been documented in the American South by Nisbett and Cohen (1996), in fascist movements (Billig, 1978), in the post-Vietnam paramilitary subcultures of America (Gibson, 1994), and in violent underclass and gang subcultures (Toch, 1992). In such cultures a pervasively salient social categorization therefore seems to be the distinction between those who are strong and superior, and therefore worthy of respect, and those who are weak, inferior, and unworthy. Threat–control cultures, on the other hand, seem quite different. Altemeyer (1988), for example, has noted how authoritarian beliefs systems are characterized by fundamentalist religiosity, moral self-righteousness, and a pervasive intolerance of nonconformity and deviance. These were prominent features of the White Afrikaner culture that gave rise to Apartheid in South Africa.
IDEOLOGY AND PREJUDICE
93
MacCrone (1937) has described how this culture was originally forged in the South African frontier situation. Once the Dutch colonists had spread from the original settlement at the Cape to the interior, the frontier situation and an enduring conflict with the more numerous indigenous Black people and the more powerful force of British colonialism developed. The essential characteristics of the frontier situation were isolation, danger, war, and extreme insecurity. In this situation, intensely felt religious beliefs fused with racial distinctions and the dichotomy between Christian and heathen, between good and evil, became the dichotomy between White and Black. This created a society characterized by a sense of omnipresent threat and vulnerability with an intense emphasis on group cohesion, conservatism, conformity, and a pervasive ethnocentrism. In threat–control cultures such as this, the pervasively salient social categorization therefore seems to be between “us” who are decent, normal, and good and who are threatened by “them,” who are bad, disruptive, deviant, and immoral. The characterization of White Afrikaners as a threat–control culture suggests several hypotheses that might indicate whether the constructs of social worldview and ideological attitudes can be used to describe and explain social group differences as well as individual differences.
E. COMPARING WHITE AFRIKANERS AND EUROPEAN NEW ZEALANDERS Previous research has shown that White Afrikaners tend to be particularly high on authoritarianism, ethnocentrism, and anti-Black prejudice (Duckitt, 1992a; Hampel & Krupp, 1977; Mynhardt, 1980). European/Pakeha New Zealanders, on the other hand, are generally regarded as being reasonably tolerant, and New Zealand as a society tends to be a reasonably tolerant, multicultural society in which intergroup relations, though strained at times, are not generally viewed as seriously antagonistic. White Afrikaners should therefore be significantly more ethnocentric than European New Zealanders. Moreover, if White Afrikaners’ ethnocentrism is an expression of a threat–control culture then they should be markedly higher than New Zealanders on authoritarianism and dangerous world beliefs, but not much higher on social dominance and a competitive-jungle worldview. Thus, controlling for authoritarianism and dangerous world beliefs should eliminate the difference between the two groups in ethnocentrism, whereas controlling social dominance and a competitive-jungle worldview should not. Studies 3 and 4 used measures of these constructs with European New Zealand and White Afrikaner samples, making a comparison between the two groups possible. Unfortunately the pro-in-group and anti-out-group measures in the two studies used different target groups and the content of the items was different, so that a direct comparison was not possible for ethnocentrism. However, the response
94
JOHN DUCKITT TABLE XVI ETHNOCENTRISM, SOCIAL WORLDVIEW, AND IDEOLOGICAL ATTITUDE ITEM RESPONSE SCALE MEAN SCORES FOR WHITE AFRIKANERS (n = 224–227) AND EUROPEAN/PAKEHA NEW ZEALANDERS (n = 231–236) White Afrikaners Measures
Ethnocentrism index Ideological attitudes RWA SDO Social worldviews Dangerous world Competitive world
NZ Europeans
Mean
SD
Mean
SD
F
d
1.91
1.69
.03
1.66
143.1
.98
1.68 −.82
1.20 1.45
−1.21 −1.18
1.23 1.14
650.8 9.1
1.53 .28
1.56 −2.15
1.43 1.05
−.16 −1.51
1.24 1.14
189.2 38.9
1.08 −.56
Note. All effects significant at p < .01. The item means range from +4 (highest) to −4 (lowest).
scales were the same in both studies (9-point scales from +4 to −4). This meant that summing pro-in-group and anti-out-group attitudes and dividing by the total number of items would give a very crude common index (on a scale of +4 to −4) of the degree to which White Afrikaners and European New Zealanders were endorsing ethnocentric (pro-in-group and anti-out-group) sentiments in general. Table XVI shows the means for the two groups on this index of ethnocentrism and their means for RWA, SDO, dangerous world beliefs, and competitive-jungle world beliefs (using only those items that were included in both studies). To facilitate interpretation, the means for all the measures were divided by the number of items so that all the means shown are on the same standard response scale of +4 to −4. Because the samples were large, small differences would be statistically significant, so Cohen’s (1988) d was computed to give an index of effect size. Cohen (1988) has suggested that d values of .20, .50, and .80 could be viewed as denoting weak, moderate, and strong effects respectively. Table XVI shows that Afrikaners’ mean score on the ethnocentrism index was very much higher than that for European New Zealanders, with Cohen’s d indicating a difference of almost 1 pooled standard deviation. As expected, Afrikaners were also much higher on dangerous world beliefs and RWA, with both effects very powerful. In contrast, Afrikaners were moderately lower on the competitive-jungle worldview. Although Afrikaners were significantly higher on SDO, the effect was slight (Cohen’s d indicating a weak effect), and when RWA was statistically controlled using analysis of covariance, the difference on SDO was reversed. Afrikaners’ mean SDO score adjusted for RWA (adjusted M = −1.47) was significantly lower than that for NZ Europeans (adjusted M = −1.18; F = 29.7, p <.0001). In contrast, controlling for SDO left the difference in RWA virtually unchanged (Afrikaner adjusted M = 1.62, NZ European adjusted M = −1.21; F = 694.2, p >.0001).
95
IDEOLOGY AND PREJUDICE
This suggested that Afrikaners’ higher ethnocentrism than European-origin New Zealanders might be entirely due to the their elevated dangerous world beliefs and authoritarianism. This was confirmed by a series of analyses of covariance controlling for different combinations of worldview and ideological attitudes. The adjusted ethnocentrism means obtained are shown in Table XVII. When SDO and competitive world beliefs were controlled, the differences in ethnocentrism between the two groups was little changed, with Afrikaners’ ethnocentrism means still very significantly higher. On the other hand, when RWA and dangerous world beliefs were controlled, Afrikaners’ mean ethnocentrism score was lower than that of European-origin New Zealanders, with the effect marginally significant (p = .07). This reversal in ethnocentrism level seemed likely to be due to Afrikaners being lower in SDO when RWA was controlled, and being lower in competitive-jungle world beliefs. This was confirmed by controlling for SDO and competitive-jungle world beliefs as well as RWA and dangerous world beliefs, in which case the ethnocentrism means for the two groups were similar and the difference statistically nonsignificant. These findings therefore go beyond those previous studies that have shown White Afrikaners to be high in authoritarianism and ethnocentrism (Duckitt, 1988; Hampel & Krupp, 1977; Mynhardt, 1980). They support the view that Afrikaner ethnocentrism seems to be specifically threat–control driven and characterized by a view of the world as dangerous and threatening and by authoritarian attitudes and not by a competitive–jungle worldview or by social dominance beliefs. Indeed, with differences in authoritarianism and dangerous world beliefs eliminated, Afrikaners tended to be lower than European-origin New Zealanders in competitive world beliefs, social dominance, and ethnocentrism. More broadly, these findings suggest that the constructs of worldview and ideological attitudes can usefully describe and explain group- and culture-level differences in ethnocentrism and prejudice.
TABLE XVII MEAN ETHNOCENTRISM INDEX SCORES AFTER ADJUSTING FOR SOCIAL WORLDVIEWS AND IDEOLOGICAL ATTITUDES FOR WHITE AFRIKANERS AND EUROPEAN/PAKEHA NEW ZEALANDERS Ethnocentrism means Variables adjusted for
Afrikaner
NZ European
df
F
p
SDO, CW RWA, DW SDO, CW, RWA, DW
1.80 .80 1.04
.15 1.17 .93
1,441 1,444 1,433
124.72 3.27 .36
<.0001 .07 .55
Note. SDO = social dominance orientation; CW = competitive-jungle world beliefs; RWA = right-wing authoritarianism; DW = dangerous world beliefs.
96
JOHN DUCKITT
VII. Integrating Individual and Intergroup Dynamics of Prejudice A. INTERGROUP DYNAMICS AND MOTIVE ACTIVATION If two kinds of motivational dynamics and their corresponding intergroup categorization schemas drive prejudice proneness in individuals, it seems likely that these motives or schemas would be activated by certain specific situational characteristics of intergroup relationships. Thus, threat–control motivation and authoritarian readiness to categorize as good versus bad seems likely to be activated by social or intergroup threat. A great deal of research indicates that the perception of threat from out-groups causes negativity to them (e.g., Brewer & Brown, 1998, pp. 565–566; Blake & Mouton, 1984; Stephan, Ybarra, Martinez, Schwarzwald, & Tur-kaspa, 1998; Struch & Schwartz, 1989). Intergroup threat has also been a primary focus of important group dynamic theories of prejudice and intergroup relations. Thus, the role of “real” threats to group resources and objectives has been articulated by Realistic Conflict theory (Jackson, 1993; LeVine & Campbell, 1972; Sherif, 1967). Contact theory has noted the role of intergroup contact conditions that involve threatening or unpleasant contact with out-group members, and contact involving competing goals and interests in causing and reinforcing prejudiced attitudes (Pettigrew, 1998). Terror Management theory has focused on threats to basic sociocultural values and beliefs (Solomon, Greenberg, & Pyszczynski, 1991; see also Esses et al., 1993; Zanna, 1994), and Social Identity theory on threats to acquiring and maintaining positive social identities (Tajfel & Turner, 1979). Second, competitive–dominance motivation and the readiness to categorize as superior versus inferior seems likely to be activated by intergroup power and status differentials and intergroup social competition over relative dominance and superiority. The importance of intergroup inequalities in status and power as a fundamental condition for establishing and maintaining intergroup prejudice has been a central tenet of the contact hypothesis (Allport, 1954; Stephan, 1987; Pettigrew, 1998). The role of intergroup social competition over relative status and power in influencing intergroup bias and discrimination has been largely articulated by social identity theorists (Tajfel & Turner, 1979; Turner & Giles, 1981). A great deal of empirical evidence has shown that intergroup status and power differentials and social competition over them do generate intergroup bias and discrimination (Brewer & Brown, 1998, pp. 570–571; Ellemers, 1993; Mullen, Brown, & Smith, 1992; Sachdev & Bourhis, 1987). Moreover, under certain circumstances social competition over relative status and power could also involve social or intergroup threat. This would be particularly likely when the dominance of powerful groups was threatened by competition from subordinate groups or groups relatively equal in power or status were competing
IDEOLOGY AND PREJUDICE
97
Fig. 7. Two situational intergroup dynamics that activate dual motivational goals and ideological beliefs which then generate a readiness to apply two intergroup categorization schemas underlying prejudiced attitudes and stereotypes.
for dominance or superiority. In such circumstances, competition over power and status differentials could activate both competitive–dominance and threat–control motivations. If these two kinds of intergroup dynamics cause intergroup prejudice by activating two corresponding kinds of motivationally based readiness to categorize, as shown in Fig. 7, then three important propositions seem to follow: 1. The two motivational bases of prejudice and their corresponding intergroup dynamics should define two kinds of social groups that will be typical targets of prejudice. This has important implications for understanding prejudiced intergroup attitudes in minority and majority groups. 2. The two intergroup dynamics should specifically elicit the two different dimensions of prejudice summarized in Table XV. 3. The two individual difference determinants of prejudice should be specifically activated by their corresponding intergroup dynamics; that is, these individual difference and intergroup situational dynamics should interact in determining prejudice. These three propositions are elaborated and evidence bearing on them reviewed in the next three sections.
B. PREJUDICE IN MINORITIES AND MAJORITIES If prejudice has its roots in threat–control and competitive–dominance motivational goals in individuals, and intergroup dynamics of social threat or power– status differentials in the social context, this suggests that two kinds of social groups should be the typical targets of prejudice.
98
JOHN DUCKITT
1. Prejudice deriving from threat–control motivation should be specifically directed at those social groups and categories of persons that are viewed by societies or groups as threatening their stability, cohesion, security, order, traditions, and values. In this case the socially threatening groups are viewed as bad, disruptive, threatening, immoral and deviant, and disliked. In contrast, groups or social categories that are seen as supporting and reinforcing social cohesion, security, and order will be liked and viewed as good, moral, and decent people. The typical targets of threat–control-motivated prejudice could be threatening out-groups or categories of persons within societies who are culturally dissimilar, seem socially disruptive, do not share the traditional social values and conventional beliefs, or threaten the society in some or other way. Social and cultural minority groups will therefore usually be the targets of this prejudice. Majority groups, on the other hand, will usually be closely identified with and indeed define the conventional norms, values, beliefs, and traditions of society and dominate its social control system (authorities, courts, police, and military) and will therefore be favored. 2. Prejudice deriving from competitive–dominance motivation should be directed specifically against those social groups and categories who are low in power–status (engendered by group inequality) or whom it is believed should be legitimately lower in power–status (engendered by intergroup power–status competition). These social conditions would activate motivational goals of power, superiority, and dominance in individuals. This will result in individuals seeking power and superiority by identifying with dominant groups, or groups that it is believed should be dominant, respecting and admiring them and viewing them as strong, superior, competent, and worthy. This will be accompanied by disidentification with subordinate groups; disrespecting and derogating them; and viewing them as inferior, incompetent, unworthy, and inadequate. Competitive–dominance-motivated prejudice will therefore be directed against socially subordinate groups and categories of persons, who are low in power and status, and will lead to favoritism toward dominant social groups, who are high in power and status. Once again, therefore, the result will generally be prejudice against minorities, or at least low power–status minorities, and favoritism toward majority groups, or at least high power–status majorities. A critical implication of this is that both threat–control and competitive– dominance motives should predict the same pattern of prejudiced intergroup attitudes in both majority and minority members. Thus, both minority and majority persons who are high in RWA and SDO should be more negative to minorities and more favorable to majorities despite the irony that the prejudice here is being directed against their in-group and the favoritism toward their out-group. Much evidence has been reported, particularly by Sidanius, Pratto, and their colleagues, that SDO and authoritarianism or conservatism predict promajority and
IDEOLOGY AND PREJUDICE
99
antiminority biases for both majority and minority persons (Levin, Sidanius, Rabinovitz, & Federico, 1998; Mitchell & Sidanius, 1993; Sidanius, 1993; Sidanius, Feshbach, Levin, & Pratto, 1997; Sidanius & Pratto, 1999, pp. 234–246; Sidanius, Pratto, & Mitchell, 1994; Sidanius, Pratto, & Rabinovitz, 1994). Thus, as Sidanius and Pratto (1999) have recently pointed out, whereas the classic proin-group anti-out-group ethnocentrism described by Sumner (1906) seems typical of majorities, reverse ethnocentrism has tended to be more common in minorities. If competitive–dominance and threat–control motivate promajority and antiminority attitudes, this effect should be qualified by the degree to which the minorities accept the legitimacy of the social structure and stratification system. To the extent that minorities reject the existing social structure as illegitimate, they will be less likely to view deviance from and rejection of its norms and values as threatening. Thus, minority persons high in threat–control motivation will be less likely to be positive toward groups that dominate or are closely associated with that social structure and negative toward groups which reject it. Similarly, if minority members see the relative status of dominant and subordinate groups as undeserved and illegitimate, they will be less likely to satisfy needs for power, superiority, and dominance by identifying with and favoring the dominant groups and disidentifying with and disfavoring the subordinate groups. This reasoning suggests that, whereas SDO and authoritarianism–conservatism should predict promajority and antiminority attitudes in the same direction for both majority and minority groups, the effect should typically be weaker for minorities and sometimes even absent. A number of studies have demonstrated this “ideological asymmetry effect” (Sidanius & Pratto, 1999), with SDO or authoritarian– conservatism predicting promajority and antiminority attitudes significantly less strongly for minority than majority members (Bahr & Chadwick, 1974; Combs & Comer, 1982; Levin & Sidanius, 1999; Levin, Sidanius, Rabinovitz, & Federico, 1998; Mercer & Cairns, 1981; Mitchell & Sidanius, 1993; Sidanius, Pratto, & Rabinovitz, 1994). Two studies have directly examined whether perceived legitimacy of intergroup status and power differentials between majority and minority moderates the degree to which SDO predicted promajority and antiminority attitudes in minority members. Both these studies found that at high levels of system legitimacy, SDO predicted promajority and antiminority biases equivalently in majorities and minorities, but when system legitimacy was low the effect vanished for the minority members (Levin, Sidanius, & Rabinovitz, 2000; Rabinovitz, 1999). A finding that is essentially consistent was reported by Federico (1998) for the stability versus instability of intergroup status relationships, a variable that should be closely related to perceived legitimacy and operate in the same manner, at least for minorities. Thus, he found that SDO significantly predicted promajority bias for minority participants when the intergroup status differential was perceived to be stable, but not when it was perceived to be unstable.
100
JOHN DUCKITT
This reasoning can be extended to situations of extremely high levels of perceived system illegitimacy (or instability), where an existing social structure and order is totally rejected and a majority or elites’ dominance is seen as completely illegitimate. In such a situation, ideological attitudes expressing competitive– dominance and threat–control motivations might be even associated with prominority and antimajority biases. This should happen for competitive–dominance when a minority believes that the majority’s dominance is illegitimate and believes, not merely that it should be equal to the majority, but that it should rightfully be dominant over it. It should happen for threat–control when the perceived illegitimacy of the majority’s dominance of society is combined with a powerful minority group identification and a view of the majority as threatening the cohesion, security, traditions, and values of the minority. Thus, for the White Afrikaners in Study 4, today a political minority in South Africa, authoritarianism and social dominance both predicted pro-Afrikaner and anti-Black (the new majority) attitudes. In research during the 1960s in the United States, Marx (1967) obtained significant positive correlations between the F scale and anti-White attitudes for radical separatist Black Muslims. To sum up, it was argued that both the competitive–dominance and threat– control motivational dynamics should generate promajority and antiminority attitudes in minority as well as in majority group members as long as minority members accepted the legitimacy of the social stratification system. The empirical evidence has supported this with authoritarianism–conservatism and SDO predicting pro-majority and antiminority attitudes in minorities, but usually less strongly than for majorities. Finally, several studies have directly shown that the effect does seem to be moderated as expected by perceived system legitimacy.
C. INTERGROUP DYNAMICS AND THE DUALITY OF PREJUDICE It has been argued that two motivational dynamics in individuals, competitive– dominance and threat–control, create a readiness to apply two different intergroup categorization schemas, which are expressed as two distinct forms or dimensions of prejudiced attitudes and stereotypes. If two different situational intergroup dynamics, social or intergroup threat and power–status differentials and competition, cause prejudice by activating these two motives, then these two situational intergroup dynamics should also generate the two dimensions of prejudiced attitudes and stereotypes, as shown in Fig. 7. Several studies have reported evidence that support this. Brewer and Campbell (1976), in their classic study of tribal groups in East Africa, found that out-group liking was best predicted by similarity (i.e., presumably more similar groups would be less threatening), whereas respect or evaluation was best predicted by the
IDEOLOGY AND PREJUDICE
101
out-groups’ status or level of technological advancement. Giles and Ryan (1982), who made the original distinction between competence and beneficence stereotypes, also reported evidence indicating that evaluations on the competence stereotype dimension seem to reflect groups’ relative status. Mullen, Brown, and Smith’s (1992) meta-analysis of studies of intergroup bias also produced supportive findings. They found that high-status groups showed bias against low-status groups on directly status-relevant dimensions (i.e., task competence or ability) but not on diffuse, non-status-relevant dimensions (such as liking or moral evaluation). Low-status groups, who would tend to experience social identity threat from high-status groups, on the other hand, showed bias against high-status groups on diffuse, non-status-relevant dimensions of liking and moral evaluation. More recently a series of studies by Fiske and her colleagues (Fiske, Glick, Cuddy, & Xu, 1999; Fiske, Xu, Cuddy, & Glick, 1999; see also, Glick & Fiske, this volume) have provided compelling evidence that these two kinds of intergroup dynamics do differentially predict the two kinds of intergroup attitudes hypothesized. Their research used a large number of target social groups that were evaluated on the two stereotype dimensions of competence and warmth (liking). The findings indicated that competence evaluations were predicted by group status, whereas warmth evaluations were predicted by the degree to which the out-group was seen as threatening or competitive. More specifically, they demonstrated that high-status but threatening–competing out-groups were evaluated high on competence, but low on warmth. On the other hand, low-status but nonthreatening or cooperatively interdependent groups were rated high on warmth, but low on competence. Overall, therefore, there is evidence suggesting that two well-established intergroup determinants of prejudice, out-group threat, and out-group power–status inequalities, do seem to relate differentially to two distinct dimensions of prejudiced intergroup stereotypes and attitudes: the dimensions of warmth–(dis-)liking and competence–(dis-)respecting respectively.
D. THE INTERACTION BETWEEN INDIVIDUAL DIFFERENCES AND GROUP DYNAMICS 1. Additive and Interactive Effects It has been argued that two individual difference factors, in the form of the strength of two motivational goals for individuals, and two corresponding intergroup dynamic factors, which situationally activate these two motivational goals, together determine prejudiced intergroup attitudes. These two processes seem likely to influence intergroup attitudes in both additive and interactive fashion.
102
JOHN DUCKITT
First, the nature of the intergroup relationship, that is, how threatening or unequal and competitive in power–status the out-group is, should determine how strongly threat–control or competitive–dominance motivational goals are aroused. Second, individuals’ personalities and worldviews will determine the importance of these two motivational goals for each individual. Thus, when these motivational goals are situationally activated, individual differences in their importance for the individual will modulate their strength and consequently the degree to which their corresponding prejudiced categorization schemas are applied to in-group and outgroup. Thus, both intergroup situational dynamics and individual differences should additively influence the strength of the two motivational goals in relation to any salient out-group and their expression in attitudes to that out-group. However, these individual differences and intergroup situational dynamics should also interact in determining out-group prejudice. If particular motivational goals are not situationally activated, that is, if the nature of an intergroup relationship makes particular motivational goals seem irrelevant, their importance to the individual will not be invoked in the context of that particular intergroup distinction. The less relevant these motivational goals seem to a salient intergroup distinction, the less individual differences in the importance of these motivational goals to the individual should influence attitudes to the out-group. More specifically, authoritarianism, being threat driven, should be activated by out-group or social threat and therefore more predictive of negativity to threatening than to nonthreatening out-groups or minorities. Social dominance, being motivationally driven by superiority, dominance, and power, should be activated by intergroup inequalities in power and status or competition over them. Thus, social dominance should be more predictive of intergroup attitudes when such power–status inequalities or competitiveness characterize the intergroup relationship than when they are absent. Empirical evidence for such interactive effects is summarized in Table XVIII. 2. Authoritarianism and Intergroup Threat Interactions Several studies have reported significant interactions between various indices of social or out-group threat and authoritarianism on anti-out-group and ethnocentric attitudes, with authoritarianism being more predictive when threat is higher (Feldman & Stenner, 1997; Rikert, 1998). Downing and Monaco (1986) found that high authoritarians were more negative to a directly competing out-group than low authoritarians, whereas high and low authoritarians did not differ in negativity to noncompeting out-groups. In the high-intergroup-threat context of Apartheid South Africa, correlations between the RWA scale and anti-Black prejudice for English-speaking White students were powerful, ranging from .53 to .69 (Duckitt, 1992a, p. 237). In contrast, the correlations between RWA and
IDEOLOGY AND PREJUDICE
103
TABLE XVIII EVIDENCE INDICATING INTERACTION BETWEEN RWA/SDO AND INTERGROUP DYNAMICS OF PREJUDICE Study
Result
RWA and/or SDO do not seem to predict intergroup bias in minimal group situations Duckitt et al. (1999) RWA and SDO did not predict intergroup bias in minimal group situation Sidanius, Pratto, & SDO did not predict evaluative bias in minimal group Mitchell (1994) experiment for low identifiers with their group but did for high identifiers Pratto, Shih, & Orton (1998) SDO did not predict intergroup bias in minimal group situations unless the minimal group distinction was made highly salient Hillin & McFarland (2000) Averaged correlations of RWA and SDO with multiple indices of minimal intergroup discrimination were .09 and .145 respectively Perreault & Bourhis (1999) Authoritarianism was nonsignificantly correlated with discriminatory behavior in a minimal intergroup situation SDO more predictive of out-group bias, the more intergroup competition over status differentials Federico (1998)
In three studies SDO was a stronger predictor of in-group bias for a high-status group when the intergroup status differential was unstable (s .55, .50, .45) than when it was stable (s .25, .20, and .27)
Authoritarianism stronger predictor of outgroup prejudice when outgroup more threatening Feldman & Stenner (1997) Both studies found RWA more strongly associated with ethnocentric Rikert (1998) attitudes under conditions of social or out-group threat Downing & Monaco (1986) Authoritarianism predicted out-group bias when there was intergroup win–lose competition but not when no competition Esses, Haddock & Effect of RWA on out-group prejudice mediated by out-group threat Zanna (1993) Duckitt (1992a) RWA with anti-Black prejudice in South Africa (r’s .53–.69) Alatemeyer (1996) RWA with anti-minority prejudice in Canada (r’s .30–.43) Lavine et al. (1999) High authoritarians more responsive to threat messages, low authoritarians more responsive to reward messages
antiminority prejudice obtained in Altemeyer’s (1996, p. 25) research in the lowintergroup-threat context of Canada for similar student samples over the same period tended to be markedly lower, ranging from .30 to .43. Two other findings also suggest that authoritarianism seems to be specifically reactive to threat from out-groups. Esses, Haddock, and Zanna (1993; Haddock, Zanna, & Esses, 1993) found that the association between RWA and out-group prejudice was mediated by the degree to which those groups were seen as threatening in-group values, customs, and traditions and violating important symbolic beliefs of the in-group. And more generally, Lavine et al. (1999) found that high authoritarians were more responsive to threat messages, whereas low authoritarians were more responsive to reward messages.
104
JOHN DUCKITT
3. Authoritarianism and SDO in Minimal Groups Both intergroup threat and power–status inequality or competition should be absent in standard minimal intergroup experiments (Brewer, 1979; Tajfel & Turner, 1979). Thus, authoritarianism and social dominance should not predict intergroup bias for minimal groups or do so only very weakly. On the other hand, they should be more predictive if threat or power and status concerns were introduced in such situations or group distinctions were made highly salient. Exactly this pattern of findings has been reported. Duckitt, Paton, Machen, and Vaughan (1999) found no association between either SDO or RWA and intergroup bias in a standard minimal intergroup situation. Similarly, Perreault and Bourhis (1999) found no correlation between authoritarianism and discriminatory intergroup behavior in a minimal group experiment. Hillin and McFarland (2000) used large student samples to investigate the relationship between RWA and SDO for a number of different indices of minimal intergroup discrimination. Although more correlations were significant than would be expected by chance (9 of 16), these correlations were weak, with the averaged correlation for RWA being .09 and that for SDO being .15. Finally, Pratto, Shih, and Orton (1988; see also Pratto, 1999, pp. 234–237) have reported particularly compelling findings. They conducted four minimal group studies and found that SDO did not predict intergroup discrimination in these situations, unless the group distinction was made highly salient or intergroup status concerns were aroused. An earlier study by Sidanius, Pratto, and Mitchell (1994) had obtained similar findings. SDO on its own did not predict intergroup evaluative bias significantly in a minimal group experiment, but SDO did interact significantly with the degree of identification with the minimal group in predicting bias. Thus, SDO predicted significantly only for high identifiers, for whom the intergroup distinction would presumably have been highly salient. The evidence from minimal group research therefore suggests that RWA and SDO will not predict intergroup bias in standard minimal group situations. However, SDO does seem to predict intergroup bias when status concerns are directly aroused or group distinctions are made highly salient in such situations, conditions which seem likely to arouse social competition over relative status.
4. SDO and Challenged Dominance Factors that influence intergroup competitiveness over relative status, power, and dominance should also influence the degree to which SDO predicts intergroup biases. For example, instability of status differentials should make high-status groups more competitive, and intensify power and dominance motivational goals for them, in order to defend and maintain their dominance. A recent study by Federico (1998) examined the effects of status instability on the degree to which
IDEOLOGY AND PREJUDICE
105
SDO predicted intergroup biases and obtained significant interactions between stability–instability and group status. Over three studies, the regression coefficients (betas) of SDO on in-group favoritism for the high-status group were more powerful when the intergroup status differential was believed to be unstable (being .55, .50, and .47) than when it was believed to be stable (being .25, .20, and .27). Overall therefore, the evidence reviewed does seem consistent with the proposition that authoritarianism and social dominance will each predict intergroup attitudes only when the intergroup relationship or situation activates threat–control or competitive–dominance motives respectively.
VIII. Overview and Conclusions To summarize, the theoretical model proposed here suggests that prejudiced intergroup attitudes result from two motivational goals in individuals, competitively driven dominance–power–superiority motivation and threat-driven social control and group defence motivation. These motivational goals are aroused by two main kinds of situational characteristics of intergroup relationships: social and intergroup threat and inequalities in or competition over power and dominance. These two motivational goals result in two kinds of prejudiced categorization schemas being applied to in-group and out-group, with each expressed in two distinct forms or dimensions of intergroup stereotypes and attitudes, discriminatory behavior, and racism. The degree to which situational intergroup dynamics arouse these two motivational goals in individuals will be modulated by the importance of these two motivational goals to individuals, with these individual differences determined by their personalities and social worldviews. The personality dimensions of social conformity and a view of the social world as dangerous and threatening influences social control and group defence motivation and its expression in authoritarian–conservative ideological attitudes. Personality tough-mindedness and a competitive-jungle social worldview influence dominance–power–superiority motivation, which is expressed in social dominance ideological attitudes. The model therefore integrates into one general framework the two most important individual difference theories of prejudice, authoritarian personality and social dominance theory, with important contemporary group-level perspectives on prejudice; specifically, those focusing on intergroup threat, inequalities in status and power, and social competition over status and power. The model is most fundamentally a motivational one, seeing prejudiced social and intergroup attitudes as emerging from powerful and basic human motivational goals. Motivation is therefore not viewed as an intrapsychic given or force, but in terms of motivational goals that are cognitively activated, or made salient, by individuals’
106
JOHN DUCKITT
and groups’ perceptions and understandings of their social and intergroup situations. In concluding, it should be noted that the model does not attempt to encompass all possible kinds of intergroup social categorizations, attitudes, and behavior, merely two particular kinds that seem powerfully associated with prejudiced intergroup attitudes and behavior. For example, the categorizations described here go beyond the simple, basic “us–them” intergroup social categorization studied in minimal groups. This kind of minimal intergroup social categorization seems to be typically associated with in-group enhancement and not out-group derogation; that is, viewing in-group members more favorably but not out-group members more negatively (Brewer, 1979; Perdue, Dovidio, Gurtman, & Tyler, 1990). In contrast, the two modes of categorization described here involve intergroup derogation or prejudice. The categorization and bias obtained in minimal intergroup situations may therefore be more appropriately viewed as a precondition for, or precursor of, prejudice. Finally, the focus in this chapter has been on how two basic cognitive-motivational dynamics can explain negative intergroup attitudes, prejudice, discrimination, and oppression. Although this was not explicitly explored here, these dynamics would also have relevance for explaining those social and intergroup attitudes and behaviors associated with resistance to oppression, discrimination, and injustice. These would arise out of the two motivational goal-schemas opposite to threat–control and competitive–dominance; that is, security–autonomy and cooperation–altruism and their expression in antiauthoritarian libertarian and egalitarian–altruistic ideological attitudes.
Acknowledgments I am grateful to Claire Wagner, Illouize du Plessis, Adrian Furnham, and Abigail Poore for collecting data reported in this chapter. I also thank Mark Zanna for his helpful comments and suggestions on an earlier draft.
References Adorno, T., Frenkel-Brunswick, E., Levinson, D., & Sanford, N. (1950). The authoritarian personality. New York: Harper. Allport, G. (1954). The nature of prejudice. Reading, MA: Addison–Wesley. Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg, Canada Univ. of Manitoba Press. Altemeyer, B. (1988). Enemies of freedom: Understanding right-wing authoritarianism. San Francisco: Jossey–Bass. Altemeyer, B. (1996). The authoritarian specter. Cambridge, MA: Harvard Univ. Press.
IDEOLOGY AND PREJUDICE
107
Altemeyer, B. (1998). The other “authoritarian personality.” In M. P. Zanna (Ed.), Advances in experimental social psychology, (Vol. 30, pp. 47–92). San Diego: Academic Press. Bahr, H., & Chadwick, B. (1974). Conservatism, racial intolerance, and attitudes toward racial assimilation among Whites and American Indians. Journal of Social Psychology, 94, 45–56 . Bierly, M. (1985). Prejudice toward contemporary outgroups as a generalised attitude. Journal of Applied Social Psychology, 15, 189–199. Billig, M. (1978). Fascists: A social psychological view of the National Front. London: Academic. Blake, R., & Mouton, J. (1984). Solving costly organizational conflicts. San Francisco: Jossey– Bass. Bond, M. (1988). Finding universal dimensions of individual variation in multi-cultural studies of values: The Rokeach and Chinese value surveys. Journal of Personality and Social Psychology, 55, 1009–1015. Bonta, B. (1997). Cooperation and competition in peaceful societies. Psychological Bulletin, 121, 299–320. Braithwaite, V. (1994). Beyond Rokeach’s equality–freedom model: Two dimensional values in a one-dimensional world. Journal of Social Issues, 50, 67–94. Brewer, M. (1979). In-group bias in the minimal intergroup situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307–324. Brewer, M., & Brown, R. (1998). Intergroup relations. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 554–594). New York: McGraw– Hill. Brewer, M., & Campbell, D. (1976). Ethnocentrism and intergroup attitudes: East African evidence. New York: Wiley. Brown, R. (1965). Social psychology. New York: Free Press. Brown, R. (1995). Prejudice: Its social psychology. Oxford, UK: Blackwell. Bullock, A. (1962). Hitler: A study in tyranny. Harmondsworth, UK: Pelican. Combs, M., & Comer, J. (1982). Race and capital punishment: A longitudinal analysis. Phylon, 12, 350–359. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. D’Andrade, R. (1992). Schemas and motivation. In R. D’Andrade & C. Strauss (Eds.), Human motives and cultural models (pp. 23–44). Cambridge: Cambridge Univ. Press. de Kiewiet, C. (1957). A history of South Africa: Social and economic. London: Oxford Univ. Press. Dijker, A. (1987). Emotional reactions to ethnic minorities. European Journal of Social Psychology, 17, 305–325. Doty, R., Peterson, B., & Winter, D. (1991). Threat and authoritarianism in the United States, 1978– 1987. Journal of Personality and Social Psychology, 61, 629–640. Downing, L., & Monaco, N. (1986). In-group/out-group bias as a function of differential contact and authoritarian personality. Journal of Social Psychology, 126, 445–452. Duckitt, J. (1989). Authoritarianism and group identification: A new view of an old construct. Political Psychology, 10, 63–84. Duckitt, J. (1992a). The social psychology of prejudice. New York: Praeger. Duckitt, J. (1992b). Patterns of prejudice: Group interests and intergroup attitudes. South African Journal of Psychology, 22, 147–156. Duckitt, J. (1993). Right-wing authoritarianism among white South African students: Its measurement and correlates. Journal of Social Psychology, 133, 553–563. Duckitt, J. (1998). Measuring ethnic attitudes in New Zealand. Unpublished manuscript. Duckitt, J. (2000). Culture, personality, and prejudice. In S. Renshon & J. Duckitt, (Eds.), Political psychology: Cultural and crosscultural foundations (pp. 89–107). New York: New York Univ. Press.
108
JOHN DUCKITT
Duckitt, J., & Mphuthing, T. (1998). Group identification and intergroup attitudes: A longitudinal analysis in South Africa. Journal of Personality and Social Psychology, 74, 80–85. Duckitt, J., Paton, S., Machen, M., & Vaughan, G. (1999). Minimal intergroup bias and real-world prejudice. Politics, Groups, and the Individual, 8, 111–121. Ellemers, N. (1993). The influence of socio-structural variables on identity management strategies. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 4). London: Wiley. Esses, V., Haddock, G., & Zanna, M. (1993). Values, stereotypes, and emotions as determinants of intergroup attitudes. In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and stereotyping (Vol. 17, pp. 137–166). San Diego, CA: Academic Press. Eysenck, H. (1954). The psychology of politics. London: Routledge & Keagan Paul. Federico, C. (1998). The interactive effects of social dominance orientation, group status, and perceived stability on favoritism for high status groups. Group Processes and Intergroup Relations, 2, 119– 143. Feldman, S., & Stenner, K. (1997). Perceived threat and authoritarianism. Political Psychology, 18, 741–770. Fink, H. (1971). Fictitious groups and the generality of prejudice: An artefact of scales without neutral categories. Psychological Reports, 29, 359–365. Fiske, S. (1998). Stereotyping, prejudice, and discrimination. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.), The handbook of social psychology, (4th ed., Vol. 2, pp. 357–411). New York: McGraw– Hill. Fiske, S., Glick, P., Cuddy, A., & Xu, P. (1999). Ambivalent stereotypes predicted by social structure: Status and competition predict competence and warmth. Unpublished manuscript. Fiske, S., Xu, J., Cuddy, A., & Glick, P. (1999). (Dis)respecting versus (dis)liking: Status and interdependence predict ambivalent stereotypes of competence and warmth. Journal of Social Issues, 55, 473–489. Fiske, A., Kitayama, S., Markus, H., & Nisbett, R. (1998). The cultural matrix of social psychology. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 915–981). New York: McGraw–Hill. Forsyth, D. (1980). A taxonomy of ethical ideologies. Journal of Personality and Social Psychology, 39, 175–184. Gaertner, S. (1973). Helping behavior and racial discrimination among liberals and conservatives. Journal of Personality and Social Psychology, 25, 335–341. Gibson, J. (1994). Warrior dreams: Paramilitary culture in post-Vietnam. New York: Hill and Wang. Giles, H., & Ryan, E. (1982). Prolegomena for developing a social psychological theory of language attitudes. In E. Ryan, & H. Giles (Eds.), Attitudes toward language variation (pp. 208–223). London: Edward Arnold. Glock, C., Wuthnow, R., Piliavin, J., & Spencer, M. (1975). Adolescent prejudice. New York: Harper & Row. Goertzel, T. (1987). Authoritarianism of personality and political attitudes. Journal of Social Psychology, 127, 7–18. Guelke, A. (1996). Dissecting the South African miracle. Nationalism and Ethnic Politics, 2, 141–154. Haddock, G., Zanna, M., & Esses, V. (1993). Assessing the structure of prejudiced attitudes: The case of attitudes toward homosexuals. Journal of Personality and Social Psychology, 65, 1105– 1118. Hampel, R., & Krupp, B. (1977). The cultural and political framework of prejudice in South Africa and Great Britain. Journal of Social Psychology, 103, 193–202. Hartley, E. (1946). Problems in prejudice. New York: King’s Crown Press. Hinkle, S., & Brown, R. (1990). Intergroup comparisons and social identity: Some links and lacunae. In D. Abrams, & M. Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 48–70). New York: Springer-Verlag.
IDEOLOGY AND PREJUDICE
109
Hillin, S., & McFarland, S. (2000). Authoritarianism, social dominance orientation, empathy, and favoritism in majority and minority minimal groups. Paper presented at the 23rd annual meeting of the International Society of Political Psychology, Seattle, WA. Hofstede, G. (1980). Culture’s consequences. Beverly Hills, CA: Sage. Hu, L., & Bentler, (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modelling, 6, 1–55. Hughes, A. (1975). Psychology and the political experience. Cambridge: Cambridge Univ. Press. Jackson, J. (1993). Realistic group conflict theory: A review and evaluation of the theoretical and empirical literature. The Psychological Record, 43, 395–414. Johnson, R., & Schlemmer, L. (1994). Launching democracy in South Africa. London/New Haven, CT: Yale Univ. Press. Katz, I., & Hass, R. (1988). Racial ambivalence and American value conflict: Correlational and priming studies of dual cognitive structures. Journal of Personality and Social Psychology, 55, 893–905. Kerlinger, F. (1984). Liberalism and conservatism: The nature and structure of social attitudes. Hillsdale NJ: Erlbaum. Kinder, D., & Sears, D. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life. Journal of Personality and Social Psychology, 40, 414–431. Kovel, J. (1970). White racism: A psychological history. New York: Pantheon. Lavine, H., Burgess, D., Snyder, M., Transue, J., Sullivan, J., Haney, B., & Wagner, S. (1999). Threat, authoritarianism, and voting: An investigation of personality and persuasion. Personality and Social Psychology Bulletin, 25, 337–347. Levin, S., & Sidanius, J. (1999). Social dominance and social identity in the United States and Israel: Ingroup favoritism or outgroup derogation. Political Psychology, 20, 99–126. Levin, S., Sidanius, J., & Rabinovitz, J. (2000). Ideological asymmetry or ideological symmetry? The interactive effects of status, social dominance, and system legitimacy. Paper presented at the Annual Scientific Meeting of the International Society of Political Psychology, Seattle, WA. Levin, S., Sidanius, J., Rabinovitz, J., & Federico, C. (1998). Ethnic identity, legitimizing ideologies, and social status: A matter of ideological asymmetry. Political Psychology, 19, 373–404. LeVine, R., & Campbell, D. (1972). Ethnocentrism: Theories of conflict, ethnic attitudes, and group behavior. New York: Wiley. MacCrone, I. (1937). Race attitudes in South Africa. London: Oxford Univ. Press. Maiti, S., & Mukerjee, B. (1990). A note on the distributional properties of the Joresk¨og–S¨orbom fit indices. Psychometrika, 55, 721–726. Marx, G. (1967). Protest and prejudice. New York: Harper & Row. McConahay, J., & Hough, J. (1976). Symbolic racism. Journal of Social Issues, 32, 23–45. McFarland, S. (1998). Toward a typology of prejudiced persons. Paper presented at the annual meeting of the International Society of Political Psychology, Montreal Canada. McFarland, S., & Adelson, S. (1996). An omnibus study of personality, values, and prejudice. Paper presented at the annual meeting of the International Society of Political Psychology, Vancouver, Canada. McHosky, J. (1996). Authoritarianism and ethical ideology. Journal of Social Psychology, 138, 709– 717. Meloen, J. (1983). De autoritaire reaktie in tijden van welvaart en krisis [The authoritarian response in times of prosperity and crisis]. Unpublished doctoral dissertation. University of Amsterdam. Meloen, J. (2000). The political culture of state authoritarianism. In S. Renshon & J. Duckitt (Eds.), Political psychology: Cultural and crosscultural foundations (pp. 108–127). New York: New York Univ. Press. Mercer, G., & Cairns, E. (1981). Conservatism and its relationship to general and specific ethnocentrism in Northern Ireland. British Journal of Social Psychology, 20, 13–16. Middendorp, C. (1991). Ideology in Dutch politics. Assen: The Netherlands: Van G. Corcum.
110
JOHN DUCKITT
Milburn, M., Conrad, S., Sala, F., & Carberry, S. (1995). Childhood punishment, denial, and political attitudes. Political Psychology, 16, 447–478. Mischel, W. (1968). Personality and assessment. New York: Wiley. Mitchell, M., & Sidanius, J. (1993). Group status and asymmetry in the relationship between ideology and death penalty support: A social dominance perspective. National Journal of Sociology, 7, 67–93. Moscovici, S. (1983). The phenomenon of social representations. In R. Farr & S. Moscovici (Eds.), Social representations (pp.3–69). Cambridge, UK: Cambridge Univ. Press. 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. Mynhardt, J. (1980). Prejudice among Afrikaans- and English-speaking South African students. Journal of Social Psychology, 110, 9–17. Nisbett, R., & Cohen, D. (1996). Culture of honor: The psychology of violence in the South. Boulder, CO: Westview. Peterson, B., Doty, R., & Winter, D. (1993). Authoritarianism and attitudes toward contemporary social issues. Personality and Social Psychology Bulletin, 19, 174–184. Perdue, C., Dovidio, J., Gurtman, M., & Tyler, R. (1990). Us and them: Social categorisation and the process of intergroup bias. Journal of Personality and Social Psychology, 59, 475– 486. Perreault, S., & Bourhis, R. (1999). Ethnocentrism, social identification, and discrimination. Personality and Social Psychology Bulletin, 25, 92–103. Pettigrew, T. (1958). Personality and socio-cultural factors in intergroup attitudes: A cross-national comparison. Journal of Conflict Resolution, 2, 29–42. Pettigrew, T. (1960). Social distance attitudes of South African students. Social Forces, 38, 246–253. Pettigrew, T. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65–85. Pettigrew, T., & Meertens, R. (1995). Subtle and blatant prejudice in western Europe. European Journal of Social Psychology, 25, 57–75. Poppe, E., & Linssen, H. (1999). Ingroup favouritism and the reflection of realistic dimensions of difference between national states in Central and Eastern European nationality stereotypes. British Journal of Social Psychology, 38, 85–102. Potthast, M. (1993). Confirmatory factor analysis of ordered categorical variables with large models. British Journal of Mathematical & Statistical Psychology, 46, 273–286. Pratto, F. (1999). The puzzle of continuing group inequality: Piecing together psychological, social, and cultural forces in social dominance theory. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 31, pp. 191–263). San Diego: Academic Press. Pratto, F., Shih, M., & Orton, J. (1998). Social dominance orientation and group context in implicit and in explicit group discrimination. Unpublished manuscript, University of Connecticut, Storrs, CT. Pratto, F., Sidanius, J., Stallworth, L., & Malle, B. (1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67, 741–763. Prothro, E., & Jensen, J. (1950). Interrelations of religious and ethnic attitudes in selected southern populations. Journal of Social Psychology, 34, 252–258. Rabinovitz, J. (1999). Go with the flow or fight the power? The interactive effects of social dominance orientation and perceived injustice on support for the status quo. Political Psychology, 20, 1–24. Raden, D. (1999). Is anti-Semitism currently part of an authoritarian attitude syndrome. Political Psychology, 20, 323–244. Ray, J. (1985a). Authoritarianism of the left revisited. Personality and Individual Differences, 6, 271– 272.
IDEOLOGY AND PREJUDICE
111
Ray, J. (1985b). Defective validity in the Altemeyer authoritarianism scale. Journal of Social Psychology, 125, 271–272. Ray, J., & Lovejoy, F. (1986). The generality of racial prejudice. Journal of Social Psychology, 126, 563–564. Rikert, E. (1998). Authoritarianism and economic threat: Implications for political behavior. Political Psychology, 19, 707–720. Rigdon, E. (1994). Demonstrating the effects of unmodelled random measurement error. Structural Equation Modelling: A Multidisciplinary Journal, 1, 375–380. Rokeach, M. (1973). The nature of human values. New York: Free Press. Rogers, R., & Prentice-Dunn, S. (1981). Deindividuation and anger-mediated interracial aggression: Unmasking regressive racism. Journal of Personality and Social Psychology, 41, 63–73. Ross, M. (1993). The culture of conflict. New Haven, CT: Yale Univ. Press. Sachdev, I., & Bourhis, R. (1987). Status differentials and intergroup behaviour. British Journal of Social Psychology, 17, 277–293. Sales, S. (1973). Threat as a factor in authoritarianism. Journal of Personality and Social Psychology, 28, 44–57. Sales, S., & Friend, K. (1973). Success and failure as determinants of level of authoritarianism. Behavioral Science, 18, 163–172. Saucier, G. (1994). Separating description and evaluation in the structure of personality attributes. Journal of Personality and Social Psychology, 66, 141–154. Saucier, G. (2000). Isms and the structure of social attitudes. Journal of Personality and Social Psychology, 78, 366–385. Sears, D. (1988). Symbolic racism. In P. Katz & D. Taylor (Eds.), Eliminating racism: Profiles in controversy (pp. 53–84). New York: Plenum. Shumacker, R., & Lomax, R. (1996). A beginner’s guide to structural equation modelling. Mahwah, NJ: Erlbaum. Schwartz, S. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. P. Zanna (Ed.), Advances in experimental social psychology, (Vol. 25, pp. 1–65). San Diego: Academic Press. Schwartz, S. (1996). Value priorities and behavior: Applying a theory of integrated value systems. In C. Seligman, J. Olson, & M. Zanna (Eds.), The psychology of values: The Ontario symposium (Vol. 8, pp. 1–24). Hillsdale NJ: Erlbaum. Sherif, M. (1967). Group conflict and cooperation. London: Routledge & Kegan Paul. Sidanius, J. (1993). The psychology of group conflict and the dynamics of oppression: A social dominance perspective. In S. Ivengar & W. McGuire (Eds.), Explorations in political psychology, (pp. 183–219). Durham, NC: Duke Univ. Press. Sidanius, J., & Ekehammar, B. (1976). Cognitive functioning and socio-political ideology. Scandinavian Journal of Psychology, 17, 205–216. Sidanius, J., Feshbach, S., Levin, S., & Pratto, F. (1997). The interface between ethnic and national attachment: Ethnic pluralism or ethnic dominance. Public Opinion Quarterly, 61, 102–133. Sidanius, J., & Pratto, F. (1993). The dynamics of social dominance and the inevitability of oppression. In P. Sniderman & P. Tetlock (Eds.), Prejudice, politics, and race in America today (pp. 173–211). Stanford, CA: Stanford Univ. Press. Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. Cambridge, UK: Cambridge Univ. Press. Sidanius, J., Pratto, F., & Bobo, L. (1994). Social dominance orientation and the political psychology of gender: A case of invariance? Journal of Personality and Social Psychology, 67, 998–1011. Sidanius, J., Pratto, F., & Mitchell, M. (1994). In-group identification, social dominance orientation, and differential intergroup social allocation. Journal of Social Psychology, 134, 151–167.
112
JOHN DUCKITT
Sidanius, J., Pratto, F., & Rabinovitz, J. (1994). Gender, ethnic status, and ideological asymmetry: A social dominance interpretation. Journal of Cross-Cultural Psychology, 25, 194–216. Singh, R., Choo, W., & Poh, L. (1998). Ingroup bias and fair-mindedness as strategies of selfpresentation in intergroup perception. Personality and Social Psychology Bulletin, 24, 147–162. Smith, P., Dugan, S., & Trompenaars, F. (1996). National culture and values of organizational employees: A dimensional analysis across 43 nations. Journal of Cross-Cultural Psychology, 27, 231–264. Sniderman, P., & Tetlock, P. (1986). Reflections on American racism. Journal of Social Issues, 42, 173–187. Solomon, S., Greenberg, J., & Pyszczynski, T. (1991). A terror management theory of social behavior: The psychological function of self-esteem and cultural worldviews. Advances in Experimental Social Psychology, 24, 93–159. Stangor, C., Sullivan, L., & Ford, T. (1991). Affective and cognitive determinants of prejudice. Social Cognition, 9, 59–80. Steiger, J. (1989). EzPATH: A supplementary module for SYSTAT and SYGRAPH. Evanston, IL: SYSTAT, Inc. Stephan, W. (1987). The contact hypothesis in intergroup relations. In C. Hendrick (Ed.), Group processes and intergroup relations: Review of personality and social psychology (pp. 599–638). New York: Random House. Stephan, W., Ybarra, O., Martinez, C., Schwarzwald, J., & Tur-kaspa, M. (1998). Prejudice toward immigrants to Spain and Israel: An integrated threat theory analysis. Journal of Cross-Cultural Psychology, 29, 559–576. Stone, W., Lederer, G., & Christie, R. (1993). The status of authoritarianism. In W. Stone, G. Lederer, R. Christie (Eds.), Strength and weakness: The authoritarian personality today (pp. 229–245). New York: Springer-Verlag. Strauss, C. (1992). Models and motives. In R. D’Andrade & C. Strauss (Eds.), Human motives and cultural models (pp. 1–20). Cambridge, UK: Cambridge Univ. Press. Struch, N., & Schwartz, S. (1989). Intergroup aggression: It’s predictors and distinctness from in-group bias. Journal of Personality and Social Psychology, 62, 564–576. Sumner, W. (1906). Folkways. New York: Ginn. Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup conflict. In W. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole. Tomkins, S. (1964). The polarity scale. New York: Springer-Verlag. Triandis, H., & Gelfand, M. (1998). Converging measurement of horizontal and vertical individualism and collectivism. Journal of Personality and Social Psychology, 74, 118–128. Trompenaars, F. (1993). Riding the waves of culture. London: Brealey. Turner, J., & Giles, H. (1981). Introduction. In J. Turner & H. Giles (Eds.), Intergroup behaviour, (pp. 1–32). Oxford, UK: Blackwell. Turner, J., Hogg, M., Oakes, P., Reicher, S., & Wetherell, M. (1987). Rediscovering the social group: A self-categorisation theory. Oxford, UK: Blackwell. Verkuyten, M., & Hagendoorn, L. (1998). Prejudice and self-categorisation: The variable role of authoritarianism and in-group stereotypes. Personality and Social Psychology Bulletin, 24, 99– 110. Vollebergh, W., Iedema, J., & Meeus, W. (1999). The emerging gender gap: Cultural and economic conservatism in the Netherlands: 1970–1992. Political Psychology, 20, 291–320. Weigel, R., & Howes, P. (1985). Conceptions of racial prejudice: Symbolic racism reconsidered. Journal of Social Issues, 41, 117–138. Winter, D. (2000). Power, sex, and violence: A psychological reconstruction of the 20th century and an intellectual agenda for social psychology. Political Psychology, 21, 383–404.
IDEOLOGY AND PREJUDICE
113
Wilson, G. (1970). Is there a general factor in social attitudes? Evidence from a factor analysis of the conservatism scale. British Journal of Social and Clinical Psychology, 9, 101–107. Yung, Y., & Bentler, P. (1994). Bootstrap-corrected ADF test statistics in covariance structure analysis. British Journal of Mathematical & Statistical Psychology, 47, 63–84. Zanna, M. P. (1994). On the nature of prejudice. Canadian Psychology, 35, 11–23.
This Page Intentionally Left Blank
AMBIVALENT SEXISM
Peter Glick Susan T. Fiske
I. Introduction Depending on the observer’s cynicism or romanticism, relationships between women and men constitute battle, war, and “intimate terrorism” or epitomize a loving, ideal, and necessary union (with men and women forming two halves of a whole). Both views, of course, capture a portion of the truth about gender relations and conventional gender ideologies. If one strips away the rhetoric and exposes the basic structure of gender intergroup relations, the source of these polarized views emerges; at the heart of gender relations lies a curious combination of power difference and intimate interdependence. Within social psychology, prejudice researchers have emphasized how power differences precipitate conflict and antipathy in gender relations, whereas close relationship researchers have explored interdependence in the forms of heterosexual attraction and intimacy. It would be exaggerating to claim that each subfield represents contradictory characterizations of men’s and women’s relationships (given the attention that prejudice researchers pay to contact, attraction, and friendship and that close relationship researchers pay to conflict and power within intimate relationships). Nevertheless, two na¨ıve aliens who divvied up their task of understanding Earthlings by each reading different parts of a typical social psychology text would come to quite different conclusions about relations between the sexes, if one had consumed the prejudice and intergroup chapters and the other had read the attraction and close relationships chapters. Here, we hope to begin bridging the gap between the prejudice and close relationship approaches to gender relations. We start with a theoretical account of how structural relations between the sexes generate ambivalent attitudes by each ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 33
115
C 2001 by Academic Press, Inc. Copyright All rights of reproduction in any form reserved. 0065-2601/01 $35.00
116
PETER GLICK AND SUSAN T. FISKE
sex toward the other, capturing the two poles of relations between the sexes. This approach, ambivalent sexism theory, highlights the striking coexistence of power difference and intimate interdependence between the sexes and suggests that these conditions are cross-culturally pervasive. This seemingly peculiar (though not unique) combination, we argue, creates hostile and benevolent ideologies about both men and women. These ideologies are cultural beliefs that, at a systemic level, help to legitimize conventional gender relations and roles. At the individual level, men and women adopt these beliefs to differing degrees, shaping how they perceive and treat members of each sex. Our research on gender ideologies employs two self-report inventories. The Ambivalent Sexism Inventory (ASI) measures hostile sexism (HS), sexist antipathy toward women, and benevolent sexism (BS), subjectively favorable, yet patronizing, beliefs about women. The Ambivalence toward Men Inventory (AMI) measures hostility toward men (HM), which presumes men’s greater power but expresses a resentment of it, and benevolence toward men (BM), subjectively favorable beliefs about men as protectors and providers. Although HS and BS were derived to capture men’s attitudes toward women, whereas HM and BM were designed to tap women’s attitudes toward men, members of either sex can hold the ideologies they assess. People can adopt or reject these cultural beliefs; the likelihood of acceptance depends, at least partly, on how much it flatters one’s own group (i.e., people are less likely to adopt subjectively hostile ideologies about their own sex than about the other). Although attitudes toward men appear toward the chapter’s end, we mainly focus on attitudes toward women (as tapped by the ASI), both because we have researched these attitudes more extensively and because of their importance for understanding the continuation of gender inequality. Over 15,000 respondents in 19 nations have completed the ASI, allowing empirical examination of HS and BS on national as well as individual levels of analysis. In contradiction to equating prejudice with antipathy (Allport, 1954), one of the most striking aspects of this research is the manner in which subjectively benevolent beliefs about women serve to justify, promote, and maintain gender inequality. Subjectively benevolent beliefs about subordinated groups can have far from benevolent effects. At the national level, the subjectively positive, but patronizing, BS beliefs about women (e.g., that they are pure, delicate creatures in need of men’s provision and protection) correlate with both HS and objective indicators of gender inequality. Nevertheless—because of the subjectively benevolent tone of BS and the possibility that promulgation of such views can work to women’s individual advantage (by securing rewards from men)—many women find BS attractive, even though it ultimately works to their group disadvantage by rewarding them for conventional female roles and portraying them as unsuited to high status, powerful roles. The beliefs tapped by the ASI revolve around the twin facts of gender relations, namely men’s greater structural power and intimate interdependence between the
AMBIVALENT SEXISM
117
sexes. HS justifies men’s power through sexist antipathy, whereas BS more subtly legitimates men’s power by promising women that men will take care of them within intimate relationships. At the cultural level of analysis, the evidence suggests that HS and BS are an interlocking set of beliefs that reflect a system of rewards (BS) and punishment (HS) that give women strong incentive to accept, rather than to challenge, power differences between the sexes. At the individual level of analysis, HS predicts hostility toward women who are perceived as challenging men’s power, whereas BS predicts benevolence toward women who enact conventional gender roles, helping to fulfill men’s needs. The AMI assesses beliefs about men that similarly reflect power difference and interdependence. HM taps a generalized resentment of men’s position of greater power and the abuses that can result (e.g., sexual aggression). This resentment, however, is not a form of feminist consciousness because it relates to BM, which includes beliefs in men’s need to have women care for them at home, in men as protectors and providers, and in the centrality of heterosexual romantic relationships for a woman’s fulfillment (all beliefs strongly complementary to BS). Both HM and BM probably increase with exaggerated power differences (e.g., in more sexist cultures or within more sexist marriages), which enhance women’s dependence on men as protectors and providers, giving women more to resent as well as to admire about men. Although HM itself may not justify patriarchy, it presumes men’s power as a given (now and always). In the final section, we put ambivalent attitudes toward men and women into a more general theoretical perspective on prejudice, a theory of stereotype content (Glick & Fiske, 2001; Fiske, Glick, Cuddy, & Xu, 1999a; Fiske, Xu, Cuddy, & Glick, 1999b). This approach, like ambivalent sexism theory, emphasizes the importance of structural relations between groups for understanding and predicting intergroup attitudes, emotions, and stereotypes. We suggest that ambivalent attitudes toward groups are not uncommon, with many groups (including men and women) being viewed as either warm, but incompetent (paternalistic prejudice) or competent, but cold (envious prejudice). This approach suggests that prejudice is not a unitary construct, but rather comprises qualitatively different types.
II. Attitudes toward Women: Ambivalent Sexism Theory Cross-culturally, men dominate. By the most obvious measures of societal power, such as earnings or presence in high-status roles (in business, government, and religious institutions), men rule (United Nations Development Programme, 1999). Patriarchy, if not universal, is strikingly common (Harris, 1991; Pratto, 1996). Yet, at the same time, men are highly dependent on women (as women are on men). The inappropriateness of social distance measures of prejudice for
118
PETER GLICK AND SUSAN T. FISKE
assessing sexism is telling. The most intimate, and therefore least prejudiced, social distance respondents can permit is marriage with a member of the out-group; most highly sexist men no doubt desire to conduct their most intimate relationship with a woman, yet might be loath to have one as a coworker (a less intimate social distance). That a low-status group might depend on a higher status group (e.g., that rewards trickle down depending on the whim of high-status group members) is expected. Yet, in gender relations, as Guttentag and Secord (1983) pointed out, dependence is a two-way street, yielding the curious and interesting fact that heterosexual men’s structural power coexists with a strong dependence on women as romantic partners, mothers, and homemakers, which in turn lends women some degree of dyadic power within relationships.
A. THE STRUCTURE OF GENDER RELATIONS The combination of power differences and interdependence between the sexes goes a long way toward explaining ambivalent attitudes on the part of each sex toward the other. Increasing attention addresses how the structure of group relations gives rise to legitimizing ideologies, systems of belief that justify the relative positions and relationships of social groups (Sidanius, Pratto, & Bobo, 1994). For the most part, theorists have been concerned with the justification of status and power differences (e.g., Jost & Banaji, 1994; Jost, Burgess, & Mosso, 2001; Tajfel, 1981). Research has confirmed that status differences foster beliefs in the relative superiority of high-status groups, even among members of low-status groups (see Jost & Banaji, 1994), at least in terms of “status-relevant” traits associated with competence (e.g., intellectual abilities and ambitiousness). Thus, male dominance results in stereotypes of male superiority on a status-relevant cluster of traits, often labeled in gender research as “agency” (Eagly, 1987). The belief in subordinates’ relative inferiority on competence-related dimensions suggests a sexist hostility that is consistent with the antipathy model of prejudice. Prejudice researchers have failed, however, to take account of a second dimension of structural group relations that also generates legitimizing ideologies: whether perceived relations are cooperative or competitive. Based on the antipathy model of prejudice, which presumes intergroup competition and conflict, the assumption has been that power differences occur within overtly competitive group relations (and that only with equality can there be cooperation). But convention views male–female group relations as cooperative due to interdependence between the sexes. In perceived interdependence, both dominants’ and subordinates’ interests lie in adopting subjectively benevolent, system-justifying ideologies that characterize the subordinate group as superior on a social warmth dimension (cf. Jackman, 1994; Ridgeway, 1992, 2001). Warm traits, such as nurturing and trusting, constitute Peeters and Czapinski’s (1990) “other-profitable” traits,
AMBIVALENT SEXISM
119
in contrast to the “self-profitable” traits that compose the competence dimension on which high-status groups allegedly excel. Relatedly, warm traits are also the traits of deference and subordination that, when displayed socially, cede power to those who enact traits associated with competence (Berger, 1982; Ridgeway, 2001). Thus, when dominants depend on interaction with subordinate group members, they have powerful incentives to view subordinates as possessing and to reward them for exhibiting warm traits, thereby encouraging subordinates to fulfill dominants’ needs without social conflict. This effect has been evident in research on stereotypes as legitimizing beliefs. Several researchers have noted that lower status groups tend to be seen by dominants and to see themselves as being nicer and friendlier (Jost et al., 2001; Spears & Manstead, 1989). A social identity theory framework interprets this as subordinate group members’ attempt to gain collective self-esteem by seeing themselves as superior on status-irrelevant dimensions. We agree that this is one function such self-stereotyping serves; subordinate group members have adopted warm traits for their self and group identities as a socially sanctioned way of achieving individual and collective self-esteem. Additionally, however, they do so to receive affection and potential rewards from the more powerful members of the dominant group, who are content to view subordinates as possessing greater warmth because it suits those subordinates to lower status roles and minimizes the likelihood of resistance or rebellion. Dominants gravitate toward ostensibly benevolent, patriarchal ideologies (e.g., protective attitudes toward women) to justify this arrangement, offering to govern and protect subordinate group members who acquiesce and defer to, rather than challenge, dominants’ social privileges (Jackman, 1994). In sum, power differences create intergroup antipathy and hostile ideologies of dominant group members’ superiority on competence-related traits, a fact that has been well appreciated and frequently demonstrated by prejudice researchers working in the tradition of the antipathy model of prejudice. What has not been clearly understood is how easily power differences can coexist with intimate interdependence between dominant and subordinate groups, fostering ostensibly benevolent ideologies that, though they presume superiority of dominants on competencerelated traits, cast cooperative subordinate group members as warm (and therefore compliant) and promise dominants’ protection and affection (contingent on subordinates’ compliance). Jackman (1994) has persuasively articulated and demonstrated the particularly effective form of coercion that the “iron hand in a velvet glove” can wield within gender, race, and class relations. It is within gender relations, however, that the coexistence of power differences and intimate interdependence (and therefore of both hostile and benevolent prejudices) most strongly and universally appears. To understand why this is the case, one must examine three domains of male–female relations that condition the content of both HS and BS: patriarchy, gender differentiation, and heterosexual reproduction.
120
PETER GLICK AND SUSAN T. FISKE
B. DOMAINS OF AMBIVALENT SEXISM AND THEIR STRUCTURAL FOUNDATIONS 1. Patriarchy: Dominative and Protective Both evolution and society can explain the prevalence of patriarchy across cultures. Evolutionary psychologists (Kenrick & Trost, 1993; Trivers, 1972) suggest that differences in reproductive investment have resulted in selective pressures that shaped men, in comparison to women, to be more strongly oriented toward obtaining social status and resources. Other theorists dispute the relative importance of evolutionary selection. Eagly and Woods (1999) suggest that although physical sexual dimorphism, particularly men’s greater size and upper body strength, originally helped men to establish a dominant social position, male dominance continues mainly via gendered divisions of labor (an issue explored at more length below). Industrialization has diminished the role of physical strength, allowing women to challenge male dominance more successfully. Nevertheless, role differences perpetuate via gender ideologies (e.g., gender stereotypes) that affect socialization and breed social pressures that in turn lead men and women to develop different traits and values (Eagly, 1987). Regardless of the relative contributions of evolution and society, patriarchy clearly prevails and, by some measures, remains strong even within highly egalitarian nations (e.g., consider the small number of female U.S. Senators and Chief Executive Officers). Yet patriarchy is not inevitable. More sophisticated evolutionary theorists note that “evolutionary” does not mean “immutable.” Pratto (1996) and Smuts (1996), for example, both point out the considerable cross-cultural diversity in the degree to which gender inequality exists. In relatively nonhierarchical cultures, male status striving is minimized, whereas in steeply hierarchical cultures, status striving among men is exacerbated (as is sexism toward women). Eagly’s (1987) social role analysis also suggests capacity for change; as women have increasingly worked outside the home, changing gender stereotypes now impute the competence-related traits of agency to at least some types of women (e.g., career women). Patriarchy, male structural power, has important implications for the content of hostile and benevolent gender ideologies. In attitudes toward women, the ideological manifestation of patriarchy is paternalism, the justification of male dominance. The hostile component of this ideology is dominative paternalism, the belief that men ought to have more power than women and the corresponding fear that women might manage to usurp men’s power. This attitude can be found both in the public domain (e.g., the notion that women complain too much about discrimination at work) and in the private domain (e.g., the belief that the male partner in an intimate heterosexual relationship ought to make the most important decisions). Dominative paternalism is tempered, however, by protective paternalism, the view that
AMBIVALENT SEXISM
121
men ought to protect and provide for the women on whom they depend. This too spans gender relations, both public (e.g., women ought to be rescued before men in emergencies) and private (e.g., the man of the household is the primary provider and protector of the family). 2. Gender Differentiation: Competitive and Complementary Although cultures vary considerably in the degree to which they view men and women as differing fundamentally or inhabiting separate social roles and occupations (Eagly & Wood, 1999), all cultures make social distinctions between the sexes and imbue gender identity with some importance (Harris, 1991). Arguably, gender constitutes the most fundamental group distinction; children learn gender as one of the first dimensions of social categorization (Maccoby, 1988), and gender becomes an automatic classification that profoundly affects adult interactions (Fiske, 1998). Well before physical differences make gender chronically salient, differential treatment shapes and reinforces boys’ and girls’ gender identities (e.g., parents dressing newborns to communicate gender; Fagot & Leinbach, 1993). Antipathy theories of prejudice, such as social identity theory (Tajfel & Turner, 1986), hold that group categorization produces in-group favoritism and intergroup competition, generating hostility toward out-groups. Because of their low status, subordinate groups suffer stereotypes of inferiority and incompetence. Stereotypes about women, in part, rationalize their alleged unsuitability for high-power roles (see Hoffman & Hurst, 1990). Thus, differentiating men and women has a hostile side, with men making downward comparisons to women that justify men’s power and inflate their collective self-esteem (Crocker & Luhtanen, 1990). Competitive gender differentiation includes the underlying belief that, as a group, women are ultimately inferior to men on competence-related dimensions (e.g., they generally could not win high-status roles in a fair competition or “male” forms of competence involving strength outflank “female” forms of competence involving sensitivity). Stereotypes of women, however, are not uniformly hostile; indeed, as Eagly and Mladinic (1993) have shown, women receive more favorable (overall) stereotypes than men. The favorable aspects of female stereotypes stem from men’s and women’s divergent reproductive and social roles. Eagly’s (1987) social role theory posits that gender stereotypes follow from a gendered division of labor (as well as status differences). Women are associated with domestic duties and child rearing that are viewed as requiring communal traits, such as warmth and understanding. In addition, women are underrepresented in high-status or leadership roles that require agentic traits, such as independence and self-confidence. Focusing solely on the agency or competence dimension of gender stereotypes would suggest that men are viewed more favorably, but both men and women think “women are wonderful” (Eagly & Mladinic, 1993), in that the warm, communal traits associated with women rate positively. Furthermore, conventional women’s roles
122
PETER GLICK AND SUSAN T. FISKE
complement and cooperate with men’s; women’s work in the home allows men to concentrate on their careers. This interdependence of conventional gender roles creates the subjectively benevolent attitude of complementary gender differentiation, the belief that women are the better sex, but only in ways suiting lower status, gender-conventional roles (e.g., they are nurturing, pure, and delicate). 3. Heterosexuality: Hostile and Intimate Prejudice-oriented researchers have tended to focus on the darker side of sexual relationships between men and women, such as sexual harassment and sexual violence. This approach emphasizes the feminist insight that sexuality is not independent of power, but rather can be a tool of domination. At the same time, heterosexuality undoubtedly creates one of the strongest sources of male–female interdependence and intimacy. Most men and women alike rate having a heterosexual romantic relationship as one of the most important factors in achieving happiness (Berscheid & Peplau, 1983). Long a culturally sanctioned norm, rooted in human reproduction, heterosexuality likely will remain a potent determinant of gender ideology. That dominance and violence often intertwine with heterosexual pair bonding is an unfortunate fact of life. Smuts (1996) even suggests that heterosexual pair bonding evolved as a female strategy for combating the threat of male sexual violence, which characterizes at least some mating in several primate species. By forming an exclusive alliance with one man, women potentially gain resources provided by the partner and protection from the threat of rape by other men. The male partner receives the benefits of sexual access to at least one woman and of paternity certainty (if the female partner is faithful and if he is successful at deterring other men from sexually coercing his partner). Whether or not the evolutionary argument holds water, a social psychological analysis yields the same conclusion: In a society in which sexual violence is common, women find it in their self-interests to procure male protectors by forming exclusive romantic attachments to a male partner (Jackman, 1994). Thus, far from being incompatible, at a systemic level, sexual violence and pair-bonded heterosexual intimate relationships go hand-inhand. Even within individual relationships, the two are not as incompatible as would be confortable to believe. Although pair bonding offers some measure of protection, it also puts women at risk of violence from their partners. Given their overall social dominance, men often view women proprietarily, and affection can turn to violence, particularly given suspected infidelity (Kenrick & Trost, 1993; Smuts, 1996). As constructs, power and sexuality can fuse for at least those men who evince an automatic power–sexuality association (Bargh & Raymond, 1995). This association may feed on women’s tendency to find power and status cues sexually attractive in men, a preference that can be explained through both evolutionary (Kenrick &
AMBIVALENT SEXISM
123
Trost, 1993) and social-psychological theories (Eagly & Wood, 1999). Perpetrators use traditional notions stressing wives’ obedience and fidelity to justify violence toward domestic partners (Jackman, 1994, 2001). Additionally, women’s sexuality has long lurked as dangerous because of its potential to alter the power equation, giving women, as sexual gatekeepers, the upper hand. Cross-cultural evidence suggests that men’s fearful and hostile attitudes toward seductive women or “femme fatales” is pervasive. This stereotypical subtype of dangerous, sexually alluring women who can weave a spell that not only ensnares, but emasculates, men (by making them dependent and thereby usurping their power) was evident in 94% of the 78 cultures examined by Jankowink and Ramsey (2000). In contrast, only 42% of these cultures had corresponding images of sexually dangerous men (e.g., Don Juan). Fear of the femme fatale is evident in cautionary tales ranging from ancient (e.g., Sampson and Delilah and the Sirens) to modern (e.g., contemporary films such as “Basic Instinct”). Thus, sexual attraction promotes not just intimacy, but heterosexual hostility toward women, a component of hostile sexism that fuses sex with power and expresses the belief in women’s sexuality as dangerous to men. At the same time, heterosexual romantic relationships can inspire passion, intimacy, and devotion, long romanticized (e.g., in medieval notions of chivalry and happily-ever-after fairy tales) and still idealized (e.g., in popular culture’s music, films, and television). Conventional gender ideologies are strongly heterosexist, lauding these relationships as normative, necessary, and the highest personal gratification for both men and women. Thus, heterosexual hostility has a benevolent counterpart, heterosexual intimacy, comprising the belief that heterosexual romantic relationships are essential for true happiness in life for both sexes.
III. Measuring Ambivalent Sexism: The ASI A. CONTENT OF THE ASI The ASI (see Table I) is a 22-item self-report measure consisting of two 11-item scales designed to tap the ideologies we have labeled HS and BS. Both the HS and BS scales were intended to represent the three domains hypothesized to underlie hostile and benevolent attitudes toward women: patriarchy, gender differentiation, and heterosexuality. Theoretical considerations, ideas expressed in popular culture in the United States, as well as informal interviews with men and women generated a large pool of potential ASI items. The items appear as statements to which respondents indicate their agreement or disagreement on a scale from 0 (disagree strongly) to 5 (agree strongly). Exploratory and confirmatory factor analyses whittled the initial 140-item pool down to the final scale. To control for possible acquiescence bias, we later reverse-coded three items on each
124
PETER GLICK AND SUSAN T. FISKE TABLE I THE AMBIVALENT SEXISM INVENTORY Relationships between men and women
Below is a series of statements concerning men and women and their relationships in contemporary society. Please indicate the degree to which you agree or disagree with each statement using the scale below: 0 1 2 3 4 5 disagree disagree disagree agree agree agree strongly somewhat slightly slightly somewhat strongly B(I) H B(P) H H B(I) H B(G) B(P) H H B(I) B(I) H H H B(P) H B(G) B(P) H B(G)
— 1. No matter how accomplished he is, a man is not truly complete as a person unless he has the love of a woman. — 2. Many women are actually seeking special favors, such as hiring policies that favor them over men, under the guise of asking for “equality.” — 3. In a disaster, women ought to be rescued before men. — 4. Most women interpret innocent remarks or acts as being sexist. — 5. Women are too easily offended. — 6. People are not truly happy in life without being romantically involved with a member of the other sex. — 7. Feminists are seeking for women to have more power than men. — 8. Many women have a quality of purity that few men possess. — 9. Women should be cherished and protected by men. — 10. Most women fail to appreciate fully all that men do for them. — 11. Women seek to gain power by getting control over men. — 12. Every man ought to have a woman whom he adores. — 13. Men are incomplete without women. — 14. Women exaggerate problems they have at work. — 15. Once a woman gets a man to commit to her, she usually tries to put him on a tight leash. — 16. When women lose to men in a fair competition, they typically complain about being discriminated against. — 17. A good woman should be set on a pedestal by her man. — 18. Many women get a kick out of teasing men by seeming sexually available and then refusing male advances. — 19. Women, compared to men, tend to have a superior moral sensibility. — 20. Men should be willing to sacrifice their own well being in order to provide financially for the women in their lives. — 21. Feminists are making unreasonable demands of men. — 22. Women, as compared to men, tend to have a more refined sense of culture and good taste.
Note. Copyright 1996 by Peter Glick and Susan T. Fiske. H = Hostile Sexism; B = Benevolent Sexism; P = Protective Paternalism; G = Complementary Gender Differentiation; I = Heterosexual Intimacy. Items 3, 6, 7, 13, 18, and 21 are reverse-worded in the original version of the ASI (though not in the version that appears here). Scoring: Hostile Sexism = average of items 2, 4, 5, 7, 10, 11, 14, 15, 16, 18, 21; Benevolent Sexism = average of items 1, 3, 6, 8, 9, 12, 13, 17, 19, 20, 22.
AMBIVALENT SEXISM
125
subscale, but subsequent cross-cultural research suggested that these negatively worded items do not perform well in translation.1 The nonreversed wording appears here, but the scale with reversed wording appears in Glick and Fiske (1996). The HS scale attempts to capture dominative paternalism, competitive gender differentiation, and heterosexual hostility, but both male and female respondents uniformly disagreed with potential HS items that overtly asserted these aspects of sexist antipathy. Hence, the final scale excluded such items (e.g., “Men ought to be in charge,” “Women are inferior to men,” and “Women are only useful to men as sexual objects”). Their rejection fits progress in addressing the most overt sexism, documented by the steady decline in scores on the Attitudes toward Women Scale (AWS; Spence & Helmreich, 1972), a measure of traditional versus egalitarian attitudes about women’s roles, from the 1970s through the 1990s (Twenge, 1997). Thus, HS items had to be relatively subtle to tap contemporary sexist antipathy in the United States, where the scale was developed. In contrast, presumably because benevolent ideologies do not normally seem a form of prejudice and, indeed, may seem polite, kind, and desirable, BS items did not have to be tempered to bow to contemporary political correctness. Thus, BS items more directly tap the intended domains. Protective paternalism is evident in items suggesting that women deserve greater protection than men and ought to be set on a pedestal; complementary gender differentiation comes out in items that would not seem out of place in a medieval tome on chivalry, asserting that women are more pure, morally sensitive, and refined than men; heterosexual intimacy emerges clearly in items stating that men are incomplete unless they have the love of a woman. With respect to complementary gender differentiation, items stressing women’s communal qualities (e.g., women are more nurturant than men) did not load highly on the scale because these stereotypes are strongly endorsed by sexists and nonsexists alike, perhaps because they reflect a social reality shaped by strong and pervasive prescriptions for women to exhibit these warm traits (Eagly, 1987; Rudman, 1998; Rudman & Glick, 1999).
B. THE ASI COMPARED TO OTHER SEXISM MEASURES Older sexism measures have concentrated on assessing individual differences on a traditional versus egalitarian orientation toward gender roles. The most prominent of these is the AWS, which assesses attitudes about women’s roles. More recently, partly as the result of respondents’ increasing reluctance to endorse highly traditional opinions about the proper roles of women, newer scales have attempted 1
No evidence of acquiesence bias was found. The correlation of HS and BS was similar whether reverse-worded were included or excluded.
126
PETER GLICK AND SUSAN T. FISKE
to measure more subtle, contemporary sexism. These include the Modern Sexism (MS; Swim, Aikin, Hall, & Hunter, 1995) and the Neo-sexism (NS; Tougas, Brown, Beaton, & Joly, 1995) scales. Inspired by on the concept of Modern Racism (McConahay, 1986), the MS and NS scales assess a modern form of sexism which denies that women continue to face discrimination and so opposes affirmative action policies designed to redress discrimination. These scales focus more on political attitudes and less on relationships between men and women. Other scales do not attempt to distinguish between antiegalitarian attitudes that are associated with subjectively positive and those that imply subjectively negative views of women. In concordance with the antipathy model of prejudice, traditional gender beliefs have presumably represented negative attitudes toward women. Eagly and Mladinic (1989) questioned this assumption, showing that the AWS was not correlated with overall evaluative attitudes toward women. In other words, traditionalists did not evince more negative attitudes about women as a group than did egalitarians. Although others’ research has found relationships between measures of sex-role traditionalism and overall attitudes toward women (Swim & Cohen, 1997), correlations are modest and most likely explained by the focus on hostile aspects of sexism (see Glick & Fiske, 1997). Given our approach, the lack of a strong relationship between traditional beliefs and antipathy toward women is not surprising. The AWS, as its subtitle explicitly notes, measures beliefs about “the rights and roles of women.” Although many items reflect a hostile reaction against perceived changes in women’s roles (e.g., “Women should worry less about their rights and more about becoming good wives and mothers”), we have noted that traditional attitudes about women’s roles can coincide with subjectively favorable views of women (e.g., as kind and understanding). The two newer sexism scales, MS and NS, are directed more at political attitudes about gender equality and aim to assess a subtle form of sexism that, like Modern Racism, masquerades as egalitarianism. High scorers on these scales are not true egalitarians, as demonstrated by the correlation of MS with overt sexism scales, such as the AWS (Glick & Fiske, 1996; Swim & Cohen, 1997). The MS scale predicts negative affect toward women in general and toward feminists in particular. The ASI shares some similarities with older scales, such as the AWS, in that it assesses traditional versus egalitarian views about gender relations and with the newer scales (MS and NS) in that the HS (though not the BS) scale uses less reactive items suited to the current political climate. The ASI is distinctive, however, in that it divides sexist views into subjectively hostile and benevolent components (i.e., those that express positive versus negative affect toward women). Furthermore, whereas the MS and NS scales focus on political attitudes about whether gender equality has been achieved and whether affirmative action programs are warranted, the ASI assesses attitudes that span the personal as well as political. The ASI covers concerns about gender relations not only in the public realm (e.g., gender discrimination at work), but also in intimate relationships. The ASI, as detailed
AMBIVALENT SEXISM
127
above, explicitly aims to cover the areas specified by our theoretical approach, including issues related to power, gender roles, and sexuality.
IV. Initial Validation of the ASI in the United States Six studies with over 2000 male and female respondents initially validated the ASI in the United States (Glick & Fiske, 1996). Most of the samples were restricted to college undergraduates, but two small community samples (with older, working adults) also participated. These studies established the factor structure of the scale and provided preliminary support for the convergent and discriminant validity of the ASI.
A. FACTOR STRUCTURE OF THE ASI As theoretically derived, the ASI was to measure two distinct, yet related, aspects of sexism, HS and BS, each intended to cover three content domains dealing with power differences (paternalism), gender differentiation, and heterosexuality. Thus, the hypothesized factor structure of the scale was complex and layered, with two correlated superordinate factors (HS and BS) each with three subfactors (paternalism, gender differentiation, and heterosexuality). Exploratory factor analyses, however, failed to yield evidence of any HS subfactors. The unidimensionality of the HS scale may be due to the need for more subtle items (because, as noted previously, respondents are reluctant to endorse overtly hostile views about women) and because of the tight linkage among the hypothesized HS subdomains. Dominative paternalism and competitive gender differentiation are conceptually difficult to distinguish because men’s greater power (the basis for paternalism) is instantiated in and reinforced by gender roles and stereotypes (gender differentiation). Moreover, men’s power derives from their relative monopoly of high-status roles, and the traits that stereotypically distinguish men and women are the same traits that distinguish the powerful from the powerless. Finally, recent research demonstrates that men who are sexually hostile to women have a tight and automatic link between their conceptions of sex and of power (Bargh & Raymond, 1995; Pryor, Giedd, & Williams, 1995). Thus, we did not expect the preferred model to differentiate among HS factors. 1. Preferred and Alternative Models Based on ambivalent sexism theory and the results of exploratory factor analyses, the preferred factor model of the ASI (see Fig. 1) has separate, though potentially
128
PETER GLICK AND SUSAN T. FISKE
Fig. 1. Preferred factor model for the Ambivalent Sexism Inventory. Reprinted from Glick and Fiske (1996).
correlated, HS and BS factors, with three subfactors nested within BS. This model was tested against various alternative models using confirmatory factor analysis (CFA) in LISREL 8.0 (J¨oreskog & S¨orbom, 1993). CFA has distinct advantages for testing the factor structure of this model as, unlike exploratory factor analysis, it allows (a) specification of theoretically relevant models whose fit can be tested; (b) models that are layered, such that there are superordinate factors and subfactors; and (c) the comparison of alternative models to determine which has the best fit. Analyses using CFA can restrict models to provide a stringent test of a hypothesized structure, seeing how well this reproduces the observed correlations between variables (i.e., responses to individual ASI items). That is, the hypothesized model can greatly risk falsification by restricting items to load only on specific, theoretically derived factors and seeing how well this fits the observed interitem correlations. This, combined with the ability of CFA to test the relative fit of alternative models, makes it an excellent tool for examining the ASI’s construct validity. The preferred model was pitted against alternative models of the ASI’s factor structure in five U.S. samples. The one-factor alternative, in which all items load on a single sexism factor, tests the most parsimonious model in which sexism is all of a piece, a unitary construct. A two-factor alternative (HS and BS, but no BS subfactors) tests a less restricted model than the preferred one, distinguishing between HS and BS (though allowing them to correlate), but not between the hypothesized BS subfactors. In this model, items are restricted to loading on HS or BS, but not both. The preferred model is the most restricted (i.e., the most stringent). Like the two-factor model, it includes (possibly correlated) HS and BS factors, but also adds BS subfactors and restricts individual items to loading on either HS or one of the BS subfactors (with no items allowed to load on more than one subfactor). The primary measure of model fit is the Goodness of Fit Index (GFI), which varies from 0 to 1 (with values close to 1 indicating good fit). Two large U.S.
AMBIVALENT SEXISM
129
samples (more than 800 participants each) compared the preferred factor model to the one-factor and two-factor alternatives; GFIs for the preferred model averaged .93 and were consistently higher than the alternative models, with significantly better fit according to the -square difference test (Bollen, 1989). The superior fit of the preferred model replicated in several smaller samples (even though these samples were smaller than the 300-participant minimum recommended for these factor analyses). The two large samples were of sufficient size to examine the fit of the model separately for women and men, which yielded GFIs of .92 and .91, indicating similar factor structure for women and men. 2. Correlation of HS and BS Each of the factor analyses left HS and BS free to correlate and, in each case, showed positive factor correlations, ranging from .37 to .74. These positive correlations fit our hypothesis that, though separable, HS and BS both are forms of sexism. Correlations of HS and BS raw scale scores, like the factor correlations, were consistently positive and statistically significant for undergraduate women and men, for whom correlations were typically similar and ranged from .31 to .57. Among two small community samples, the HS–BS correlation remained strong among women (averaging .54), but disappeared for men, with correlations in both samples being slightly, but trivially (and nonsignificantly) negative for men (−.12 and −.15). The lower correlation among a more diverse male sample is an issue to which we return below. Reliabilities for raw scores on HS and BS were acceptable across six U.S. samples. The reliability of the unidimensional HS scale tends to be (not surprisingly) stronger (␣s ranged from 0.80 to 0.92) than the reliability of the multidimensional BS scale (␣s ranged from 0.73 to 0.85). 3. Subfactors and Item Loadings Median CFA item loadings for five U.S. samples appear in Table II, along with the range of loadings for each item across the samples. Gist wording indicates the items. Some items were reverse-worded in some samples (see note at bottom of Table II). Keep in mind that the preferred factor model was a stringent one in which items could load on only one factor, either HS or one of the BS subfactors, with loadings on all other factors set to zero. Item loadings tended to be consistently strong across samples, although items that were reverse-worded tended toward somewhat poorer loadings. 4. Summary In general, the initial factor analytic results offered good preliminary support for the construct validity of the ASI. The preferred model (HS and BS with
130
PETER GLICK AND SUSAN T. FISKE TABLE II FACTOR LOADINGS FOR ASI ITEMS Factor loadings across five U.S. samples
Scale items
Median
Range Hostile Sexism
Women exaggerate problems at work Women are too easily offended Most women interpret innocent remarks as sexist When women lose fairly, they claim discriminationa Women seek special favors under guise of equality Feminists are making unreasonable demandsb Feminists are seeking more power than menb Women seek power by gaining control over men Many women tease men sexuallyb Once man commits, she puts him on a tight leash Women fail to appreciate all men do for them
.73 .69 .69 .66 .70 .50 .56 .69 .46 .73 .66
.70–.80 .66–.81 .55–.74 .31–.77 .59–.74 .42–.75 .47–.73 .64–.72 .25–.60 .65–.81 .58–.69 Benevolent Sexism
Protective Paternalism A good woman should be set on a pedestal Women should be cherished and protected by mena Men should sacrifice to provide for women In a disaster, women should be rescued firstb
.62 .49 .69 .47
.58–.68 .28–.69 .54–.73 .33–.62
Complementary Gender Differentiation Women have a superior moral sensibility Women have a quality of purity few men possess Women have more refined sense of culture, taste
.74 .80 .71
.56–.77 .61–.82 .67–.72
Heterosexual Intimacy Every man ought to have a woman he adores Men are complete without womena Despite accomplishment, men incomplete without women People are not happy without heterosexual romancea
.64 .63 .75 .44
.55–.69 .51–.70 .66–.84 .36–.67
Note. Table adapted from Glick and Fiske (1996). a Indicates items for which reversed wording (and reversed scoring) was used in two of five samples, but which were returned to their original wording for the final version of the scale reported by Glick and Fiske (1996) due to poor loadings when reversed. b Indicates items reverse-worded (and reverse-scored) for four of five samples and on the final scale reported by Glick and Fiske (1996).
three subfactors) consistently outperformed alternative models of the ASI’s factor structure. The superior fit of the preferred and two-factor models to a onefactor model demonstrates the importance of distinguishing between HS and BS (i.e., sexism is not a unidimensional construct). The increase of fit from the two-factor to the preferred model suggests that the subdomains specified by
AMBIVALENT SEXISM
131
ambivalent sexism theory (paternalism, gender differentiation, and heterosexuality) are tapped by the BS scale. The fit of the preferred model was similar for men and women, suggesting that the structure of sexist beliefs does not vary according to respondents’ gender. And, in support of the notion that HS and BS are two sides of the same sexist coin, HS and BS consistently showed a positive factor correlation. The only problematic aspect of the ASI’s structure was the failure to find the theoretically expected HS subfactors. We argue that this may be due to both the necessity of using more subtle, contemporary items to assess HS as well as the strong conceptual links among the hypothesized HS subdomains, connected by a concern with power differences between women and men. In the current climate of gender politics, judging from the content of the items that emerged on the scale, men high in HS appear to be reacting against challenges to male power (expressing the fear that men’s power is eroding) and to conventional gender roles in both public (work) and private (romantic) relationships with women. Even though the HS scale is factorially unidimensional, themes related to the hypothesized subdomains are nonetheless embedded in the scale: Dominative paternalism emerges in antifeminism and the general notion that women are seeking power over men; competitive gender differentiation shows in the idea that women cannot handle true competition with men, want special favors, and cry discrimination when they lose in a fair fight; and heterosexual hostility appears in the characterization of many women as “sexual teases” and the belief that women seek to control men within heterosexual romantic relationships.
B. SEX DIFFERENCES IN HS AND BS SCORES For previous sexism measures, the most consistent finding is that men score higher in sexism than women. We expected to replicate this finding for HS: Even if women show some degree of system justification, they would still find it in their self- and group interests to reject HS. In contrast, because BS represents a subjectively positive attitude and implies certain benefits for women, most notably men’s protection and provision, it is likely to be a more palatable form of system justification for women, one that may be perceived by many women as being in their interest to accept. In each of the six U.S. samples that initially validated the ASI, men score significantly higher than women on HS. The same gender gap, with men scoring significantly higher, was evident for BS scores in five of the six samples, but for all samples the gap in BS scores was significantly smaller. In other words, even though men consistently scored higher than women on both HS and BS, men’s and women’s scores on BS were consistently closer together than were their scores on HS, supporting the idea that women are more prone to reject HS than BS.
132
PETER GLICK AND SUSAN T. FISKE
C. CONVERGENT AND DISCRIMINANT VALIDITY OF THE HS AND BS SCALES One U.S. sample (Glick & Fiske, 1996, 1997) was part of a large pretest in which male and female undergraduates completed a variety of other sexism measures: the AWS, Old-Fashioned Sexism scale (OFS; Swim et al., 1995), MS, and Rape Myth Acceptance (RMA; Burt, 1980) scales. Both the AWS and OFS are more overt measures of sexism that have moderately correlated with more subtle measures of sexism, such as MS. Additionally, participants completed Paulhus’s (1988) Balanced Inventory of Desirable Responding (BIDR), a measure of socially desirable response tendencies. Because sexism scales have typically been based on the antipathy model of prejudice, we expected HS to exhibit convergent validity, but as BS has generally been ignored, we expected the BS scale, once its relationship with HS was partialed out, to show discriminant validity (i.e., weak or nonsignificant correlations with other sexism measures). Because gender relations are a politically and socially sensitive issue, perhaps especially among college students, we included the BIDR to see whether ASI scores were strongly negatively related to socially desirable response tendencies. Although HS and BS each showed no correlation with the Self-Deception subscale of the BIDR (r s = −.04 and .04, respectively, ns), they showed modestly negative correlations with the Impression Management (IM) subscale (r s = −.29 and −.26, respectively, ps < .05). Thus, among college students, agreeing with sexist items, whether hostile or benevolent, was a more socially undesirable response. Although these correlations are significant, they are modest. Further, controlling for the IM subscale of the BIDR did not affect the correlation of HS or BS with other sexism measures. Correlations of the ASI with the other sexism scales were generally similar for men and women. These correlations revealed the expected convergent validity for the HS scale, which correlated strongly and significantly with each of the other sexism measures. The correlations of HS with AWS (scored so that high scores equal more traditional views), MS, and RMA ranged from .61 to .68; the weakest correlation was with OFS (.48). Partialing out the relationships of HS to IM or to BS did not affect these correlations. In contrast, BS showed weaker (though still significant) zero-order correlations with the other measures of sexism. Correlations with AWS, MS, and RMA ranged from .32 to .40; again, the weakest correlation occurred with OFS (.24). As with HS, partialing out IM did not affect the correlations of BS with the other scales (i.e., they were not due to desirable responding), but they did disappear when the relationship of BS and HS was controlled (each dropping to near zero and nonsignificance). Similar results establishing the BS scale’s discriminant validity have emerged in correlations of the ASI with NS. In two student and one community sample in England, Masser and Abrams (1999) found that although HS and BS both had significant zero-order correlations with NS, the relationship of BS to NS disappeared
AMBIVALENT SEXISM
133
after partialing out HS (whereas the relationship of HS and NS remained significant after controlling for BS). Similarly, in a community sample of over 1000 Spanish men, Exp´osito, Moya, and Glick (1998) found that the BS–NS correlation dropped from a zero-order value of .36 to .17 when HS was controlled. In short, BS seems to be a distinct form of sexism that is not directly assessed by other sexism scales. Although the correlation of HS and BS with impression management is a concern, it did not affect how the scales correlated with other measures of sexism. Further, the modest correlations with socially desirable responding may well be weaker in other samples because college campuses (and particularly the University of Massachusetts at Amherst, where the study was conducted) are noted for their special sensitivity to gender politics. Nevertheless, researchers may want to include the IM scale of the BIDR when using the ASI so they can partial out socially desirable response tendencies. In any event, socially desirable responding did not affect the tendency of HS to show convergent and BS to show divergent validity with other sexism measures.
V. Cross-Cultural Validation of the ASI Ambivalent sexism theory posits that HS and BS arise due to social and biological factors that are common across human cultures: patriarchy, gender differentiation, and heterosexuality. Thus, the constructs of HS and BS ought to generalize across a variety of cultures. Cross-cultural psychologists refer to universal phenomena as etics, whereas culturally specific processes are labeled as emics (Triandis & Mar´ın, 1983). Even if, at the construct level, HS and BS are etics, the ASI scales might be culturally specific measures of these constructs as they were initially developed and validated only in the United States. To assess the cross-cultural validity of the HS and BS scales, multiple collaborators (Glick et al., 2000) administered the ASI to over 15,000 respondents in 19 nations. The countries in which the ASI was administered are culturally and economically diverse and included representation from all major geographical areas: North America (United States), the Caribbean and South America (Brazil, Chile, Colombia, and Cuba), Western Europe (Belgium, England, Germany, Italy, The Netherlands, Portugal, and Spain), Africa (Botswana, Nigeria, and South Africa), the Middle East (Turkey), Asia (Japan and South Korea), and the Pacific (Australia). Samples within each country were not representative, with the exception of The Netherlands. In most cases, the samples consisted mainly of college undergraduates, though working adults also participated in many countries. In non-English-speaking samples, the ASI was translated and back-translated to check on the appropriateness of item wording. Because items that were originally reverse-worded (three HS and three BS items)
134
PETER GLICK AND SUSAN T. FISKE
often proved difficult to translate and had poor factor loadings, we excluded them from all subsequent analyses.
A. FACTOR STRUCTURE OF THE ASI ACROSS CULTURES If the complex factor structure of the ASI replicates across cultures, this would support not only our contention that HS and BS arise from social and biological factors that are common across human societies, but also that the ASI scales are cross-culturally appropriate measures of these constructs. The preferred factor model (HS and BS with three subfactors) once again compared favorably to alternative one-factor and two-factor models using CFA. In addition, a four-factor model (HS, Protective Paternalism, Complementary Gender Differentiation, and Heterosexual Intimacy) assessed whether the BS subfactors ought to be conceptualized as independent factors rather than nested within a superordinate BS factor (as in the preferred model). In general, the preferred model exhibited good fit to the data. In the 13 larger samples (N > 500), all GFIs were greater than .90 and generally comparable to the fit obtained in large U.S. samples. For the 6 smaller samples (N’s between 250 and 500), GFIs were all greater than .85 and similar to those obtained in smaller U.S. samples. The preferred factor model significantly outperformed each of the alternative models in almost every instance (two countries showed no significant difference between the fit of the preferred and two-factor models and one country had the preferred, two-factor, and four-factor models perform similarly). The exceptions to the preferred model’s statistical superiority were few and mostly confined to the two smallest samples (Colombia and Cuba), each of which had fewer than the recommended 300-participant minimum for computing the factor analyses. As in the United States, in the samples that were sufficiently large (over 300 participants of each sex) to compare the fit of the model for men and women, the preferred model had similar fit for men and women, especially when the HS– BS factor correlations were free to vary (GFIs ranged from .90 to .93), reflecting the tendency for HS–BS correlations to sometimes be stronger among women than among men.
B. CORRELATIONS BETWEEN HS AND BS 1. Correlations of HS and BS within Nations The tendency for respondents’ HS and BS scores to correlate positively also replicated cross-culturally. As in the United States, this correlation was occasionally stronger for women than for men (for whom it was sometimes nonsignificant). In 18 of 19 countries, the HS–BS correlation was significantly positive for women,
AMBIVALENT SEXISM
135
whereas a significantly positive correlation occurred for men in 13 of 19 nations. That the HS–BS correlation was not significant for men in some samples echoes earlier findings with men in two U.S. community samples. Furthermore, the women’s correlation was significantly stronger than the men’s correlation in 10 of the 19 national samples. Overall, averaging across the 19 samples, the HS–BS correlation for men (average r = .23) was significantly smaller than the correlation for women (average r = .37), t(18) = 5.02, p < .01. Why is the HS–BS correlation weaker, in many cases, for men than for women? Relatedly, why does the HS–BS correlation vary across cultures? Answers to these questions may stem from the observation that the HS–BS correlation tended to be smaller in nations showing high average HS and BS scores. Using nations as the unit of analysis (so that N = 19), we correlated average HS and BS scores with the HS–BS correlation within each sample. Men’s HS and BS averages respectively correlated −.65 ( p < .01) and −.49 ( p < .05) with men’s HS–BS correlations. Similarly, women’s HS and BS averages respectively correlated −.30 (ns) and −.52 ( p < .05) with women’s HS–BS correlations. Thus, in more sexist cultures, respondents’ HS and BS scores tended toward greater independence. This effect was replicated within samples by comparing the HS–BS correlations for high versus low scorers on HS (using median splits computed separately for men and women). A similar result occurred: Among more sexist individuals, the HS– BS correlation tended to disappear (average r = .07), whereas among less sexist individuals, the HS–BS correlation remained (average r = .26). Why are HS–BS correlations generally weaker among (a) men compared to women, (b) respondents in more compared to less sexist nations, and (c) more sexist compared to less sexist individuals? The connecting thread in each case is overall level of sexism (which is higher among men than among women). One possible explanation is that only the most egalitarian individuals recognize BS as a form of sexism and, therefore, reject it along with HS. These low-BSlow-HS scorers (rarer in more sexist nations and among men, as compared to women) would increase their sample’s HS–BS correlation. The lower HS–BS correlations among people who are high in sexism is consistent with our original intention to measure independent hostile and benevolent aspects of sexism: Sexist individuals can be predominantly hostile, benevolent, or ambivalent (both hostile and benevolent) in their attitudes toward women.2 2. Correlations of HS and BS Averages across Nations Although individuals who are high in HS may be low in BS and vice versa, we have suggested that, at the cultural or group level of analysis, HS and BS go 2
The lower HS–BS correlations observed among respondents in more (as opposed to less) sexist nations, among men as opposed to women and among less as opposed to more sexist individuals within nations, were not an artifact of less variance in HS and in BS scores. Standard deviations for both HS and BS scores were similar in all cases.
136
PETER GLICK AND SUSAN T. FISKE
hand-in-hand as complementary ideologies, providing a system of rewards and punishments that justify and maintain gender inequality. This is not to say that HS and BS are the product of a coordinated conspiracy, but simply that in highly patriarchal cultures all ideologies that effectively help to maintain the status quo, whether hostile or benevolent, will find adherents (cf. Jackman, 1994). Comparisons of HS and BS averages across cultures show that HS and BS do go hand-inhand such that when average scores on HS are relatively high in a nation, average BS scores are also high. In fact, with nation as the unit of analysis, HS and BS scores correlate extremely highly, both for men, r = .89 ( p < .01) and for women, r = .89 ( p < .01). The strength of these correlations fits our contention that, at a systemic level, HS and BS are complementary ideologies, even if at the individual level they show relative independence among those individuals who most strongly endorse them. Examining correlations at the national level of analysis also allows a test of system-justification theory (Jost & Banaji, 1994), which posits that subordinate group members tend to accept ideologies that dominants propagate. Accordingly, did men’s overall level of sexism predict women’s overall level of sexism? Men’s HS scores correlated strongly with women’s scores on HS, r = .84 ( p < .01), and on BS, r = .92 ( p < .01). Men’s BS scores also predicted women’s scores on HS, r = .84 ( p < .01), and on BS, r = .97 ( p < .01). These results support the systemjustification notion; when men are more sexist, women are more likely to embrace sexist ideologies, both hostile and benevolent. Although an alternative is that men take their lead from women’s sexism, system-justification theories presume (as do we) that dominant groups are in a better position to propagate ideologies. This assumption gains further support from closer examination of the magnitude of men’s and women’s average HS and BS scores within and between cultures.
C. AVERAGE HS AND BS SCORES ACROSS NATIONS Given that one cannot presume the majority of the countries’ samples to be representative and because the interpretation of ASI items may differ among respondents in different cultures (even when the ASI remained in English), we caution against fine-grained comparisons of national averages. However, comparisons of male and female respondents within nations are less likely to suffer from these problems. Average scores on HS and BS in the different nations appear in Figs. 2 and 3. Comparisons of men’s and women’s averages in each country replicate and extend the results of similar comparisons in our initial U.S. samples. In all nations, men scored significantly higher than women on HS (all ts > 2.55, p < .01), but the gender gap in BS scores tended, on the whole, to be smaller, nonsignificant, or even reversed (such that women in some nations scored significantly higher
AMBIVALENT SEXISM
137
Fig. 2. Hostile Sexism averages across 19 nations. Reprinted from Glick et al. (2000).
on BS than men). Two-way analyses of variance (Sex of Participant × ASI Scale: HS and BS as repeated measures) were computed for each national sample. All but three countries (Australia, Japan, and The Netherlands) showed a significant interaction such that the HS gender gap (with men scoring higher than women) was stronger (or occurred in a different direction) than the BS gender gap, all significant Fs > 13.05, p < .01. In nine countries, the BS gender gap was not significant. In six nations, men scored significantly higher than women in BS;
Fig. 3. Benevolent Sexism averages across 19 nations. Reprinted from Glick et al. (2000).
138
PETER GLICK AND SUSAN T. FISKE
however, for four of these six nations, a significant Sex of Participant × ASI Scale interaction indicated that the BS gender gap was significantly smaller than the HS gender gap. Finally, for four nations (Cuba, Botswana, Nigeria, and South Africa), women’s BS scores were significantly higher than men’s. The general trend, then, is for women, relative to men, to reject HS, but to be more accepting of BS, which is characterized by a positive tone and the promise of potential rewards for women. These findings suggest important qualifications to the system-justification interpretation of the correlation between men’s and women’s sexism averages across nations. Recall that in nations in which men scored higher on sexism, women followed suit on both HS and BS. What this correlation concealed, however, was women’s relatively stronger opposition to HS compared to BS. Closer inspection of Figs. 2 and 3 suggests that women’s resistance to HS (i.e., to accepting HS as much as men do) increases as men’s sexism increases. In contrast, women’s resistance to BS decreases as men exhibit greater sexism to the point that women endorse BS significantly more strongly than men in the most sexist nations. These tendencies support the idea that system justification that involves the acceptance of hostile ideologies about one’s own group (in this case, women’s acceptance of HS) is tempered by self and group interests. This effect is evident in the cross-national correlations between men’s average sexism scores and the HS gender gap (men’s − women’s scores): Men’s average HS (r = .61, p < .01) and average BS (r = .65, p < .01) scores both predicted a bigger gap between men’s and women’s HS averages. In contrast, ostensibly benevolent ideologies avoid this resistance because of their affectionate and protective orientation toward the subordinate group. More pernicious is the increasing attractiveness of paternalistic ideologies to subordinate group members when these benevolent ideologies coexist with widespread hostility toward their group. Thus, we argue that it is far from coincidental that the four nations in which women endorsed BS significantly more strongly than men were the very countries that were highest in HS among men. A strong negative correlation between men’s average HS scores and the gender gap in BS scores (men’s average − women’s average) across nations (r = − .75, p < .01), supports the argument that women, to some extent, embrace BS (for the protection it promises) in response to high levels of hostility among men.
D. SUMMARY Cross-cultural comparisons support the contention that HS and BS are coherent and relevant constructs across a variety of cultures, presumably due to the underlying social and biological factors hypothesized to give rise to HS and BS. Although it is premature to claim that HS and BS are human universals, they are recognizable ideologies in a geographically and culturally diverse set of nations. These data
AMBIVALENT SEXISM
139
evidenced important cross-cultural consistencies. The cross-cultural utility of the ASI as a measure of HS and BS emerged from consistent replication of its factor structure. Other U.S. findings and trends also replicated cross-culturally. That HS and BS are positively correlated, but that the correlation tends to be higher for women, generally held true in other nations. Additionally, women’s rejection of HS compared to BS (which they reject less strongly), first documented in U.S. samples, held true for almost all of the other nations examined. In most cases, this trend was even stronger than in the United States; in all nations studied, men always scored higher than women on HS, but women frequently scored equally as highly as men and, in the most sexist nations, scored significantly higher than men on BS. These data also suggest a solution to an enigma in the original U.S. data, namely why HS and BS did not correlate among men in some samples. Across the 19 nations in our study, a tendency appeared for men, compared to women, to less consistently evince a positive HS–BS correlation. This finding paralleled two trends: for the HS–BS correlation to be lower in (a) more sexist nations and (b) among the more sexist respondents in each sample. Thus, an answer to the U.S. results suggests itself: HS and BS are not positively correlated among the most sexist respondents. Respondents low in sexism may drive the positive HS–BS correlation; if many low scores recognize both HS and BS as forms of sexism and reject both ideologies, this would create a pool of low-HSlow-BS scores that increase the HS–BS correlation. It is presumably because they are generally less sexist than men that the HS–BS correlation tends to be stronger for women. The lack of an HS–BS correlation among the most sexist respondents indicates, at the individual level of analysis, greater independence between HS and BS than was found in the initial U.S. student, though not in U.S. community male, samples (Glick & Fiske, 1996). The systemic level of analysis (comparisons of averages across nations), however, showed strong support for the notion that HS and BS are complementary faces of sexism. National averages on HS and BS across 19 nations correlated .89 (for both men and women), illustrating that at the systemic level (i.e., as cultural ideologies), HS and BS are constant companions. Although individuals may adopt one of these ideologies and not the other, when one of the ideologies is strongly present within a cultural system as a whole, the other invariably seems to accompany it. This finding supports the argument that HS and BS are complementary tools of control, the stick and the carrot, that motivate women to accept a sexist system (cf. Jackman, 1994). That the stick and the carrot are not always wielded by the same individual does not diminish the pressures exerted by such a system. These results illustrate the importance of examining legitimizing ideologies at the systemic, as well as at the individual, level by making cross-cultural comparisons. Not only were there impressive cross-cultural consistencies, but many of the significant variations across cultures were systematic and predictable. The variation
140
PETER GLICK AND SUSAN T. FISKE
in HS–BS correlations between different national samples was predictable from the overall level of sexism in each nation. Consistent with system-justification theory, women’s sexism scores were higher in cultures in which men exhibited greater sexism. Yet a fuller picture emerges by considering another predictable aspect of cross-cultural variation: The gender gaps in HS and in BS scores were related to overall levels of male sexism, but in different directions. For HS, the gender gap was always statistically significant, but the degree to which men scored higher than women on HS was positively correlated with overall sexism levels—the more sexist the men in a nation (on average), the bigger the gender gap. This suggests that although women in sexist nations, relative to women in more egalitarian countries, are more likely to accept HS, this system-justifying tendency is tempered; in more sexist nations, women show increasing reluctance (when compared to men) to accepting HS. In contrast, women’s BS averages showed the opposite effect. The more sexist the men in a nation are, the more readily women embrace BS, both relative to women in less sexist nations, but also relative to men in their own nation. This trend is so strong that the gender gap in BS scores actually reversed in the most sexist nations, with women scoring significantly higher than men. This is a convincing demonstration of the insidious seductiveness of BS. Women in sexist systems show a tendency to resist HS, but to embrace BS, presumably as a form of self-defense. When faced with a high degree of HS from men, women have considerable incentive to endorse BS, seeking protection, provision, and affection from members of the dominant group. The irony is that, in highly sexist cultures, men create both the problem (sexist hostility) and the solution (sexist benevolence) so that women are forced to seek protection from members of the very group that oppresses them.
VI. Consequences of HS and BS Even if BS is related at a systemic level to HS, given its positive tone, is BS really a social problem? After all, benevolently sexist men express a willingness to take care of women, protect them, adore them, and place them on a pedestal. Despite its subjective favorability, BS is problematic according to three lines of research reviewed in this section: First, at the systemic level, BS was not only strongly related to HS, but national BS averages were also consistently (though, with only 19 countries sampled, marginally) related to objective measures of gender inequality. Although by no means conclusive, these findings suggest BS’s importance in supporting sexist systems. Second, even though BS is related to positive stereotypes of women (whereas HS predicts negative stereotypes), the positive evaluations BS elicits are selectively directed toward women who accept conventional female roles (e.g., homemakers). Third, there is evidence that women’s belief in BS disarms
AMBIVALENT SEXISM
141
their resistance to sexist acts, especially when discriminatory actions are justified by ostensibly benevolent motives (e.g., “This is for your own protection”) or occur within intimate relationships. The second and third lines of research illuminate possible mechanisms that help to explain BS’s relationship to national measures of gender inequality. These effects are particularly important because of women’s propensity to accept BS more readily than HS, making BS an insidiously effective legitimizing ideology.
A. STRUCTURAL GENDER INEQUALITY Are both BS and HS related to gender inequality? We examined this question by correlating national averages on the two ASI scales with national indices of gender equality. The United Nations Development Programme compiles statistics on national gender equality across the globe. These statistics form two general indices of gender equality. The Gender Empowerment Measure (GEM) assesses women’s, relative to men’s, participation in the economy and in politics, focusing on the relative representation of women in elite, high-status roles (percentage of administrators and managers and professional and technical workers who are women; women’s share of earned income; and percentage of parliamentary seats held by women). The larger the GEM, the more gender equality in a country’s economic and political life. The Gender-Related Development Index (GDI) is a form of the UN’s Human Development Index (HDI), which focuses on longevity (life expectancy), knowledge (adult literacy rates and years of schooling), and standard of living (purchasing power). The GDI uses the same measures as the HDI, but the score is decreased for gender inequality (e.g., women having a lower literacy rate). The greater the gender disparity, the lower the GDI relative to the HDI.3 Glick et al. (2000) correlated men’s and women’s national averages on HS and BS in 19 nations with the UN measures of gender equality. The extremely high correlations between HS and BS averages among both men and women (both rs = .89) make it impossible to pull apart the relative contributions of HS versus BS to accounting for gender inequality. Additionally, keep in mind that the N for these correlations (19 nations) is small and that the samples within each country cannot be considered representative (of course, both of these limitations make it harder to find correlations between the ASI scales and the UN measures). We expected men’s (compared to women’s) HS and BS scores to better predict inequality because the ideological commitments of the dominant group are more likely to 3
The HDI was partialed out of correlations between the ASI scales and the GDI to control for the fact that women in more highly developed nations are likely to have increased longevity, wealth, and education compared to women in less developed nations, even if there is less gender equality within the more developed nation.
142
PETER GLICK AND SUSAN T. FISKE TABLE III CORRELATIONS BETWEEN ASI AVERAGES AND NATIONAL INDICES OF GENDER EQUALITY ACROSS 19 NATIONS
ASI Scale
Gender Development Index (GDI)
HS BS
−.47∗ −.40+
HS BS
.03 −.32
Gender Empowerment Index (GEM) Men’s Averages
−.53∗ −.43+
Women’s Averages −.38+ −.42+
Note. All correlations with GDI are partial correlations controlling for overall level of human development (HDI) in each nation. Sample sizes are 19 countries for GDI correlations and 18 countries for GEM correlations (as the GEM is not available for Nigeria). Table reprinted from Glick et al. (2000). + p < .10. ∗ p < .05.
determine social structure. Nevertheless, to the extent that women accept HS and BS, they are presumably less likely to challenge men’s power or to seek highstatus roles; therefore, women’s scores might also predict inequality. Of course, women’s average sexism scores might predict inequality for another reason—they are highly correlated with men’s average scores. Correlations between the ASI scales and the UN indices, reported in Table III, show that men’s HS predicts less gender equality and men’s BS marginally does so. The marginal tendency of women’s HS and BS scores to predict the GEM (but not the GDI) is consistent with the idea that women’s acceptance of sexist ideologies lessens the likelihood that they will pursue elite roles. In patriarchal nations, however, many women who have conservative gender ideologies may be well provided for by men. As a result, women’s sexism scores may not predict the gender gap in overall standard of living (which is what the GDI assesses). That negative correlations between the ASI scales and the UN gender equality measures emerged despite the limitations of the data (small number of nations and lack of representative samples) supports the contention that HS and BS both act to legitimize and reinforce women’s subordination. In addition to the usual caution about inferring causation, the relative contribution of HS and BS to gender inequality cannot be determined because, at the systemic level, HS and BS are so inextricably intertwined. However, this latter qualification simultaneously highlights the point that HS and BS appear in tandem in patriarchal cultures as dual justifications of gender hierarchy.
143
AMBIVALENT SEXISM
B. ATTITUDES TOWARD AND STEREOTYPES ABOUT WOMEN 1. Explicit Attitudes toward and Stereotypes about Women as a Group What are the consequences of HS and BS for individual respondents’ attitudes toward women? Although we argue that HS and BS are complementary ideologies in the justification of gender inequality, we nevertheles hypothesize them to have different subjective valences in their orientation toward women. Perhaps because of these different valences, BS and HS are relatively independent (though never significantly negatively correlated) among sexists. In our cross-cultural investigations (Glick et al., 2000), an additional measure was administered to a subset of participants in 12 of the nations studied. We asked each respondent to generate up to 10 traits he or she associates with women. After generating the traits, respondents rated each one on a −3 (extremely negative) to +3 (extremely positive) scale. This procedure has the advantage of tapping participants’ spontaneous stereotypes rather than having the researcher restrict the traits to be rated. The average positivitynegativity ratings of these spontaneous stereotypes allowed us to determine whether HS and BS predict opposing valences in stereotypes of women. Table IV reports the correlation of HS and BS with participants’ average valence ratings (because results were similar for both women and men, the data reported
TABLE IV HS AND BS PREDICT THE VALENCE OF STEREOTYPES ABOUT WOMEN IN 12 NATIONS Country Chile Japan South Africa Nigeria Botswana Spain South Korea Turkey Belgium Italy The Netherlands Australia
HS
BS
−.23∗∗ −.30∗∗ −.24∗∗ −.23∗∗ −.12∗ −.43∗∗ −.15∗ −.25∗∗ −.39∗∗ −.34∗∗
.31∗∗ .30∗∗ .21∗∗ .16∗∗ .15∗∗ .22∗∗
−.13 −.34∗∗
.11+ .33∗∗ .27∗∗ .21∗∗ .17∗ .21∗
N 735 531 499 376 366 280 219 219 209 215 151 137
Note. All correlations are partial correlations, controlling for the (typically positive) relationship between the HS and BS subscales. Table adapted from Glick et al. (2000). + p < .10. ∗ p < .05. ∗∗ p < .01.
144
PETER GLICK AND SUSAN T. FISKE
here do not distinguish male and female participants). Due to the moderately positive correlation between HS and BS, each ASI scale was partialed out of the other to reveal their opposing relations to valences of attitudes toward women. The results were strikingly consistent: All of the relationships were in the predicted direction, with only 2 of 24 correlations failing to reach significance. Overall, the results strongly support ambivalent sexism theory. HS is a subjectively negative and BS a subjectively positive (though still sexist) orientation toward women. 2. Explicit Attitudes toward Subtypes of Women Although BS is associated with favorable stereotypes of women, we have argued that such subjectively positive orientations can, nevertheless, promote inequality because these affectionate attitudes may be used to reward women for adopting conventional roles. Glick, Diebold, Bailey-Werner, and Zhu (1997) examined whether HS and BS are typically directed at different subtypes of women based on the argument that sexist hostility is reserved for women who are perceived as challenging male power, whereas sexist benevolence is bestowed upon women who are viewed as fulfilling men’s needs. Stereotype researchers have debated the question of the level at which most stereotyping takes place. Fiske and Neuberg’s (1990) continuum model of impression formation suggests that for highly familiar groups, individuals are typically classified according to subtypes. Rather than viewing an individual as being simply a woman, we are more likely to classify her as, for example, a career woman, a feminist, or a mother. Research (e.g., Deaux, Winton, Crowley, & Lewis, 1985; Six & Eckes, 1991) has demonstrated several recognizable female subtypes that range along two general dimensions that might be characterized as traditionality (e.g., homemakers vs career women) and sexual attractiveness (e.g., butch vs babe). If HS and BS target different types of women, this would explain how these attitudes can have opposing valences yet themselves be positively correlated or independent rather than negatively correlated. Possibly, the positive and negative attitudes expressed when respondents rated their spontaneous stereotypes of women (in the cross-cultural data just reported) referred not so much to different traits all women were viewed as possessing, but to different types of women who embody these traits. Glick et al. (1997) explored this possibility by having undergraduate men and women spontaneously generate the subtypes into which they normally classify women. Most of the types participants listed, consistent with past research, fit along a traditional (e.g., mother and Daddy’s girl) versus nontraditional (e.g., feminist and career woman) or a sexually attractive (e.g., babe and cheerleader) versus unattractive (e.g., unattractive and butch) dimension. Participants then rated the first eight subtypes they generated on measures adapted from Esses, Haddock, and Zanna (1993): overall evaluation (on a feeling thermometer), positive and
AMBIVALENT SEXISM
145
negative personality trait attributions (separate ratings), positive and negative affect (separate ratings) elicited when the participant interacted with women of this type, and the degree to which women of this type usually facilitated or blocked the attainment of the participant’s goals (symbolic beliefs). People who endorse both HS and BS were predicted to have more polarized subtypes of women, loving some and despising others. Indices of the polarization of participants’ views of their eight subtypes were formed by computing the variance of ratings across that participant’s subtypes. For example, the variance in negative affect ratings of eight subtypes of women made by a single participant was computed. This score, for each participant, was the dependent variable, with higher scores indicating more polarized evaluations of the different subtypes. For men, on all but one of the measures (symbolic beliefs), polarization was significantly greater for those who scored high on both HS and BS. These results replicated for women only on the negative trait and negative affect measures. Although it was the simultaneous endorsement of both HS and BS that predicted polarized views of men’s’ spontaneously generated subtypes of women, HS and BS ought separately to predict the valence of affect directed toward different female subtypes. For each participant, his or her eight subtypes yielded average ratings (e.g., the participant’s ratings of negative affect toward each of eight subtypes were averaged). For men, HS and BS predicted opposing evaluations of women, with HS accounting for more hostile (or less positive) affect, stereotypes, and symbolic beliefs and BS predicting more positive (or less negative) evaluations (although not all correlations reached significance for BS). These relationships did not emerge for women. Although the first study demonstrated that men who simultaneously endorse both HS and BS have polarized views of different female subtypes, it did not distinguish which specific subtypes elicited positive and which elicited negative affect and stereotypes. In a second study, Glick et al. (1997) examined attitudes toward two specific female subtypes that embody the traditional–nontraditional dimension of classification, homemakers and career women (we avoided feminists because antifeminism is an explicit theme of the HS scale and, therefore, demonstrating that HS predicts negative attitudes toward feminists would be merely tautological). This study tested the idea that HS is directed at nontraditional and BS at traditional types of women. If this view is correct, HS ought to exclusively predict attitudes toward career women (as these women are a threat to male dominance), and BS ought to exclusively predict attitudes toward homemakers (as these women fulfill a traditional female role). Male and female undergraduates indicated their overall evaluations of the two types of women. Additionally they generated traits, affect, and symbolic beliefs (goals that were thwarted or facilitated) they associated with each type. An overall evaluation thermometer rating and ratings of the negativity versus positivity of the traits, affect, and symbolic beliefs participants generated about each type formed the dependent measures.
146
PETER GLICK AND SUSAN T. FISKE
For both male and female participants (though more consistently for male participants), HS predicted more negative views of career women, whereas BS (with one exception) did not correlate with evaluations of this type of woman. The one exception concerned symbolic beliefs, for which men’s BS scores predicted more positive views of career women. This may reflect the expectation among these college men that, in the future, their wives would also have careers, extending some of their protective attitudes toward this type of women. In contrast, men’s attitudes toward homemakers were exclusively predicted by BS and not HS scores (though only the stereotype and symbolic belief correlations reached statistical significance). For women, evaluations of homemakers were not predicted by either scale. These results at least partially support the notion that HS and BS generally target, for men at least, different types of women. Thus, HS and BS can easily be reconciled in sexists’ minds (as they direct positive and negative affect toward different types of women) and act in concert to help maintain gender inequality by providing incentive (both reward for conforming and punishment for failing to conform) for women to stick to conventional gender roles. 3. Implicit Attitudes toward and Stereotypes about Women Although in its infancy, some research to date explores the relationship between the ASI scales and implicit gender beliefs. Both HS and BS successfully predict implicit associations of (a) men with power and women with weakness (Rudman, Greenwald, & McGhee, in press) and (b) men with high status (e.g., executive, boss) and women with low status (e.g., subordinate and aid) roles (Rudman & Kilianski, 2000). These data (based on partial correlations controlling the relationship between HS and BS) support our contention that both BS and HS justify patriarchy because both scales predict implicit associations of men with power and status and of women with weakness and low status. Thus, implicit associations confirm the notion that BS is indeed a form of prejudice. Implicit associations also support the claim that HS represents a subjectively negative orientation toward women. Rudman et al. (in press) found that HS negatively covaried with associating women to warmth and men to coldness. Thus, hostile sexists’ implicit associations (like their explicit attitudes) not only cast women as weak, but begrudge them even the favorable feminine stereotype of warmth. Whether BS is associated with subjectively favorable implicit attitudes about women is unclear. Although Rudman and Kilianski (in press) found some evidence that BS predicted simultaneous implicit stereotypes of men as more agentic and women as more communal, this effect did not replicate in another study (Rudman & Glick, in press). Furthermore, Rudman et al. (in press) did not find a significant correlation between BS and implicit associations of women with warmth, and Rudman and Kilianski (2000) found that BS predicted an implicit overall preference for men as opposed to women.
AMBIVALENT SEXISM
147
Subjectively neutral or even negative implicit attitudes may or may not underpin the explicit, subjectively positive tone of BS (though this would fit BS as a fundamentally sexist attitude). Because explicit stereotypes of women’s warmth are so widely accepted (recall that the majority of respondents so strongly endorsed potential BS items stressing women’s communal traits that the items were excluded from the scale), researchers who wish to find a relationship between BS and positive implicit stereotypes may have greater success if they examine implicit associations of women with purity, innocence, and delicacy. Additionally, implicit stereotyping studies (Rudman & Glick, in press; Rudman et al., in press; Rudman & Kilianski, 2000) suggest that researchers will have better success using HS and BS to predict implicit associations of men with traits that directly express social dominance (e.g., hard, dominant, and powerful) or status (e.g., leader, boss, and authority) rather than traits typically used to measure agency (e.g., independent, competitive, and self-sufficient). This may reflect changes in the attribution of agency traits to women as a result of their increased presence in paid employment (Cejka & Eagly, 1999). Spence and Buckner (2000) have argued that women today are permitted, even encouraged, to be independent and self-sufficient (i.e., agentic), so long as they do not go so far as to exhibit social dominance (see also Rudman & Glick, in press). Qualifications aside, the consistent message of the implicit stereotyping-ASI research conducted thus far is that both HS and BS independently predict implicit associations of men, and not women, with status and power. Finally, Carpenter (2000) has performed a series of implicit attitude studies that, even though they did not include the ASI scales, provide data consistent with our research on attitudes toward specific female subtypes. Carpenter found that men’s overall implicit attitudes toward women were slightly more favorable than their attitudes toward men, but that this effect was strongest when the target was women (versus men) or mothers (versus fathers) and was mitigated when the target was female (versus male) leaders. Presuming that the default “women” construes women in traditional roles, these results fit our contention that (for sexist men) affection toward women is contingent on whether they embrace conventional female roles.
C. BS DISARMS WOMEN’S RESISTANCE TO MEN’S SEXISM Another mechanism by which BS may promote gender inequality is by increasing women’s tolerance for acts of discrimination against them. Paternalistic justifications obscure the seriousness of discriminatory acts and deflate hostility that victims of discrimination might otherwise feel (cf. Jackman, 1994). HS alone would be likely to elicit recalcitrance and rebellion among women, giving them little to lose by resisting men. BS offers women a way of negotiating a sexist
148
PETER GLICK AND SUSAN T. FISKE
system individually without really challenging the structure of the system as a whole. Wright (2001) and Ellemers (2001) have shown that even when just a few members of subordinate groups (tokens) are allowed some degree of social mobility, collective action that challenges the system quickly becomes a disfavored option for the disadvantaged, who instead focus on individual strategies for advancement. BS similarly may deflate collective resistance. BS ideology promotes individual advancement for women who pair themselves with powerful men—protectors and providers who act as social and economic elevators for women (who stand in the background relinquishing the controls to men). Thus, BS is inimical to feminism and even to the direct empowerment of women through the attainment of high-powered jobs. Rudman and Heppen (2000) show that women who implicitly believe that “someday my prince will come” are less ambitious in their career aspirations than women who do not engage in such nonconscious romantic fantasizing. Relatedly, Moya, Exp´osito, and Casado (1999) found higher BS scores among Spanish women who did not have paid employment than among those who did. These researchers also examined women’s reactions to discriminatory scenarios (e.g., losing a promotion to a less qualified man and having their husband forbid them to go out at night). Identical acts of discrimination were viewed as less serious when the perpetrator expressed a benevolent justification (e.g., claiming that the action was protective) as opposed to a hostile one. Additionally, women who scored higher in benevolent sexism were more willing to excuse not only benevolently justified discrimination by nonintimate men (e.g., a boss), but also overtly hostile discrimination by a husband. The latter effect held true only for women without paid employment, suggesting that women who are highly dependent on male partners are prone to forgive even hostile acts, perhaps reinterpreting them as a sign of the husband’s passionate attachment. These findings support the contention that women who endorse benevolent sexism are more likely to tolerate, rather than to challenge, sexist behavior by men (especially when the sexist’s motivation can be interpreted as being protective) and demonstrate specific ways in which belief in BS not only acts as an ideological justification of women’s subordination, but actively promotes women’s acquiescence to patriarchy. Although the results reported in this section make the argument that BS promotes gender inequality, BS does not have only unambiguously negative effects on women’s position in society. Drawing the line between protective policies that are patronizing versus those that promote greater equality is not a simple matter. For example, is affirmative action a necessary step to redress a history of gender inequality or might it be a form of paternalism (implying that women cannot make it on their own merits)? Different motives can sometimes lead to similar policy positions, making strange bedfellows. For example, some feminists and conservative Christians have found themselves side-by-side on calls to ban pornography. Although BS may, overall, be associated with greater inequality, BS is also
AMBIVALENT SEXISM
149
sometimes related to what seem like progressive attitudes. Glick and Fiske (1996) found that once its relationship with HS was partialed out, BS was mildly positively related to recognition that women do face discrimination. Wiener, Hurt, Russell, Mannen, and Gasper (1997) found that people high in HS who read transcripts describing cases of sexual harassment (of women) were more likely to dismiss the behavior as less severe and pervasive. In contrast, belief in BS buffered some of these effects for participants who were also high in HS, and BS was associated with perceptions that the harassment would have a greater negative effect on the victim’s work performance. Furthermore, belief in BS does not inevitably lead women to have a greater tolerance for sexist behavior. At least in terms of nonverbal reactions, women high in BS showed greater disgust when exposed to demeaning sexist jokes (LaFrance & Woodzicka, 1998). Disentangling the various ways in which BS may sometimes seem to promote versus diminish women’s rights is a task for further research. On balance, however, current evidence suggests that BS generally does more harm than good when it comes to women’s position in society.
VII. Where Is the Ambivalence Is sexism toward women ambivalent? The ASI was so named initially because we conceived of sexism as having both a hostile and a benevolent component and thus fulfilling at least the literal meaning of ambivalence, “both valences.” The opposing valences HS and BS predict in attitudes toward women support the idea that sexism has both a positive and a negative component and suggest that an individual who scores high on both scales can be characterized as ambivalent toward women. Glick and Fiske’s (1996) initial findings of moderately positive correlations between HS and BS (at the individual level of analysis), however, posed some difficulty for the notion of ambivalent sexism. On the one hand, the positive HS–BS correlation suggested that both were sexist ideologies and that ambivalence (simultaneous endorsement of HS and BS) was likely to be common among sexist individuals. On the other hand, ambivalence is usually conceptualized as a state of conflicted, dissonant feelings that individuals attempt to avoid, and therefore (within a sample of individuals) the positive and negative components of ambivalent attitudes ought to be negatively correlated or independent, not positively related. Subsequent findings that the HS–BS correlation was significantly weaker or nonsignificant among sexist individuals (Glick et al., 2000) is more in line with what ambivalence theorists would expect (see Thompson, Zanna, & Griffin, 1995; Katz, 1981). Nevertheless, even among more sexist individuals, the HS–BS correlation has never been significantly negative, and a modestly positive correlation often remains.
150
PETER GLICK AND SUSAN T. FISKE
The initial findings of positive HS–BS correlations led us to consider how individuals might reconcile HS and BS so that, although they represent opposing valences in attitudes toward women, simultaneous endorsement of both sets of attitudes would not necessarily be psychologically dissonant or conflicting. As noted, Glick et al. (1997) showed that the two sets of beliefs are directed at different types of women (with HS being directed at nontraditional women and BS at traditional women). If so, HS and BS can be characterized as having different attitude objects, and hence both may be endorsed without entailing conflicted feelings toward the same attitude object. Polarized reactions toward an attitude object may result from ambivalent (i.e., conflicting) feelings. In the case of sexism, however, Glick et al. (1997) suggested that polarized subtyping of women avoids the experience of conflicted feelings. By dividing women into subjectively good and bad subtypes, HS and BS can aim at different targets. In other words, subtyping can be a mechanism for avoiding dissonance or psychological conflict, as it is psychologically consistent to love some women (e.g., homemakers) and despise others (e.g., feminists). If this strategy is successful and the individual never experiences conflicted feelings toward the same attitude object, one might question whether it is appropriate to label the observed polarization a sign of ambivalence. We think it unlikely, however, that the division of women into polarized subtypes captures the whole story of sexist men’s attitudes toward women. Ideologically, HS and BS may target different prototypical subtypes of women, but sexist attitudes are also played out in relationships with individual women, toward co-workers, friends, relatives, and partners in intimate heterosexual relationships. Individual women are likely to be complex and may not fit cleanly into stereotypical categories, creating ample opportunities for ambivalence on the part of men who are high in both HS and BS. Imagine the conflicted feelings such a man may have toward a homemaker who espouses feminist views or toward a sexually attractive woman who is also a high-powered lawyer. Certainly, ample evidence indicates that ambivalence of an explosive sort is not uncommon in heterosexual romantic relationships. Most dramatic is the typical pattern that occurs in domestic abuse, in which violence is followed by an apologetic honeymoon period, an alternation between violent sexist antipathy and exaggerated benevolence (Straus, Gelles, & Steinmetz, 1980). Although we have not yet documented whether HS and BS together predict ambivalent reactions to individual women, we think such an outcome likely. Furthermore, qualitative responses from Glick et al.’s (1997) studies of attitudes toward female subtypes provides suggestive evidence that sexist men’s stereotypes are not uniformly negative or positive. Recall that HS predicted negative attitudes toward career women, especially for male participants. Although the attitudes of the most sexist male respondents toward career women were negative overall, examination of the content of their stereotypes and emotions toward this type of
AMBIVALENT SEXISM
151
woman suggests an underlying ambivalence. In particular, like nonsexists, hostile sexists viewed career women as highly competent, professional, and hard-working. What distinguished the responses of men high in HS was that they also viewed career women as selfish, greedy, cold, and aggressive; additionally, they reported emotions of fear, envy, intimidation, and competitiveness toward these women. Thus, although men who scored high in HS had relatively more negative attitudes toward career women, the actual content of their attitudes came across as ambivalent, combining a grudging respect for these women’s competence with an interpersonal dislike and hostile emotions. MacDonald and Zanna (1998) conducted detailed investigations of attitudes toward feminists (the other most frequently mentioned nontraditional female type) and found similar responses—admiration for their competence, but a lack of affection or liking. The traits and emotions associated with homemakers suggest a precisely opposite view. Both sexists and nonsexists were generally positive, stressing communal and nurturing traits along with positive emotions, such as warmth and trust. The traits associated with competence, however, were conspicuous by their absence in the spontaneously generated stereotypes of homemakers. This fits with others’ observations that conventional stereotypes of women suggest a lack of respect for women’s intelligence and competence, despite warm feelings and liking for women (Eagly & Mladinic, 1993; MacDonald & Zanna, 1998). If sexists tend to view powerful women as competent, but cold, and conventional women as warm, but not competent, then they can be characterized as being ambivalent toward each of these categories of women. This kind of attitude conflict has been labeled “cross-dimensional” ambivalence (MacDonald & Zanna, 1998; Thompson et al., 1995), in reference to the conflicting valence of attitudes toward the same object on different dimensions of evaluation (e.g., competence versus warmth). We suggest, then, that sexist ambivalence is a complex phenomenon that sexist men deal with in several ways. First, some of the ambivalent feelings about women are resolved by directing positive and negative affect toward different types of women. Even so, sexist men’s evaluations of prominent female subtypes can embody a significant amount of cross-dimensional ambivalence, such that powerful or nontraditional women, despite being evaluated negatively overall by sexists, may be respected as competent, though disliked. In contrast, traditional or subordinate women, despite being viewed with an overall glow of affection, may be simultaneously perceived as incompetent. In the former case, negative affect may be increased by envy and fear of competent women, whereas in the latter, affection may be enhanced by paternalistic feelings of superiority (e.g., a woman’s attractiveness, to some men, may increase if they perceive her as not too smart). Both processes (polarized subtyping and cross-dimensional ambivalence) can be strategies for resolving ambivalence. Each involves a form of splitting: The former resolves conflicted attitudes by dividing the attitude object (women) into multiple attitude objects (subtypes), whereas the latter involves distinguishing different
152
PETER GLICK AND SUSAN T. FISKE
dimensions of evaluation when considering a specific subtype. The result avoids conflicted attitudes about the same female type on one specific dimension (the most dissonant form of ambivalence). Nevertheless, because interaction occurs not with stereotypes, but with individual women who may combine features of different categories, sexist men are likely to experience ambivalent feelings toward individual women, perhaps especially those women with whom they have the most intimate relationships.
VIII. Attitudes toward Men Although our work has mainly focused on understanding men’s attitudes toward women, we have not ignored women’s attitudes toward men, which are undoubtedly also affected by power and status differences that simultaneously coexist with intimate interdependence between the sexes. Thus, the same content dimensions—patriarchy, gender differentiation, and heterosexuality—that are important in attitudes about women ought to be evident in attitudes toward men. Whereas interdependence can be symmetrical, however, power differences, by definition, are not. Women’s attitudes toward men come from the perspective of a lower status group, altering the nature of the hostility and the benevolence that they may feel toward men.
A. THEORETICAL PERSPECTIVE Social identity theory (Tajfel, 1981), system-justification theory (Jost & Banaji, 1994), and earlier notions of traits thought to develop due to being the target of prejudice (Allport, 1954) all, in some manner, suggest that members of lower status groups experience conflicting feelings toward members of high-status groups, a mixture of negative feelings such as envy, fear, and resentment coupled with system-justifying attributions of competence and intelligence. In short, members of lower status groups are likely to both resent and admire the powerful. Women ought to prove no exception to this general rule, but their attitudes are also undoubtedly affected by their interdependence with men. As we have noted, this interdependence can puncture some of the resentment felt toward men, as male power can be used to protect and to provide for women (thus many women have a stake in supporting and enhancing the power of men who serve as their protectors and providers). At the same time, dependence on men can heighten resentment when women find themselves having to defer to men on a daily basis and to put up with various displays of male power, not only from the stranger who engages in sexual harassment on the street, but the intimate partner who assumes power over
AMBIVALENT SEXISM
153
important decisions. The end result is that women are likely to experience both hostility toward men (HM) and benevolence toward men (BM). Both sides of these ambivalent attitudes play out within each of the domains previously identified as important to male–female relations. 1. Patriarchy: Resentment of Paternalism and Maternalism Patriarchy, male dominance, is bound to create some degree of resentment among women. Relative deprivation can occur at the group level when one group sees another as experiencing better outcomes (Davis, 1959; Runciman, 1966). Social identity theory (Tajfel, 1981; Tajfel & Turner, 1986) posits that at least one strategy by which members of lower status groups may attempt to raise their status is a competitive orientation that includes derogation of other, more powerful groups (typically on dimensions not related to status). Similarly, Allport (1954) suggested that victims of prejudice sometimes respond with heightened prejudice toward dominants, though they may be forced to express their antipathy in subtle and covert ways. All of these theories suggest that a major component of HM is resentment of paternalism, resentment of men’s greater power and the attitudes of superiority that accompany it. At the same time, men’s paternalism includes the benevolently sexist notions that women ought to be provided for and protected. Our cross-cultural investigations using the ASI demonstrated that women often endorse, or at least do not as strongly reject, this aspect of paternalism. Because women may rely on men for social status (e.g., marriage as a social elevator) and economic security (even if built on a system where men’s monopolization of power and resources creates this necessity), women’s resentment of paternalism may be tempered by benevolence toward men. Although women may largely be denied public forms of power in highly paternalistic societies, they are, in exchange, ceded some degree of control and authority over domestic matters. Thus, BM may include a maternalism that is complementary to paternalism. Like its paternalistic counterpart, maternalism reflects an attitude that is affectionate and recognizes a dependence on the other sex, but is patronizing in that the other sex is presumed to be incapable in some domain and in need of help. Specifically, maternalism is the assumption that men would be lost without women to perform domestic chores and care for them at home coupled with the belief that women therefore ought to manage domestic life. Men may find it strategically effective to promote these beliefs and to avoid competence at domestic skills as long as women are willing to fulfill these tasks. Women may accept maternalism because they can gain self- and group esteem by taking pride in their superior domestic abilities. The net effect is to keep women oriented toward the home and children (and therefore out of competition with men for the status and resources women might earn in the wider job market).
154
PETER GLICK AND SUSAN T. FISKE
2. Gender Differentiation: Compensatory and Complementary Like men’s, women’s attitudes toward the other sex result from the gendered division of labor and gender roles that most societies promote. The drive to differentiate one’s group from other groups may be just as strong (or sometimes stronger) in low-status groups (Brewer, Manzi, & Shaw, 1993; Mullen, Brown, & Smith, 1992). Social reality limits the areas in which members of subordinate groups can attain positive differentiation (i.e., stereotypes that favor the group). Within a traditional system that views women as less competent when it comes to the abilities required for high-status roles, one way for women to achieve self and group esteem is to derogate men as being not truly superior. Thus, HM includes a compensatory gender differentiation that characterizes men as being not as capable as they might appear, particularly from a “behind the scenes” domestic viewpoint (e.g., the notion that men act like babies when they are sick). This attitude is related to maternalism, but is overtly hostile (i.e., it does not imply that men’s domestic incompetence is endearing and should elicit women’s help). The benevolent side of gender differentiation in women’s attitudes toward men emphasizes those areas in which men allegedly excel to the benefit of women. Thus, complementary gender differentiation toward men includes women’s recognition of men as protectors and providers who take risks for the women in their lives. This aspect of BM echoes the protective notions evident in BS. Kilianski and Rudman (1998) found that women typically rated positively a man described as a benevolent sexist, suggesting that protective men are admired by many women. Additionally, system-justification approaches suggest that dominant groups profit on status-relevant dimensions (e.g., the stereotype of men’s greater agency).
3. Heterosexuality: Hostile and Intimate Men’s greater power sometimes emerges through the sexual objectification of women and through control over important decisions in heterosexual relationships. Most women experience at least small acts of sexual harassment (e.g., catcalls) and, especially as women have entered the paid workforce in greater numbers, sexual harassment has become a salient issue in the workplace (see Swim & Stangor, 1998). We noted earlier that men who harass tend to have automatic associations between sex and power. Further, sexual aggression by men is more likely in cultures that promote gender inequality (Smuts, 1996). Many women are likely, then, to both fear and resent male sexuality, expressing an aspect of HM we label heterosexual hostility, especially to the degree that they see men’s sexual behavior as an exercise of power ( just as men high in HS fear women’s use of sexual attractiveness as a means of gaining the upper hand with men). Like heterosexual men, heterosexual women are dependent on members of the other sex as romantic partners. If anything, women are likely to receive stronger
AMBIVALENT SEXISM
155
social messages about the importance of being in a happy, long-term romantic relationship if they are to be considered normal and successful in life. Conventional gender ideology makes women the relationship experts and suggests that they need a husband and children to find fulfillment. The influence of these views, arguably, is evident in Josephs, Markus, and Tafarodi’s (1992) finding that women’s selfidentity is more bound up with relations to others (i.e., interdependence with others) compared to that of men, whose self-identity revolves more strongly around areas of competence and work (i.e., their ability to be independent). Thus a belief in the importance of heterosexual intimacy should be a benevolent component of women’s attitudes toward men, just as this is a component of men’s benevolent attitudes toward women.
B. CONTENT OF THE AMI The Ambivalence toward Men Inventory (AMI; Glick & Fiske, 1999) (see Table V) is a 20-item self-report measure comprising two 10-item scales assessing HM and BM. Each of the AMI subscales aims to cover the three domains concerning power differences, gender differentiation, and heterosexuality that, as with the HS and BS scales, we hypothesized to be important content dimensions of hostile and benevolent beliefs. A large pool of potential AMI items resulted from theoretical considerations, supplemented by notes from a group of women discussing what they most and least liked about men. As with the ASI, all AMI items were phrased as statements with which respondents indicated agreement or disagreement on a scale from 0 (strongly disagree) to 5 (strongly agree). The initial 133 AMI items shrank to the final 20-item version based on exploratory and confirmatory factor analyses. The HM scale theoretically taps resentment of paternalism (e.g., men will always strive to maintain control over women), compensatory gender differentiation (e.g., men are like children), and heterosexual hostility (e.g., most men are potential harassers and want to dominate in relationships with women). The BM scale was designed to measure paternalism’s female counterpart, maternalism (e.g., women ought to take care of men at home), complementary gender differentiation (e.g., men are better in emergency situations than women), and heterosexual intimacy (e.g., every woman needs a romantic relationship with a man to be truly happy).
C. VALIDATION OF THE AMI SCALE IN THE UNITED STATES The AMI scale was administered to three U.S. samples, two comprising undergraduate college students and one a community sample of older adults (Glick & Fiske, 1999). Analyses then explored the factor structure of the AMI and
156
PETER GLICK AND SUSAN T. FISKE TABLE V THE AMBIVALENCE TOWARD MEN INVENTORY Relationships between men and women
Below is a series of statements concerning men and women and their relationships in contemporary society. Please indicate the degree to which you agree or disagree with each statement using the scale below: 0 1 2 3 4 5 disagree disagree disagree agree agree agree strongly somewhat slightly slightly somewhat strongly B(M) H(S) B(G) H(S) B(S) H(G) B(S) H(G) H(P) B(M) H(P)
B(S) B(G) H(S) H(P) B(S) H(G) B(G) H(S) B(M)
— 1. Even if both members of a couple work, the woman ought to be more attentive to taking care of her man at home. — 2. A man who is sexually attracted to a woman typically has no morals about doing whatever it takes to get her in bed. — 3. Men are less likely to fall apart in emergencies than women are. — 4. When men act to “help” women, they are often trying to prove they are better than women. — 5. Every woman needs a male partner who will cherish her. — 6. Men would be lost in this world if women weren’t there to guide them. — 7. A woman will never be truly fulfilled in life if she doesn’t have a committed, long-term relationship with a man. — 8. Men act like babies when they are sick. — 9. Men will always fight to have greater control in society than women. — 10. Men are mainly useful to provide financial security for women. — 11. Even men who claim to be sensitive to women’s rights really want a traditional relationship at home, with the woman performing most of the housekeeping and child care. — 12. Every woman ought to have a man she adores. — 13. Men are more willing to put themselves in danger to protect others. — 14. Men usually try to dominate conversations when talking to women. — 15. Most men pay lip service to equality for women, but can’t handle having a woman as an equal. — 16. Women are incomplete without men. — 17. When it comes down to it, most men are really like children. — 18. Men are more willing to take risks than women. — 19. Most men sexually harass women, even if only in subtle ways, once they are in a position of power over them. — 20. Women ought to take care of their men at home, because men would fall apart if they had to fend for themselves.
Note. Copyright, 1999 by Peter Glick and Susan T. Fiske. The HM subscales are indicated by the following notation: H(P) = Resentment of Paternalism; H(G) = Compensatory Gender Differentiation; H(S) = Hetereosexual Hostility. The BM subscales are indicated by the following notation: B(M) = Maternalism; B(G) = Complementary Gender Differentiation; B(S) = Heterosexual Intimacy. Scoring: Hostility toward Men = average of items 2, 4, 6, 8, 9, 11, 14, 15, 17, 19; Benevolence toward Men = average of items 1, 3, 5, 7, 10, 12, 13, 16, 18, 20.
AMBIVALENT SEXISM
157
tested the convergent, discriminant, and predictive validity of the HM and BM scales. 1. Factor Structure of the AMI Exploratory factor analyses suggested that the theoretically preferred model in which the AMI consists of separate, but related, HM and BM scales, each with three subfactors, might prove to be viable. This preferred model is depicted in Fig. 4. Why did the three desired subfactors emerge for both the hostile and benevolent AMI factors, whereas the desired subfactors only emerged for the BS scale of the ASI? Recall that one difficulty in developing the HS scale, which may have blurred the lines among the subfactors, was the need to develop more subtle items because hostility toward women is a politically sensitive topic (whereas benevolently sexist beliefs about women are more widely acceptable). These difficulties were not evident in creating AMI items—respondents do not seem to monitor as vigorously overtly hostile beliefs about men, the dominant group. The preferred factor model contrasted with the simpler (and therefore less restricted) alternatives, a two-factor (HM and BS, no subfactors) and a one-factor model, using CFA in LISREL 8.0 (J¨oreskog & S¨orbom, 1993). The preferred theoretical model had the best fit in all three samples, even though two of the samples were relatively small (Ns = 205 and 168) for conducting factor analysis. In the one large sample (N = 489), the GFI for the preferred model was .93 and comparisons of the fit of the model for women and men revealed similar factor structures (GFI = .92). Based on three samples, median loadings of individual items (with gist wording) appear in Table VI (note that each item was allowed to load on only one subfactor of HM or of BM). Factor correlations for HM and BM were positive in all three samples, ranging from .45 to .54. Similarly, correlations for raw scores on HM and BM were positive and moderate in all samples, ranging from .34 to .39 for women and .33 to .65 for men. Why might a positive correlation emerge between HM and BM? Ironically, women who embrace conventional gender roles may exhibit the most hostility as well as benevolence toward men. Such women are more likely to experience,
Fig. 4. Preferred model for the Ambivalence toward Men Inventory.
158
PETER GLICK AND SUSAN T. FISKE TABLE VI FACTOR LOADINGS FOR AMI ITEMS Factor loadings across three U.S. samples
Scale Items
Median
Range Hostility toward Men
Resentment of Paternalism Men will always fight for greater control in society Even sensitive men want traditional relationships Men pay lip service to equality, but can’t handle it
.61 .65 .65
.56–.64 .56–.71 .54–.68
Compensatory Gender Differentiation Most men are really like children Men would be lost without women to guide them Men act like babies when they are sick
.77 .59 .68
.70–.78 .59–.64 .65–.73
Heterosexual Hostility When in positions of power, men sexually harass women Men have no morals in what they will do to get sex When men “help” women it is to prove they are better Men usually try to dominate conversations with women
.61 .60 .66 .56
.53–.67 .58–.67 .62–.67 .54–.63 Benevolence toward Men
Maternalism Even if both work, woman should take care of man at home Men are mainly useful to provide financial security for women Women should take care of man at home, else he’d fall apart
.73
.69–.80
.56
.55–.69
.60
.39–.68
Complementary Gender Differentiation Men are more willingly to risk self to protect others Men are more willing to take risks than women Men are less likely to fall apart in emergencies
.77 .71 .68
.61–.80 .70–.76 .55–.72
Heterosexual Intimacy Every woman needs a male partner who will cherish her Woman is never fulfilled without romantic relationship Every woman ought to have a man she adores Women are incomplete without men
.71 .73 .69 .65
.69–.83 .66–.75 .63–.75 .63–.70
Note. Table adapted from Glick and Fiske (1999).
on a daily basis, male partners who exert greater power, and their dependence on men may actually increase resentment of men ( just as, for example, hostile sexists resent the power women’s sexual attractiveness may have over them). For example, dependence on men reinforces women’s low-power position and makes them more susceptible to being victims of proprietary violence (Jackman, 1994, 2001). Consider Virginia Woolf’s sense of liberation, not simply economically,
AMBIVALENT SEXISM
159
but psychologically, as a result of an inheritance that provided her with enough to live on. “I need not hate any man; he cannot hurt me,” Woolf wrote, “I need not flatter any man; he has nothing to give me” (19291981, p. 38). Woolf’s sentiments nicely illustrate the connection between HM and BM for women who are highly dependent on men. 2. Sex Differences in HM and BM Scores For all three samples, 2 (Sex of Participant) × 2 (AMI Scale: HM and BM) repeated-measures analyses of variance on AMI scores revealed significant Sex of Participant × AMI Scale interactions, all Fs > 26.15, p < .01, such that women scored considerably higher on HM than men did, whereas the sex difference in BM scores was reversed. Averaged across three samples, women’s, as compared to men’s, greater HM (M = 2.33 for women, M = 1.86 for men) mirrors men’s tendency to score higher than women on HS. Men consistently scored higher than women on BM (averaged across three samples, M = 1.85 for men and M = 1.47 for women). Most striking, however, is the difference between women’s endorsement of HM and BM: Whereas men’s averages on HS and BS are generally close, women (on average) evinced more hostile than benevolent attitudes toward men. In general, BM scores were extremely low, even for male respondents. 3. Convergent and Discriminant Validity of the HM and BM Scales Of the relatively few published measures of attitudes toward men, two of the more psychometrically sound and comprehensive scales were compared to the AMI. Both labeled the Attitudes toward Men Scale (AMS; Downs & Engleson, 1982; Iazzo, 1983), the scales aim to measure traditional attitudes toward men across a variety of domains (e.g., work, sexuality, and parenting), but unlike the AMI, do not distinguish between positive and negative attitudes. Respondents in one undergraduate sample (Glick & Fiske, 1999) completed the AMI and the ASI along with the two attitudes toward men scales, which we label AMS1 (Downs & Engleson, 1982) and AMS2 (Iazzo, 1983). If the AMI measures conventional attitudes about men, HM and BM both ought to correlate with AMS1 and AMS2. In addition, we expected HM and BM to correlate with the two ASI scales, HS and BS. In particular, BM and BS ought to be strongly related, given that both emphasize the complementarity of the sexes (e.g., the idea that men and women complete each other), although BM focuses on the implications for attitudes toward men and BS assesses these beliefs as attitudes toward women. BM ought also to correlate positively with HS because hostile sexists tend to view men as superior to women and would therefore be likely to exhibit benevolence toward men. HM, on the other hand, was not predicted to correlate strongly with HS or BS, given that it does not justify men’s power (and interdependence between the sexes),
160
PETER GLICK AND SUSAN T. FISKE TABLE VII CORRELATIONS BETWEEN AMI SCALES AND OTHER MEASURES OF ATTITUDES TOWARD MEN AND WOMEN AMI
AMS1
AMS2
ASI
HS
BS
HM BM HM controlling for BM BM controlling for HM
.72∗ .82∗ .46∗ .64∗
Female respondents (n = 114) .61∗ .56∗ .49∗ .68∗ .79∗ .74∗ .34∗ .15 .09 .45∗ .67∗ .63∗
.47∗ .61∗ .14 .45∗
HM BM HM controlling for BM BM controlling for HM
.56∗ .57∗ .43∗ .48∗
Male respondents (n = 65) .67∗ .36∗ .17 .48∗ .80∗ .65∗ .60∗ .11 −.13 .34∗ .76∗ .64∗
.48∗ .74∗ .32∗ .67∗
Note. Table reprinted from Glick and Fiske (1999). ∗ p < .01.
representing instead a reaction against men’s power. Nevertheless, the presumption pervading HM items is that men will always remain in command, so it does not, we think, reflect a feminist call for change. In other words, HM is resentment that assumes the continuation of power differences between men and women, making it a traditional attitude that may correlate moderately with HS and BS. Correlations appear in Table VII. HM and BM alike strongly correlated with both AMS1 and AMS2, with correlations remaining significant after partialing each of the AMI scales from the other (to control for the positive correlation between HM and BM). The correlations with AMS1 and AMS2 suggest that HM and BM have convergent validity with existing measures. That HM and BM each independently relate to AMS1 and AMS2 suggests that the prior scales include both benevolent and hostile attitudes toward men, which HM and BM aim to distinguish. HM and BM were also positively correlated with HS and BS (with one exception: for men, HM was unrelated to HS). The correlation of HM with both HS and BS generally dropped to nonsignificance when the influence of BM was partialed out. Because HM taps a resentment of men’s power, its relative independence from HS and BS (once its relationship with BM was controlled) was expected, as was the convergence of BM with the ASI scales. The correlations of BM with the ASI scales are strong, but not so extreme as to suggest that they measure identical constructs.
4. Predictive Validity of the AMI In one undergraduate and one community sample (Glick & Fiske, 1999), respondents generated 10 traits they associate with men and then rated the positivity
AMBIVALENT SEXISM
161
or negativity of each on a scale from a −2 (extremely negative) to +2 (extremely positive). We averaged these valence ratings for each respondent to see whether HM and BM actually predict opposing valences in attitudes toward men. For both male and female respondents, HM predicted negative and BM predicted positive stereotypes about men (with 7 of 8 correlations reaching statistical significance across 2 samples). Neither of the two AMS scales consistently predicted the valence of stereotypes (correlations with valence of stereotypes were generally weak and nonsignificant). The HS scale of the ASI was, not surprisingly, related to more positive views of men, but only the AMI scales successfully predicted both subjectively hostile and benevolent stereotypes of men.
D. CROSS-CULTURAL VALIDATION OF THE AMI Cross-cultural data for the AMI are much less extensive than for the ASI. In addition to the U.S. samples just reported, Glick, Fiske, Lameiras, Brunner, Willemsen, L´opez, Castro and Sotelo (2000) administered the AMI along with the ASI in six additional countries: Argentina, Colombia, Cuba, The Netherlands, Portugal, and Spain. With the exception of The Netherlands, which constituted a large, randomly selected national sample, the samples are relatively small for factor analysis. Nevertheless, these data offer some preliminary information about the cross-cultural generalizability of the scales and of findings in the United States. Confirmatory factor analyses using LISREL 8.0 tested the preferred model against the simpler two-factor and one-factor alternatives. The preferred model had statistically superior fit to the alternative models in all nations; however, although the GFIs were greater than .90 in both the United States and The Netherlands, GFIs for the preferred model (ranging from .76 to .80) were lower than desired in Argentina, Colombia, Cuba, Portugal, and Spain. This may be due to small sample sizes that are below the recommended minimum, especially problematic given the complexity of the preferred model (with small samples in the United States, the fit of the preferred model for both the AMI and the ASI is typically in the .80, rather than the .90, range). We cannot rule out the possibility, however, that the structure of the scale, in contrast to the ASI, may vary across cultures. Despite the somewhat poor GFIs in some countries, the reliabilities of the HM and BM scales were generally high, with ␣s ranging from 0.72 to 0.83 for HM and from 0.66 to 0.84 for BM (with all ␣s being greater than 0.80 except for the Cuban sample). Although future research will have to resolve the question of whether the factor structure of the AMI generalizes sufficiently well, because of the relative superiority of the preferred model and the acceptable reliabilities of the HM and BM subscales, it seems worthwhile to look at other AMI results in cross-cultural comparisons. HM and BM consistently correlate positively in other countries; the r across national samples averaged .57 for men and .52 for women, comparable to U.S. correlations. Also, as in the United States, HM and BM scores tended to
162
PETER GLICK AND SUSAN T. FISKE TABLE VIII AVERAGE CORRELATIONS OF AMI AND ASI SCALES IN SIX NATIONS (EXCLUDING THE UNITED STATES) Male respondents BS
Female respondents
AMI Scale
HS
HS
HM BM
.34 .53
HM BM
Controlling for the correlation between HM and BM .07 .20 .20 .27 .41 .43 .42 .56
Zero-order correlations .47 .40 .56 .54
BS
.53 .69
correlate positively with HS and BS scores for both women and men. The average of these correlations in the six countries, excluding the United States (already reported), appear in Table VIII. As in the United States, the correlations of BM, as opposed to HM, with the ASI scales were generally stronger. Partial correlations revealed a similar pattern to the U.S. data in that, after controlling the relationship of HM and BM, HM only weakly or nonsignificantly correlated with the ASI scales, whereas BM remained significantly correlated with both HS and BS. Mean scores on HM and BM across seven nations are presented in Figs. 5 and 6. A 2 (Sex of Participant) × 2 (AMI Scale: HM and BM) repeated-measures analysis of variance within each national sample revealed a significant Sex of Participant × AMI Scale interaction, all Fs > 22.44, p < .01. In each case, the interaction was of a similar type. As in the United States, women consistently scored higher than men on HM (differences in all countries were significant at p < .01), but this sex difference generally reversed for BM, such that men scored higher. The difference
Fig. 5. Hostility toward men averages across seven nations.
AMBIVALENT SEXISM
163
Fig. 6. Benevolence toward men averages across seven nations.
in men’s and women’s BM means was significant (at p < .01) in four of seven countries, but was nonsignificant in Argentina, Spain, and Colombia (in Colombia men’s and women’s BM means were virtually identical). Although the small number of countries sampled restricts any comparisons at the national level of analysis to N = 7 and therefore cross-national analyses must be viewed with caution, we expected that just as national HS and BS averages are strongly correlated with each other, national averages on the AMI scales would be strongly correlated with each other. Furthermore, we expected strong correlations of national averages on the ASI scales with national averages on the AMI scales. The reasoning for these predictions is that both HM and BM are likely to be most prevalent in the most sexist cultures. The resentful attitude toward men that HM represents is predicated on greater male power and a high degree of gender differentiation, as are the attitudes represented by BM, which assume male dominance outside the home, but that women ought to care for men within the home. Interestingly, the cross-national correlations between HM and BM averages, as well as between averages on the AMI and the ASI scales (reported in Table IX), were so strong that most were statistically significant despite being based on a set of only seven countries. That the correlations of the HM scale with HS and BS was generally lower for men makes sense in that it is women who ought to resent men (HM) if men are perceived to be more powerful (i.e., the country is more sexist).
E. SUMMARY Power differences and mutual interdependence between the sexes produce both hostile and benevolent ideologies on the part of each sex toward the other. Like the ASI, the AMI yields evidence for both hostile and benevolent ideologies on the
164
PETER GLICK AND SUSAN T. FISKE TABLE IX CORRELATIONS OF MEN’S AND WOMEN’S NATIONAL AVERAGES ON AMI AND ASI SCALES ACROSS SEVEN NATIONS AMI Scale
BM
HS
BS
HM BM
.91∗∗ —
Men’s averages .49 .77∗
.59 .82∗
HM BM
.98∗∗ —
Women’s averages .97∗∗ .93∗∗
.93∗∗ .92∗∗
∗p
< .05. < .01.
∗∗ p
dimensions that ambivalent sexism theory identifies: power differences (paternalism), gender differentiation, and heterosexuality. Although the factor structure of the AMI outperformed alternative models in all seven countries we have examined, the overall fit of the model, though high in the United States and The Netherlands, was not as high as desired in the remaining nations. Resolving whether failure to obtain a good fit in the other (mostly Spanish-speaking) countries was due to the relatively small sample sizes obtained in these nations or a difference in the structure of beliefs about men awaits further research. Nevertheless, the AMI scales had good reliability in all cases. Further, sex differences in scores and correlations between the AMI’s and ASI’s subscales were cross-culturally consistent. BM is, clearly, an ideology that is compatible with traditional gender relations and, like HS and BS, potentially a form of system justification. BM offers justification of conventional gender roles in which women take care of the home (and of their male partners) in exchange for men playing the protector and provider role. Additionally, the correlation of BM with positive views of men demonstrates that women who endorse BM exhibit some degree of out-group favoritism (which, when exhibited by subordinates toward dominants, signals system justification). Women, both in the United States and in other nations tested, however, scored considerably higher than men on HM and (in most cases) lower on BM, suggesting that their overall attitudes toward men may be more predominantly hostile than ambivalent, though the absolute level of hostility is generally low (with averages in all countries being below the HM scale’s midpoint, with the exception of Cuba). With the spread of more egalitarian gender ideologies, it may be that women in many nations are less likely to justify men’s dominance through admiration and affection for men as a group (BM). Given its content, HM (unlike BM, HS, or BS) does not directly seem to be a legitimizing ideology for traditional gender relations, but instead reflects the resentment that a powerful group can evoke (cf. Feather, 1994). As noted, we do
AMBIVALENT SEXISM
165
not believe, however, that HM is a form of feminist consciousness or correlates with social activism. Women within highly patriarchal cultures may resent the traits and behaviors that accompany men’s power, even if they believe that men ought to be in charge. HM resembles the notion that “boys will be boys” (or an exasperated “men!”). It presumes that men will always possess greater power and abuse that power, an attitude that is more likely to promote cynicism than activism. Thus, HM is likely to be higher when gender-based power differences are more extreme (i.e., within more patriarchal systems) because this is when women are more likely to be restricted by men or subjected to violence and harassment. In terms of gender ideology, the most conservative country (on both the ASI and AMI) was Cuba, and it was here that women’s scores on both BM and HM were greatest. That HM is not a form of feminism is evidenced by its positive correlation, at both the individual and national levels of analysis, with BM and with both HS and BS. Although the correlations among respondents (i.e., within samples) of HM with HS and BS tend to be weak or nonsignificant when HM’s relationship with BM is controlled, the correlations are never significantly negative. If HM were a type of feminist consciousness or a rejection of sexism, the correlation of HM with measures of sexism ought to be negative. These results were highly consistent across countries. Furthermore, the correlations of national averages on the AMI and the ASI scales (i.e., treating nation, rather than individual respondents, as the unit of analysis) were extremely strong (both for men’s and for women’s averages), suggesting that all are traditional forms of gender ideology. In other words, despite the contention of some recent social commentators, hostility toward men is not a result of feminism; rather, our preliminary data suggest that it is strongest in countries with the most traditional gender ideologies. Successful feminism ought potentially to reduce HM by promoting greater equality and undermining sexist ideologies among both genders.
IX. Development of Gender Prejudice Our empirical work on hostile and benevolent attitudes toward each sex has employed late adolescent and adult samples because the theoretical foundations of this research suggest that ambivalent attitudes are likely to appear only during and after adolescence. Glick and Hilt (2000) argued that ambivalent sexism theory can shed new light on the developmental course of gender prejudice, helping to interpret existing data on children’s, as compared to adults’, attitudes toward members of the other sex. Specifically, Glick and Hilt posit that gender relations change not just from cognitive simplicity to complexity, but from overt, univalent hostility to the more subtle and ambivalent attitudes documented in our research.
166
PETER GLICK AND SUSAN T. FISKE
Although cognitive development is, in part, responsible for the greater complexity of adult attitudes, ambivalent sexism theory highlights another important change that has received insufficient research attention—the development of interdependence between the sexes during adolescence. Further, although cognitive complexity of attitudes increases with age, Glick and Hilt argue that this does not necessarily represent a progression from “more” to “less” gender prejudice from childhood to adulthood; rather, ambivalent sexism theory suggests qualitative changes in the nature of childhood and adult prejudices about the other sex.
A. CHILDHOOD PREJUDICE Most parents will not be surprised that developmental researchers have demonstrated strong childhood gender prejudice, of the sort that fits well with the standard antipathy model (Maccoby, 1990). This prejudice is abundantly evident in children’s cognitions, affect, and behavior toward the other sex. Cognitively, children endorse strong stereotypes about each sex at a relatively young age. At 2–3 years of age, children already possess gender stereotypes concerning children’s toys and adult roles; by 5 years of age, children begin to associate agentic traits with men and communal traits with women (Ruble & Martin, 1998). These stereotypes, like children’s other cognitive structures, are overgeneralized and simplistic; they are also highly resistant to well-meaning adult intervention (Bigler, 1999). Similarly, children’s affect and behavior toward the other sex are generally antipathetic. Young children’s evaluations demonstrate not merely in-group bias, but a hostile rejection of the other sex (Powlishta, 1995). This antipathy only increases with age until children are close to adolescence (Yee & Brown, 1994). The antipathy is not one-sided; both boys and girls demonstrate overt hostility toward the other sex, though girls may manifest these feelings a bit earlier and more intensely than do boys (Maccoby, 1990). Behaviorally, childhood play is strikingly segregated, a tendency that also increases with age in early childhood, becoming so strong that 6-year-olds have 11 times as much playtime with same-sex compared to other-sex children (Maccoby & Jacklin, 1987). Maccoby and Jacklin found that segregation in play is not typically imposed by adults, but by the children themselves as they exhibit same-sex play preferences more strongly during free playtimes than when adults structure play. In short, young children’s (unlike adult’s) gender attitudes fit the antipathy model of prejudice like a glove: They exhibit overt hostility, simplistic and persistent stereotypes, and behavioral discrimination toward the other sex. Ambivalent sexism theory would predict these findings because the structure of young children’s gender relations entails no interdependence to temper their hostility with benevolent feelings. Although young children may depend on adults of the other sex, they do not depend on other-sex peers. The relative absence of strong heterosexual
AMBIVALENT SEXISM
167
romantic attraction and lack of a gendered division of labor among childhood peers means that boys and girls do not typically need to interact or to cooperate with each other (unless adults so demand). By itself, lack of interdependence might lead merely to indifference, but some of the seeds of hostile attitudes about the other sex evident among adults are sown in childhood. In particular, as early as 3 years of age, young boys (perhaps for biological as well as social reasons) already exhibit a concern for dominance and competition (Maccoby, 1990) that echoes the power difference that ambivalent sexism theory posits as central to hostility in adult gender relations. Maccoby suggests that girls show the first signs of intergender hostility because of the frustration they experience when interacting with boys, whose dominant style (e.g., grabbing what he wants) gives them the upper hand in cross-gender interaction. Girls find that their more polite style (e.g., requesting and suggesting) is ineffective in interaction with boys and begin to develop hostile attitudes toward boys and to segregate from them, with boys returning the favor on both counts. Although young boys are a long way from holding high-status roles, their interaction styles lend them greater power in cross-gender encounters, creating resentment of that power among girls (an early-childhood version of HM’s resentment of patriarchy) and fostering a contempt for girls among boys (corresponding to HS’s dominative paternalism). Feelings of antipathy toward the other sex begin at a crucial time in development when gender constancy (children’s understanding that gender is a permanent characteristic of their identity) is being established (Lutz & Ruble, 1995). Awareness of gender constancy is likely to heighten children’s desire for a positive social identity based on gender, creating a competitive drive to positively differentiate one’s own gender from the alternaive, which can be accomplished by derogating members of the other sex (Tajfel, 1981). Thus, competitive gender differentiation, which, like budding power differences, also corresponds to components of adult intergender hostility, occurs early in childhood.
B. ADOLESCENT AND ADULT AMBIVALENCE The most obvious change that occurs in adolescent gender relationships is heightened interest in members of the other sex as romantic partners, creating the interdependence that ambivalent sexism theory identifies as crucial to fostering benevolent attitudes toward the other sex. Benevolence toward the other sex ought to be most evident in dating relationships, as this is where interdependence first develops. In fact, although some evidence suggests increased cross-sex friendships during adolescence, most friendships continue to be with members of the same sex (Maccoby, 1990). Benevolent attitudes on the part of adolescent boys toward girls appear in the continued popularity of chivalrous behavior as
168
PETER GLICK AND SUSAN T. FISKE
normative in dating scripts (Rose & Frieze, 1993) and in the tendency for male college students, in most nations studied, to endorse BS about as strongly or more strongly than women (Glick & Fiske, 1996; Glick et al., 2000). In contrast, AMI results suggest much less benevolence on the part of both younger and older women toward men (Glick & Fiske, 1999). Women’s less benevolent attitude toward the other sex may reflect a difference between the manner in which, in egalitarian cultures, prejudices are expressed toward higher status as opposed to lower status groups—higher status groups may be considered fair game for criticism, whereas criticizing lower status groups may threaten an egalitarian self-image. Furthermore, given men’s more dominant and women’s more accommodating interaction styles, women may tend to be particularly frustrated in their dealings with men, creating less benevolence toward men. Recall Maccoby’s (1990) hypothesis that young girls initially withdraw from interaction with boys because of boy’s dominant interaction styles. The gender segregation of early childhood means that girls and boys face different social worlds (among peers at least) that hone very different styles of interaction (and that adults’ expectations may also reinforce). When sexual attraction leads adolescent boys and girls to interact again, girls may once again be frustrated by boys’ competitive and dominant behavior. In contrast, the more nurturant, stereotypically feminine interaction style developed by girls is relatively accommodating and cedes status and therefore is less likely to frustrate boys. The changes that begin in adolescence anticipate the continuing romantic interdependence that heterosexual men and women experience in adulthood, often institutionalized through marriage. Interdependence in adulthood expands from romance to a reliance (for many individuals) on a gendered division of labor (e.g., husband as primary provider and wife as primary caretaker of children and household). In both adolescence and adulthood, however, power differences and competitive relationships between the sexes remain, often with the stakes increasing. Power differences between men and women no longer function solely as different interaction styles, but are reinforced by differences in earnings and status of occupation. The increasing movement of women into paid employment may reduce women’s overall dependence on men, but only increase competition between the sexes, as women gain in power (and men, in a relative sense, perceive their power to be reduced).
C. SUMMARY Ambivalent sexism theory suggests that childhood attitudes on the part of each sex toward the other entail cognitive simplicity and an overt, univalent hostility that undergoes a dramatic (but unfortunately underresearched) transformation to greater complexity and ambivalence. This transition rests not only on the
AMBIVALENT SEXISM
169
development of cognitive abilities, but on a fundamental shift in the structure of peer relationships, from segregated to interdependent, as adolescent sexuality begins. Although it is easy to read this developmental story as one of progress in which adolescents and adults become less prejudiced toward the other sex, we caution against assuming that this is necessarily so. We suggest that it makes more sense to view this transition as qualitative, from one form of prejudice to another, for the larger lesson of ambivalent sexism is that prejudice is not a one-size-fits-all construct. In the final section, we argue that not only do qualitatively different kinds of prejudice apply to intergroup relations more generally, not just to gender relations, but also that the character and the occurrence of these prejudices responds to the structure of intergroup relations.
X. General Implications for Prejudice: A Theory of Stereotype Content We have suggested (and presented supporting data) that the structure of adult male–female relationships gives rise to both hostile and benevolent ideologies on the part of each sex toward the other. In doing so, we have emphasized the peculiar distinctiveness of gender relations, in which both extreme power differences and profound interdependence typically coexist. Although gender relations may be unique in terms of the degree to which they combine these structural characteristics, other group relations can have a similar quality. Furthermore, although power difference and simultaneous cooperative interdependence characterize gender relations as a whole, relations between individuals or subtypes of each sex do not inevitably combine these two qualities (e.g., many men may see themselves as being in a competitive power struggle with feminists or career women rather than being dependent on them). In this section, we offer an overview of a more general theory of prejudice that provides a broader interpretive framework for ambivalent attitudes between groups. Like ambivalent sexism theory, the more general approach examines how intergroup attitudes, emotions, and stereotypes reflect the structure of group relations. In contrast to the antipathy model, this theory posits multiple forms of prejudice characterized by distinct patterns of stereotypes, behaviors, and emotions in response to outgroups.
A. TWO DIMENSIONS OF STEREOTYPE CONTENT Stereotypes about men and women fall along two personality trait dimensions, often labeled communion and agency (Eagly, 1987). The overall stereotype of women holds that they are warm, understanding, and nurturing, whereas men are stereotypically independent, ambitious, and competitive. These trait dimensions,
170
PETER GLICK AND SUSAN T. FISKE
however, are by no means confined to stereotypes about men and women. Fiske (1998) and her colleagues (Fiske et al. 1999a, 1999b) note that these same trait clusters crop up repeatedly as central dimensions in impression formation, both in interpersonal (e.g., Asch, 1946; Rosenberg, Nelson, & Vivekananthan, 1968) and in intragroup and intergroup perception (see Fiske, 1998, pp. 377–381 for a review). We label these general dimensions “warmth” and “competence” (Glick & Fiske, 2001), with the former assessing the target’s socioemotional orientation toward others (warm versus cold) and the latter indicating the target’s perceived ability to be successful at the kinds of tasks that bring the most status. Target’s possession of warm traits is particularly important in determining feelings of affection or liking, whereas perceived competence determines whether targets are respected (Hamilton & Fallot, 1974; Jaimeson, Lydon, & Zanna, 1987; Lydon, Jamieson, & Zanna, 1988). Overall, gender stereotypes characterize men as high in competence, but low in warmth and women as low in competence, but high in warmth. Eagly and Mladinic (1993) showed that women’s greater perceived warmth generates the “women are wonderful” effect. These researchers also demonstrated that although men’s stereotypical competence is admired, these traits, when viewed as excessive, have a downside—men are often viewed as too competitive, arrogant, and egotistical. The general stereotypes of men and women suggest a trade-off: Be perceived as warm, but not competent, or as competent, but not warm (i.e., be liked, but disrespected or respected, but disliked). Competence and warmth are not mutually exclusive; extensive work on the halo effect (see Cooper, 1981) shows that in interpersonal judgments we often respect those we like. When it comes to stereotypes about men and women, however, the two trait dimensions are negatively correlated (Spence & Buckner, 2000). The competence cluster includes traits such as ambition and competitiveness that (at least when carried to excess) are incompatible with being warm and caring. Peeters and Czapinski (1990) have labeled the competence traits as “self-profitable” because their aim is typically to further self-interest (e.g., to gain power and status), whereas they label the warmth cluster as “other-profitable” traits because these traits inherently reflect an orientation toward helping others (see also Sidanius et al., 1994, on masculine “ranking” versus feminine “linking” orientations). Even at the level of gender subtyping, the same dimensions and, often, the same trade-offs between traits that garner liking versus those that elicit respect, are evident. The overall warm-but-not-competent stereotype of women is consistent with traditional female subtypes (e.g., mothers), whereas perceptions of most nontraditional subtypes essentially reverse the stereotype—career women are not considered warm, but are viewed as competent (Eagly, 1987; Glick et al., 1997) and the same is true of feminists (MacDonald & Zanna, 1998). In essence, nontraditional women are stereotyped as having (perhaps exaggeratedly) masculine personalities, losing feminine warmth and gaining competence. Thereby, they
AMBIVALENT SEXISM
171
forfeit the affection directed toward traditional women, though they simultaneously gain more respect. The problem for nontraditional women is that the negatives that come with perceived competence may target them more strongly than men who act in comparable ways because enacting competence (especially when done with an authoritarian, masculine style) violates prescriptions for feminine niceness (Eagly, Makhijani, & Klonsky, 1992; Rudman, 1998; Rudman & Glick, 1999).
B. AMBIVALENCE AS A COMMON REACTION TO OUT-GROUPS Our recent research shows that stereotypes of a variety of groups can meaningfully map out in a two-dimensional space defined by competence and warmth. Fiske et al. (1999a) had participants nominate social groups they consider important in U.S. society. These groups were subsequently rated by new samples of both undergraduate students and older adults on a variety of measures, including traits that fall on the warmth and competence dimensions, for which they indicated their understanding of the consensual stereotype of these groups (regardless of whether they themselves subscribed to stereotypical beliefs). Cluster analyses revealed 5 clusters on warmth and competence axes. Apart from 1 cluster centered around the origin, representing a few groups with (on average) neutral ratings on the competence and warmth dimensions, the other 4 clusters define what can be viewed as a 2 (competence: high vs low) × 2 (warmth: high vs low) table. Established models of intergroup relations (e.g., social identity theory) suggest that in-group members (e.g., in most U.S. college student samples, middle-class, Whites, and Christians) would tend to rate favorably on all dimensions (i.e., land in the high competence–high warmth cluster) and, indeed, this occurred. The antipathy model of prejudice would further suggest that out-groups would tend to rate low on both dimensions (i.e., group into the low competence–low warmth cluster). This did not occur for many groups. Rather, a number of out-groups were placed into the off-diagonals and, like men and women, came across as competent, but not warm or warm, but not competent, suggesting that cross-dimesionally ambivalent characterizations of groups are common, not rare. The cluster analyses suggest that such ambivalent reactions to out-groups are frequent. Fiske et al. (1999a) choose 17 wide-ranging and important social groups (as determined by consensus on pretests in which participants nominated groups, supplemented by groups frequently studied in the prejudice research literature); participants rated the groups on several traits composing competence and warmth dimensions. Most groups, 13 of 17, were perceived to be significantly more warm than they were competent or significantly more competent than warm (as determined by t- tests). Groups that have consistently, across a number of studies, emerged as being perceived as warm, but not competent, include the prototypical traditional female subtype of housewives along with the elderly and those who
172
PETER GLICK AND SUSAN T. FISKE
possess disabilities, both physical (disabled, blind) and mental (retarded). The groups that are repeatedly rated as competent, but not warm, include the prototypical nontraditional female subtype of feminists, model minorities (Jews and Asians), and rich people. Low competence–low warmth ratings tend to describe only the poor, welfare recipients, and the homeless.
C. SOCIAL STRUCTURE PREDICTS STEREOTYPES Why is ambivalence toward groups a relatively common phenomenon? Ambivalent sexism theory suggests that ambivalent characterizations of women and men follow from structural relations between these groups. Similarly, our broader theory of stereotype content posits that the content of stereotypes, as well as emotions toward groups, flow from the structure of intergroup relations (Fiske et al., 1999a, 1999b; Glick & Fiske, 2001). Some earlier theories of stereotype content have focused on stereotypes as a consequence of roles (e.g., LeVine & Campbell, 1972). Our approach considers the effects of two more generalized structural variables, both previously identified (but never combined) as important aspects of intergroup relations: relative status of groups and the nature of their interdependence (competitive versus cooperative). The relative status of groups has been central focus of prejudice theorists and researchers, who have emphasized prejudice directed by dominant toward subordinate groups. The interdependence dimension has also received a great deal of attention, but, conceptually, positive interdependence (cooperation between groups) has falsely entailed intergroup equality and a lack of prejudice, whereas competitive intergroup relations have surfaced as both a cause and a consequence of prejudice (e.g., realistic group conflict theory; LeVine & Campbell, 1972). As ambivalent sexism theory shows (see also Jackman, 1994), however, positive interdependence can coexist with prejudice and with strong and stable differences in status. Thus, our innovation is not the introduction of previously ignored structural dimensions of intergroup relations, but the specification of a 2 (relative status: high vs low) × 2 (type of interdependence: competitive vs cooperative) table that unconfounds these dimensions to specify how these structural aspects of intergroup relations determine stereotype content and to show how the structure of intergroup relations creates qualitatively different forms of prejudice previously obscured by social psychologists’ commitment to the notion of prejudice as an antipathy (see also Alexander, Brewer, & Herrmann, 1999, for a related, but independently developed theory of group images). Table X illustrates our theory of stereotype content. In this model, inferences about a group’s competence follow from their relative social status. High-status groups, whether they represent the culturally dominant group or a socioeconomically successful minority group, are stereotyped as competent. This effect also fits,
173
AMBIVALENT SEXISM TABLE X A THEORY OF STEREOTYPE CONTENT Type of interdependence Status High Stereotype Negative emotions Positive emotions Behavior Experienced by
Cooperative Admiration Competent, warm Respect, admiration, affection Defer
Groups in category
subordinates toward generous dominants upon whom they depend and dominants toward their own group and its allies Middle-class Whites, Christians
Low Stereotype Negative Emotions Positive Emotions Behavior
Paternalistic prejudice Warm, but incompetent Disrespect, condescension Patronizing affection, pity, liking Intimacy, but role differentiation
Experienced by
Dominants to subordinates upon whom they depend and toward groups that pose no threat
Groups in category
Retarded, housewives, disabled, elderly, blind
Competitive Envious prejudice Competent, but not warm Envy, fear, resentment, hostility Respect, admiration Avoid, exclude, segregate, exterminate Dominants whose status is slipping and low status groups toward successful minorities; low status groups toward dominants Jews, Asians, feminists, career women, rich people Contemptuous prejudice Not warm, incompetent Disrespect, resentment, hostility Avoid, exclude, segregate, exterminate Dominants toward subordinates who are seen as a threat or as illegitimate dependents (a drain on social resources) Poor Whites, poor Blacks, welfare recipients
Note. Adapted from Glick and Fiske (2001).
of course, correspondence bias (Jones & Harris, 1967; Ross, Amabile, & Steinmetz, 1977), system-justification theory (Jost & Banaji, 1994), and social identity theory (Tajfel, 1981). The stereotype of high-status groups as relatively more competent than low-status groups has been replicated many times (see Jost et al., 2001). The model hypothesizes the form of interdependence between groups (cooperative versus competitive) specifically to affect stereotypes about a group’s warmth, with groups viewed as competitors being stereotyped as cold and self-interested and with groups viewed as cooperators being stereotyped as warm. This prediction also has ample social psychological precedence, most notably, Sherif’s (1996) focus on shared goals as promoting friendly intergroup relationships. The results of this 2 × 2 scheme are qualitatively different types of prejudice, two of which are ambivalent (see Table X). In this model, prejudice as a uniform antipathy is directed only at groups that are low in status and that are perceived as
174
PETER GLICK AND SUSAN T. FISKE
being in a zero-sum game with the in-group (e.g., poor people may be viewed as manipulators who deliberately rely on welfare payments subsidized by taxpayers). Such contemptuous prejudice stereotypes its targets as lacking both competence and warmth. Emotional reactions toward targets of contemptuous prejudice are predicted to be wholly negative, including disgust, contempt, resentment, and dislike. But contempt (univalent antipathy) becomes only one form of prejudice in this model. Two other forms of prejudice, which we believe to be quite common, are ambivalent. Paternalistic prejudice, of which BS is a prime example, aims at groups perceived as low in status, but cooperative (or, at least, nonthreatening). We hypothesize that stereotypes of the lower status group members’ warmth will be especially strong when high-status group members depend on them (e.g., for their labor) and have intimate contact with them (e.g., because low-status group members fulfill service jobs, such as cleaning, gardening, or childcare). As Jackman (1994) argues, it is in dominants’ interests to have affectionate and comfortable relationships with their subordinates, who may well be in a position to wreak havoc on or sabotage various aspects of dominant group members’ lives through covert acts of rebellion (e.g., spitting in the soup). Envious prejudice targets out-groups who are socioeconomically successful, but perceived as competitors. Although admired for their competence (inferred from their socioeconomic success), they inspire fear. So-called “model minorities” make prime targets of envious prejudice, such as anti-Semitic notions of Jewish conspiracy to control world commerce (see Glick, in press). Prejudice toward Asians is similar. Lin and Fiske (1999) found that Asian Americans are perceived as highly successful, but resented as power-hungry competitors. Further, they are stereotyped as lacking in social skills (i.e., as not possessing socioemotional warmth). Moreover, HS fits the envious prejudice category; it is directed at women perceived as wresting control from men (e.g., career women, feminists, and temptresses). These women are thought to be competent (even by hostile sexists), but cold and manipulative (rather than warm). Similarly, HM is an envious prejudice because it represents resentment of men’s power and stereotypes men as cold, selfish, and arrogant. The final cell of the table is an attitude of admiration. Admired groups are stereotyped as both competent and warm. This cell captures in-group bias, at least for those groups that have sufficient social standing—groups with low social status are known to cede the competence traits to higher status groups [a central finding and tenet of the system-justification approach (Jost et al., 2001)]. Admiration, however, may also be directed by members of subordinate groups toward the dominant cultural group if the dominant group is perceived to be benevolent (i.e., cooperative). BM fits this category, as it represents the attitude of women who believe that men are benevolent providers deserving of their higher status. Such admiration is most likely to be present in stable systems that justify long-standing status inequalities through widely accepted legitimizing ideologies (e.g., BS). The
AMBIVALENT SEXISM
175
low level of BM scores among women in most countries studied thus far may reflect the breakdown of stability in male–female status differences and increasing skepticism about the legitimacy of these differences. We have found evidence consistent with the claim that relative status and type of interdependence determine stereotypes in correlational work that examines stereotypes of actual groups and in experiments that manipulate the relationships between fictional groups (Fiske et al., 1999a, 1999b). In the correlational work with real social groups, the correlation between perceptions of group status (perceived socioeconomic success of groups) and consensual stereotypes of competence have been extremely strong. Perceived intergroup competition also correlates well with consensual stereotypes of warmth. Descriptions of fictitious groups further allow the manipulation of structural relations without interference of prior knowledge about the groups. Two studies support the hypothesized causal connection between the structural variables of status (see also Conway, Pizzamiglio, & Mount, 1996) and interdependence, respectively, on perceptions of competence and warmth.
D. COMPARISON WITH RECENT THEORIES OF PREJUDICE Two recent theories of prejudice also propose qualitatively different kinds of stereotypes of and prejudice toward out-groups. Alexander et al.’s (1999) functional theory of stereotypes was developed out of image theory in international relations. In its basic contentions, it resembles our theory of stereotype content because it holds that out-group stereotypes fall into qualitatively different categories determined by the structure of group relations. Furthermore, Alexander et al. emphasize the same structural variables that we identify as important—status and interdependence (which they label as goal compatibility)—though they also add power to the equation as separate from status (whereas we simplify the picture by assuming that perceived power and status typically go together). In general, though developed separately, the Alexander et al. model and the empirical evidence these theorists have gathered to illustrate its utility are highly compatible with our theory of stereotype content. Although there is considerable convergence between our model and Alexander et al.’s, the theories differ in several ways. In addition to the difference in the number of structural factors each model emphasizes, our model explicitly identifies two general dimensions of stereotype content (competence and warmth) that correspond to the two structural dimensions of group relations (status and type of interdependence). Another difference is our focus on the ambivalence of many outgroup stereotypes as a result of positive evaluation on one and negative evaluation on the other dimension of stereotypes (e.g., high warmth, but low competence), and the role that this ambivalence plays in legitimizing prejudice. Duckitt (this volume) has recently proposed a theory of prejudice that also suggests qualitatively different types of prejudice. Unlike Alexander et al.’s (1999) or
176
PETER GLICK AND SUSAN T. FISKE
our own model, Duckitt focuses on individual differences in ideology—specifically Right-Wing Authoritarianism (Altemeyer, 1998) and Social Dominance Orientation (Sidanius & Pratto, 1999)—rather than on structural relations between groups. Nevertheless, despite this very different starting point, Duckitt identifies types of prejudice, motivations that result in these forms of prejudice, and dimensions of stereotyping that are generally consistent with those we have proposed. Specifically, Duckitt’s “threat-control” motivation (associated with authoritarianism) is related to what we have termed envious prejudice. We agree with Duckitt that the perception of threat (especially, we believe, when the out-group is perceived as highly competent) tends to generate anger, fear, anxiety, and what Fiske (1998) has termed “hot” discrimination toward the out-group. Because low-status groups can also sometimes be perceived as a threat (e.g., in cases where they are a large minority or even a majority), threat-control motivation can also be part of (what we have termed) contemptuous prejudice. Duckitt’s “competitive-dominance” motivation (associated with social dominance orientation) is relevant to the prejudices that, in our model, target lower status groups, paternalistic and contemptuous. Like Duckitt, we argue that these prejudices uphold the status quo, maintaining an existing social hierarchy. Because Duckitt’s model is amenable to incorporating the influence of structural relations between groups (as he proposes that intergroup dynamics, not simply preexisting ideologies, may activate the motivations he has identified), it is compatible with our approach. In this respect, the models can be viewed as enriching each other: Duckitt’s approach focuses on individual differences and socialization practices that undoubtedly interact wih the intergroup dynamics we identify. For example, in a culture in which child-rearing practices foster authoritarian beliefs, people ought to be prone to responding in a highly reactive and aggressive manner to groups that (in our model) evoke envious prejudice. Duckitt agrees that our models are compatible. Indeed, he has embraced and also expanded upon our typology of prejudice. One modification Duckitt suggests is renaming “envious prejudice” as “resentful dislike.” Although we had originally considered a similar label, recent data confirm that envy, rather than resentment (which is more strongly direct toward low-status, competitive groups, e.g., welfare recipients), is the primary emotion directed toward groups in the high-statuscompetitive category (e.g., Jews, Asians, and the rich), suggesting that “envious prejudice” is the more appropriate term. More substantially, Duckitt proposes to expand our typology to a 3 × 2 scheme, adding an “equal” level on the status dimension that yields another category of prejudice he terms “hostile.” This proposal assumes a specific group as a frame of reference against which status is to be judged. In other words, Duckitt assumes an in-group to which out-groups are compared. In contrast, our model was created without such a specific frame of reference in mind. Rather, we assumed a bird’s eye view of the social structure, without using a specific group as a standard
AMBIVALENT SEXISM
177
against which others are compared—an approach that makes the term “equal” meaningless, as there is no answer to the question “equal to what?” We took this approach because stereotypes of groups are, to a great extent, consensual (Tajfel, 1981; Jost & Banaji, 1994) due to system-justifying tendencies and reality constraints, especially on the status dimensions (i.e., low-status group members often endorse stereotypes of their group’s lesser competence). Nevertheless, we think our model can be successfully employed when a frame of reference is assumed. Stereotypes are not completely consensual; prejudiced attitudes are also strongly determined by the perspective and group membership of the perceiver. Thus, for example, members of low-status groups may experience envious prejudice toward groups that are not especially high in status, but are higher up in the social hierarchy. When using our model with a specific frame of reference (i.e., an in-group) in mind, the “equal status” addition could potentially be useful. However, we are not convinced that “hostile prejudice” as described by Duckitt is a necessary addition, as the attitudes (hostile), images (enemy), and behaviors (conflict) he describes are not unique, but rather capture that which is common to both envious and contemptuous prejudice as we have described them (i.e., this new category does not define a qualitatively different form of prejudice, but merely captures the overlap between the two prejudices, envious and contemptuous, that we suggest result from intergroup competition). Furthermore, the dynamics of competitive forms of prejudice are, we believe, likely to cause in-group members to perceive the out-group as either hypercompetent (i.e., to move toward envious prejudice) or as inferior (i.e., to move toward contemptuous prejudice). The latter dynamic is well known: conflict often leads to disparagement of the out-group. The former dynamic is evident in historical examples, such as forms of anti-Semitism in which the power (and therefore the threat posed by) the Jews has been grossly exaggerated (Glick, in press). Interestingly, Duckitt’s social dominance and threat motivations may help predict in which direction out-group stereotypes will tend to move. When social dominance concerns become salient, in-groups are likely to emphasize out-group inferiority (to enhance the in-group’s sense of dominance), whereas when threat is activated, in-groups may exaggerate the competence and superior power of the out-group (just as anxiety often leads individuals to exaggerate perceived threats). Although Duckitt has been willing to accept our basic typology of prejudices (with some additions), we note that his approach is still strongly influenced by the antipathy model of prejudice. Duckitt tends to focus on the subjectively negative aspects of prejudice. For example, he emphasizes hostile stereotypes on the warmcold dimension directed toward feared (i.e., “threatening”) out-groups rather than how these groups are perceived on the competence dimension. This is only natural from a prejudice-equals-antipathy view because, according to our model, the out-groups that are viewed as most threatening are stereotyped positively (not negatively) on the competence dimension. Yet we argue that these attributions of
178
PETER GLICK AND SUSAN T. FISKE
competence may actually exacerbate the perception of threat and, therefore, hostility toward the out-group. Similarly, Duckitt proposes that competitive-dominance motivation results in negative stereotypes of an outgroup’s competence and “low positive affect.” This fails fully to recognize that dominance may be maintained over groups that are disrespected, but liked (even loved). We believe that we have convincingly demonstrated in the case of sexism that BS is a subjectively positive attitude of affection toward women (a lower status group) that simultaneously reflects Duckitt’s competitive-dominance motivation. In both examples, the “positive” aspects of out-group stereotypes can serve to legitimate and maintain discrimination. Exaggerating the competence of Jews or Asians can inflame hostility toward them as competitors. Stressing the empathy and kindness of women makes them seem especially suited to taking care of husband and children at home and unsuited to the competitive world of the boardroom. In short, perhaps the most important distinguishing feature of our model compared to Duckitt’s or Alexander et al.’s approaches is its explicit recognition of and focus on the hidden dangers of seemingly favorable aspects of out-group stereotypes.
E. CONCLUSION: POSITIVE STEREOTYPES ARE NOT NECESSARILY BENIGN Although contemptuous prejudice may seem to have more serious social consequences than the ambivalent prejudices (paternalistic and envious), we argue that this is not the case (see Glick, in press; Glick & Fiske, 2001). Paternalistic prejudices, like BS, may be insidiously effective at legitimizing inequality because they pacify many members of subordinate groups as well as dominant group members who might otherwise question the legitimacy of their privileges. For example, the belief that colonization and Christianization represented both the material and spiritual improvement of non-European peoples retarded support for anticolonial movements in Europe (Hochschild, 1998). A form of prejudice that served as a primary justification for the horrors of both colonization and slavery (see Jackman, 1994) deserves to be taken seriously. Likewise, envious prejudice can have terrible consequences. Nazi anti-Semitism was an envious prejudice that viewed Jews as extremely competent, but coldly calculating and malicious. Socioeconomically successful minority groups may be at grave risk for violence because members of other groups can justify aggression against these minorities as a matter of self-defense. Glick (in press) argues that scapegoating—blaming a group for social and economic turmoil—is more likely to be directed at successful than unsuccessful minorities. In difficult times, people actively search for an explanation of their shared misfortune (Staub, 1989; Tajfel, 1981). It is more psychologically plausible that groups who wield influence out of proportion to their numbers will be seen as a source of widespread difficulties
AMBIVALENT SEXISM
179
(e.g., economic depression). For example, the Nazis believed that a Jewish conspiracy was responsible for Germany’s loss of WWI, and the Nazis characterized the war against the Jews as ‘“the battle for the destiny of the German people”’ (Arendt, 1964, p. 52). Why target the Jews? Because their perceived prominence in banking, industry, science, and the media outstripped their small numbers; the Jews were the only minority that seemed successful and powerful enough to have caused Germany’s post-WWI economic collapse. More generally, genocide is typically directed at groups that are perceived to be powerful, who cannot therefore (in the eyes of perpetrators) be enslaved or domesticated (see Staub, 1989) The perspective of the antipathy model of prejudice could falsely paint ambivalent feelings and stereotypes toward groups (as opposed to univalent antipathy) as signs of progress. Our model provides a contrast. The “positive” aspects of paternalistic stereotypes, the attribution of warmth to members of low-status groups, may merely salve the consciences of people in the privileged majority and, worse still, defuse resistance to inequality. When it comes to pressuring women to remain in restricted roles or to discourage them from acting ambitiously, prescriptions of feminine warmth and caring (traits that are subjectively favorable, but perceived to be incompatible with ambition) can be insidiously effective at curtailing women’s aspirations (cf. Rudman, 1998; Rudman & Glick, 1999). Imagine the popular response to a politician who pronounces that women are the best nurturers (and therefore it is better for mothers to minimize their investment in careers outside the home) as opposed to one who asserts outright that women do not have the ability to excel in high-status careers. The “positive” side of envious stereotypes can be more dangerous still. The statements “Jews are clever” or “Asians are hardworking” are not necessarily simple admiration, but may be an integral part of a virulent anti-Semitism or anti-Asian prejudice. For envious prejudices, exaggerated beliefs about another group’s ability only add to fear and resentment of groups that are thought to be malicious enemies. Self-defense is a powerful rationalization for even the most extreme forms of violent discrimination.
XI. Conclusions No other groups have shared such a long history of intimate interdependence coexistent with power and status inequality as have women and men. It is not surprising, then, that gender intergroup relations are complex and ambivalent or that social commentators have used metaphors of war about as frequently as myths of romance to characterize relations between the sexes. The research program described here has attempted to show how structural aspects of male–female relations create both hostile and benevolent ideologies about each gender that encompass the domains instantiating power differences and mutual interdependence—
180
PETER GLICK AND SUSAN T. FISKE
patriarchy, gender stereotypes and roles, and sexuality. Contrary to the predominant antipathy model of prejudice, in which power and status differences precipitate overt conflict and hostility between groups, our research is more compatible with Jackman’s (1994) contention that long-term inequality is more often maintained by paternalistic systems that accompany (and disguise) expropriation with affection by dominant groups toward subordinate groups. Exploitation and affection are not mutually exclusive; rather, this combination is the predictable outcome of the structure of male–female relations. Other recent theories, though still steeped in the tradition of prejudice as an antipathy, similarly suggest that prejudiced feeling are conflicted, complex, and camouflaged. For example, modern racism seems to be aversive (Gaertner & Dovidio, 1986), subtle (McConahay, 1986), and symbolic (Kinder & Sears, 1981) because of conflicts between a desire among most Whites to be egalitarian, despite having hostile affect toward and negative stereotypes of Blacks. New conceptions of sexism (Swim et al., 1995; Tougas et al., 1995), embodied in the MS and NS scales, take a related theoretical perspective. All of these approaches contrast “old-fashioned” forms of prejudice, presumed to have been overtly and univalently hostile, with more subtle and ambivalent modern prejudices. Our approach differs in that we presume ambivalent sexism not to be a recent development, but rooted in structural aspects of gender relations that have long been a part of the human condition. Benevolent sexism, for example, may be subtle, but it is far from being a uniquely contemporary concept. The notion of women as pure, delicate, and essential complements to men is evident in ancient sources, such as classic texts (Pomeroy, 1975) and traditional religious images (e.g., the Madonna). Presumably, it is because the structural conditions that promote sexist ambivalence are virtually universal that the ASI has shown such strong cross-cultural validity. Perhaps the most important contribution of our research is to point out the role that subjectively positive views of women have in perpetuating gender inequality. Because social psychologists have typically equated prejudice with antipathy, BS has previously been ignored. Although BS is subtle and may prompt behavior that benefits individual women in the short term (e.g., chivalrous offers of help to damsels in distress), the weight of the evidence suggests that BS is kinder, gentler tool of oppression. At the systemic level, BS and HS go hand-in-hand (with national averages on HS and BS, correlating .89 for both men and women). And both BS and HS predict gender inequality on objective national indicators (e.g., degree to which women inhabit high-status positions) in a diverse set of nations. At the individual level of analysis, the means by which HS and BS act as complementary tools of control is evident in findings that HS is targeted at nontraditional and BS at traditional women, thereby providing punishment (hostility) for women who fail to conform and rewards (provision, protection, and affection) for women who do conform to the conventional gender roles that serve men’s interests. More perniciously, BS tends to disarm women’s resistance to patriarchy, promoting
AMBIVALENT SEXISM
181
rationalizations that make discriminatory behavior tolerable—women who endorse BS are more likely to excuse sexist acts (e.g., being forbidden to work outside the home) that can be perceived as protectively benevolent (Moya et al., 1999). Women’s relatively greater acceptance of BS, as opposed to HS, can be viewed as an unwillingness to bite the hand that provides and protects. What this obscures, however, is the fact that men’s HS and their monopolization of power and resources coerce women into accepting BS. Having low power within a hostile environment creates strong incentives to accept ideologies that promise provision, protection, and affection from members of the dominant group. This effect was evident in the increasing tendency of women (as compared to men) to accept BS the more sexist the culture, to the point that in the most sexist cultures the usual gender gap reversed and women endorsed BS more than men. The combination of HS and BS resembles a protection racket in which men provide both the threat (HS) and the solution to that threat (BS), with the price being women’s compliance with conventional gender roles and acceptance of patriarchy. Although women do show evidence of system justification, in that their acceptance of both HS and BS is correlated, cross-culturally, with men’s endorsement of these ideologies, resistance to accepting HS was apparent; the gender gap in HS scores was larger in more sexist cultures. In addition, in countries where men’s sexism scores were relatively low, the gender gap in BS was greater, suggesting that women in these countries may be starting to reject BS as well. Women’s resistance to (at least the hostile aspects of) patriarchy is evident even in the traditional views about men tapped by the AMI. The HM scale (which correlates with traditional gender ideologies) suggests considerable resentment of men’s greater power, even as it presumes that inequality will continue. Furthermore, we have generally found very low scores among women on BM, in both the United States and most other nations studied (although the set of nations in which the AMI has been administered is much smaller than for the ASI). These low BM scores suggest that female college students (at least) are not currently buying into the idea that men are superior in the nondomestic world or that men should be cared for by women at home. Although no historical data allow us to compare current responses, we suspect that the low BM scores among women reflect the changes in gender roles and attitudes that have occurred in the past few decades in many industrialized nations (e.g., women’s movement into the paid workforce and increased questioning of the legitimacy of gender inequalities). Although our original conception was that ambivalence occurs when people simultaneously hold both hostile and benevolent ideologies (e.g., score highly on both HS and BS), subsequent research has suggested that each type of ideology may itself encompass a form of cross-dimensional ambivalence (MacDonald & Zanna, 1998; Thompson et al., 1995). In terms of attitudes toward women, HS combines a grudging admiration of potentially powerful women’s competence with envy, fear, and animosity because such women (e.g., feminists) are viewed as
182
PETER GLICK AND SUSAN T. FISKE
dangerous competitors in a struggle for power. In contrast, BS combines affection toward conventional female types (who are viewed by men as helpmates) with a lack of respect. More generally, we have proposed that prejudice toward many groups reflects these two forms of cross-dimensional ambivalence: being liked, but disrespected versus being respected, but disliked. Our recent theory of stereotype content (Fiske et al., 1999a, 1999b; Glick & Fiske, 2001) proposes that the same general structural factors that affect gender relations—relative status and type of interdependence (cooperative versus competitive)—similarly condition the stereotypes, emotions, and behaviors directed at other social groups. In this model, only groups that are perceived to be both lower in status and a competitive threat (perhaps as a drain on social resources) are subjected to univalent hostility or contempt. In contrast, low-status groups that are viewed as cooperative (either nonthreatening or actively helpful) are perceived as warm and rewarded with affection, but are also viewed as incompetent and therefore not worthy of respect. Successful minority groups are respected as competent, but viewed as lacking warmth because they are perceived as a competitive threat, evoking respect, but also dislike, envy, and fear. These two ambivalent prejudices, paternalistic and envious, have not previously been distinguished. Evidence that these forms of prejudice are widespread (Fiske et al., 1999a, 1999b) suggests that, in contrast to the standard antipathy model, there are qualitatively different kinds of prejudices (cf. Young-Breuhl, 1996). Although we propose the addition of new conceptions of prejudice, our model is parsimonious and posits that the type of prejudice (and the overall content of stereotypes) directed at a group is predictable from its structural relationship with other groups on two dimensions whose importance has long been recognized: status and interdependence. Recognizing that the form of prejudice directed at, for example, Jews and Asians differs from that directed at traditional women and the elderly is a first step (we hope) to clearing up important conceptual and empirical puzzles such as the one that launched our research on ambivalent sexism—how can women be both (virtually universally) oppressed, yet loved? The antipathy conception of prejudice encouraged researchers to look for uniformity in affect, cognitions, and behavior toward targets of prejudice—all were assumed to be characterized by an undifferentiated amalgam of hostility, conflict, aversion, and opposition. As a result, the antipathy model was ill-equipped to explain results showing a more complicated picture in which the various components of attitudes do not always cohere in the expected manner (see Zanna & Rempel, 1988). Our proposed theory of stereotype content makes more precise predictions that differentiate between dimensions (warmth and competence) on which opposite evaluations are often expected. Furthermore, it highlights the perils of at least some “positive” stereotypes which may exacerbate ambivalent prejudices. Stereotypes of subordinate groups as warm (but not competent) may be a sign of affectionate coercion that maintains inequality, whereas stereotypes of successful
AMBIVALENT SEXISM
183
minorities as competent may only serve to justify aggression and discrimination against groups deemed to be dangerous competitors. Social psychology must move beyond the equation of prejudice with an undifferentiated antipathy to better understand not only relations between men and women, but also intergroup relations more generally.
References Alexander, M. G., Brewer, M. B., & Herrmann, R. K. (1999). Images and affect: A functional analysis of outgroup stereotypes. Journal of Personality and Social Psychology, 77, 78–93. Allport, G. (1954). The nature of prejudice. Reading, MA: Addison–Wesley. Altemeyer, B. (1998). The other authoritarian personality. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 30, pp. 47–92). San Diego: Academic Press. Arendt, H. (1964). Eichmann in Jerusalem: A report on the banality of evil. New York: Penguin. Asch, S. E. (1946). Forming impressions of personality. Journal of Abnormal and Social Psychology, 41, 1230–1240. Bargh, J. A., & Raymond, P. (1995). The na¨ıve misuse of power: Nonconscious sources of sexual harassment. Journal of Social Issues, 51, 85–96. Berger, J. (1982). Expectation states theory: A theoretical research program. New York: University Press of America. Berscheid, E., & Peplau, L. A. (1983). The emerging science of close relationships. In H. H. Kelley, E. Berscheid, A. Christensen, J. H. Harvey, T. L. Huston, G. Levinger, E. McClintock, L. A. Peplau, & D. R. Petersen (Eds.), Close relationships (pp. 1–19). New York: Freeman. Bigler, R. S. (1999). Psychological interventions designed to counter sexism in children: Empirical limitations and theoretical foundations. In W. B. Swann, Jr., L. A. Gilbert, & J. Langlois (Eds.), Sexism and stereotypes in modern society: The gender science of Janet Taylor Spence (pp. 129– 151). Washington, DC: American Psychological Association. Bollen, K. A. (1989). Structural equations with latent variables, New York: Wiley. Brewer, M. B., Manzi, J. M., & Shaw, J. S. (1993). In-group identification as a function of depersonalization, distinctiveness, and status. Psychological Science, 4, 88–92. Burt, M. R. (1980). Cultural myths and support for rape. Journal of Personality and Social Psychology, 38, 217–230. Carpenter, S. (2000). Implicit gender attitudes: Group membership, cultural construal, valenceconsistency, and stability. Unpublished doctoral dissertation. Yale University. Cejka, M. A., & Eagly, A. H. (1999). Gender-stereotypic images of occupations correspond to the sex segregation of employment. Personality and Social Psychology Bulletin, 25, 413–423. Conway, M., Pizzamiglio, M. T., & Mount, L. (1996). Status, communality, and agency: Implications for stereotypes of gender and other groups. Journal of Personality and Social Psychology, 71, 25–38. Cooper, W. H. (1981). Ubiquitous halo. Psychological Bulletin, 90, 218–224. Crocker, J., & Luhtanen, R. (1990). Collective self-esteem and ingroup bias. Journal of Personality and Social Psychology, 58, 60–67. Davis, J. A. (1959). A formal interpretation of the theory of relative deprivation. Sociometry, 22, 280–296. Deaux, K., Winton, W., Crowley, M., & Lewis, L. L. (1985). Level of categorization and content of gender stereotypes. Social Cognition, 3, 145–167.
184
PETER GLICK AND SUSAN T. FISKE
Downs, A. C., & Engelson, S. A. (1982). The Attitudes Toward Men Scale (AMS): An analysis of the role and status of men and masculinity. Catalog of Selected Documents in Psychology, 12(4), 45 (Ms. No. 2503). Duckitt, J. (2001). A cognitive-motivational theory of ideology and prejudice. In M. P. Zanna (Ed.), Advances in experimental social psychology. (vol. 33, pp. 41–112). San Diego, CA: Academic Press. Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum. Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and evalution of leaders: A metaanalysis. Psychological Bulletin, 111, 3–22. Eagly, A. H., & Mladinic, A. (1989). Gender stereotypes and attitudes toward women and men. Personality and Social Psychology Bulletin, 15, 543–558. Eagly, A. H., & Mladinic, A. (1993). Are people prejudiced against women? Some answers from research on attitudes, gender stereotypes and judgments of competence. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 5, pp. 1–35). New york: Wiley. Eagly, A. H., & Wood, W. (1999). The origins of sex differences in human behavior: Evolved dispositions versus social roles. American Psychologist, 54, 408–423. Ellemers, N. (2001). Individual upward mobility and the perceived legitimacy of intergroup relations. In J. T. Jost & B . Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations. Cambridge, UK: Cambridge Univ. Press. Esses, V., Haddock, G., & Zanna, M. P. (1993). Values, stereotypes and emotions as determinants of intergroup attitudes. In D. Mackie & D. L. Hamilton (Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 137–166). San Diego, CA: Academic Press. Exp´osito, F., Moya, M., & Glick, P. (1998). Sexismo ambivalente: Medici´on y correlatos [Ambivalent sexism: Measurement and correlates]. Revista de Psicolog´ıa Social, 13, 159–169. Fagot, B. I., & Leinbach, M. D. (1993). Gender-role development in young children: From discrimination to labeling. Developmental Review, 13, 205–224. Feather, N. T. (1994). Attitudes toward high achievers and reactions to their fall: Theory and research concerning tall poppies. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 26, pp. 1–73). San Diego, CA: Academic Press. Fiske, S. T. (1998). Prejudice, stereotyping, and discrimination. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 357–411). New York: McGraw-Hill. Fiske, S. T., Glick, P., Cuddy, A. C., & Xu, J. (1999a). Ambivalent content of stereotypes, predicted by social structure: Status and competition predict competence and warmth. Presented at XXIIth General Meetion of European Association of Experimental Social Psychology, Oxford, July 1999 and at Society for Experimental Social Psychology Conference, St. Louis, October, 1999. Fiske, S. T., & Neuberg, S. L. (1990). A continuum model of impression formation: From categorybased to individuating processes as a function of information, motivation, and attention. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 1–108). San Diego, CA: Academic Press. Fiske, S. T., Xu, J., Cuddy, A. J. C., & Glick, P. (1999b). Respect versus liking: Status and interdependence underlie ambivalent stereotypes. Journal of Social Issues, 55, 473–489. 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). San Diego, CA: Academic Press. Glick, P. (in press). Sacrificial lambs dressed in wolves’ clothing: Envious prejudice, ideology, and the scapegoating of the Jews. To appear in L. S. Newman & R. Erber (Eds.), What social psychology can tell us about the Holocaust. New York: Oxford Univ. Press. Glick, P., Diebold, J., Bailey-Werner, B., & Zhu, L. (1997). The two faces of Adam: Ambivalent sexism and polarized attitudes toward women. Personality and Social Psychology Bulletin, 23, 1323–1324.
AMBIVALENT SEXISM
185
Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70, 491–512. Glick, P., & Fiske, S. T. (1997). Hostile and benevolent sexism: Measuring ambivalent sexist attitudes toward women. Psychology of Women Quarterly (Special Issue: Measuring attitudes toward appropriate roles for men and women), 21, 119–135. Glick, P., & Fiske, S. T. (1999). The Ambivalence toward Men Inventory: Differentiating hostile and benevolent beliefs about men. Psychology of Women Quarterly, 23, 519–536. Glick, P., Fiske, S. T., Lameiras, M., Brunner, A., Willemsen, T. M., L´opez, W. L., Castro, Y. R., & Sotelo, M. J. (2000). Cross-cultural validation of the Ambivalence toward Women Inventory. Unpublished data. Lawrence University, Appleton, WI. Glick, P., Fiske, S. T., Mladinic, A., Saiz, J, Abrams, D., Masser, B., Adetoun, B., Osagie, J., Akande, A., Alao, A., Brunner, A., Willemsen, T. M., Chipeta, K., Dardenne, B., Dijksterhuis, A., Wigboldus, D., Eckes, T., Six-Materna, I., Exp´osito, F., Moya, M., Foddy, M., Kim, H-J., Lameireas, M., Sotelo, M. J., Mucchi-Faina, A., Romani, M., Sakalli, N., Udegbe, B., Yamamoto, M., Ui, M., Ferreira, M. C., & L´opez, W. L. (2000). Beyond prejudice as simple antipathy: Hostile and benevolent sexism across cultures. Journal of Personality and Social Psychology, 79, 763– 775. Glick, P., & Fiske, S. T. (2001). Ambivalent stereotypes as legitimizing ideologies: Differentiating paternalistic and resentful prejudice. In J. T. Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations. New York: Cambridge Univ. Press. Glick, P., & Hilt, L. (2000). From combative children to ambivalent adults: The development of gender prejudice. In T. Eckes & M. Trautner (Eds.), Developmental social psychology of gender. Hillsdale, NJ: Erlbaum. Guttentag, M., & Secord, P. (1983). Too many women? Beverly Hills, CA: Sage. Hamilton, D., & Fallot, R. (1974). Information salience as a weighting factor in impression formation. Journal of Personality and Social Psychology, 30, 444–448. Harris, M. (1991). Cultural anthropology (3rd ed.). New York: HarperCollins. Hochschild, A. (1998). King Leopold’s Ghost: A story of greed, terror, and heroism in colonial Africa. Boston: Houghton-Mifflin. Hoffman, C., & Hurst, N. (1990). Gender stereotypes: Perception or rationalization? Journal of Personality and Social Psychology, 58, 197–208. Iazzo, A. N. (1983). The construction and validation of the Attitudes Toward Men Scale. Psychological Record, 33, 371–378. Jackman, M. R. (1994). The velvet glove: Paternalism and conflict in gender, class, and race relations. Berkeley, CA: Univ. of California Press. Jackman, M. R. (2001). License to kill: Violence and legitimacy in expropriative social relations. In J. T. Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations. New York: Cambridge Univ. Press. Jaimeson, D. W., Lydon, J. E., & Zanna, M. P. (1987). Attitude and activity preference similarity: Differential bases of attraction for low and high self-monitors. Journal of Personality and Social Psychology, 53, 1052–1060. Jankowink, W., & Ramsey, A. (2000). Femme fatale and status fatale: A cross-cultural perspective. Cross-Cultural Research, 34, 57–69. Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental Social Psychology, 3, 1–24. J¨oreskog, K., & S¨orbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Hillsdale, NJ: Erlbaum. Josephs, R. A., Markus, H. R., & Tafarodi, R. W. (1992). Gender and self-esteem. Journal of Personality and Social Psychology, 63, 391–402.
186
PETER GLICK AND SUSAN T. FISKE
Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system-justification and the production of false-conciousness. British Journal of Social Psychology, 33, 1–27. Jost, J. T., Burgess, D., & Mosso, C. (2001). Conflicts of legitimation among self, group, and system: The integrative potential of system-justification theory. In J. T. Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations. New York: Cambridge Univ. Press. Katz, I. (1981). Stigma: A social psychological analysis. Hillsdale, NJ: Erlbaum. Kenrick, D. T., & Trost, M. R. (1993). The evolutionary perspective. In A. E. Beall & R. J. Sternberg (Eds.), The psychology of gender (pp. 148–178). New York: Guilford. Kilianski, S. E., & Rudman, L. A. (1998). Wanting it both ways: Do women approve of benevolent sexism? Sex Roles, 39, 333–352. Kinder, D. R., & Sears, D. O. (1981). Symbolic racism versus racial threats to “the good life.” Journal of Personality and Social Psycology, 40, 414–431. LaFrance, M., & Woodzicka, J. A. (1998). NO loughing matter: Women’s verbal and nonverbal reactions to sexist humor. In J. K. Swim & C. Stangor (Eds.), Prejudice: The target’s perspective (pp. 61–79). San Diego, CA: Academic Press. LeVine, R. A., & Campbell, D. T. (1972). Ethnocentrism: Theories of conflict, ethnic attitudes, and group behavior. New york: Wiley. Lin, M., & Fiske, S. T. (1999). Attitudes toward Asian-Americans. Unpublished manuscript. University of Massachusetts at Amherst. Lydon, J. E., Jaimeson, D. W., & Zanna, M. P. (1988). Interpersonal similarity and the social and intellectual dimensions of first impressions. Social Cognition, 6, 269–286. Lutz, S. E., & Ruble, D. N. (1995). Children and gender prejudice: Context, motivation, and the development of gender conceptions. Annals of Child Development, 10, 131–166. Maccoby, E. E. (1988). Gender as a social category. Developmental Psychology, 24, 755–765. Maccoby, E. E. (1990). Gender and relationships: A developmental account. American Psychologist, 45, 513–520. Maccoby, E. E., & Jacklin, C. N. (1987). Gender segregation in childhood. In E. H. Reese (Ed.), Advances in child development and behavior (Vol. 20, pp. 239–287). New York: Academic Press. MacDonald, T. K., & Zanna, M. P. (1998). Cross-dimension ambivalence toward social groups: Can ambivalence affect intentions to hire feminists ? Personality and Social Psychology Bulletin, 24, 427–441. Masser, B., & Abrams, D. (1999). Contemporary sexism: Relationships among hostile sexism, benevolent sexism, and neosexism. Psychology of Women Quarterly, 23, 503–517. McConahay, J. B. (1986). Modern racism, ambivalence, and the Modern Racism scale. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 91–125). San Diego, CA: Academic Press. Moya, M., Exp´osito, F., & Casado, P. (1999). Women’s reactions to hostile and benevolent sexist situations. Oxford, UK: European Association of Experimental Social Psychology. 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. Paulhus, D. L. (1988). Assessing self-deception and impression management in selfreports: The Balanced Inventory of Desirable Responding., University of British Columbia, Vancouver, British Columbia, Canada. Peeters, G., & Czapinski, J. (1990). Positive-negative asymmetry in evaluations: The distinction between affective and informational negativity effects. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 1, pp. 33–60). New York: Wiley. Pomeroy, S. B. (1975). Goddesses, whores, wives and slaves: Women in classical antiquity. New York: Schocken.
AMBIVALENT SEXISM
187
Powlishta, K. K. (1995a). Gender bias in children’s perceptions of personality traits. Sex Roles, 32, 17–28. Pratto, F. (1996). Sexual politics: The gender gap in the bedroom, the cupboard, and the cabinet. In D. M. Buss & N. M Malamuth (Eds.), Sex, power, and conflict (pp. 179–230). New York: Oxford Univ. Press. Pryor, J. B., Giedd, J. L., & Williams, K. B. (1995). A social-psychological model for predicting sexual harassment. Journal of Social Issues, 51, 69–84. Ridgeway, C. (1992). Gender, interaction, and inequality. New York: Springer-Verlag. Ridgeway, C. L. (2001). The emergence of status beliefs: From structural inequality to legitimizing ideology. In J. T. Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations. Cambridge, UK: Cambridge Univ. Press. Rose, S., & Frieze, I. H. (1993). Young singles’ contemporary dating scripts. Sex Roles, 28, 499–509. Rosenberg, S., Nelson, C., & Vivekananthan, P. S. (1968). A multidimensional approach to the structure of personality impressions. Journal of Personality and Social Psychology, 9, 283–294. Ross, L., Amabile, T. M., & Steinmetz, J. L. (1977). Social roles, social control, and biases in the social perception process. Journal of Personality and Social Psychology, 35, 485–494. Ruble, D. N., & Martin, C. L. (1988). Gender development. In W. Damon (Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology. Vol. 3: Social, emotional, and personality development (5th ed., pp. 933–1016). New York: Wiley. Rudman, L. A. (1998). Self-promotion as a risk factor for women: The costs and benefits of counterstereotypical impression management. Journal of Personality and Social Psychology, 74, 629– 645. Rudman, L. A., & Glick, P. (1999). Feminized management and backlash toward agentic women: The hidden costs to women of a kinder, gentler image of middle-managers. Journal of Personality and Social Psychology, 77, 1004–1010. Rudman, L. A., & Glick, P. (in press). Prescriptive gender stereotypes and backlash toward agentic women. Journal of Social Issues. Rudman, L. A., Greenwald, A. G., & McGhee, D. E. (in press). Implicit self-concept and evaluative implicit gender stereotypes: Self and ingroup share desirable traits. Personality and Social Psychology Bulletin. Rudman, L. A., & Heppen, J. (2000). Some day my prince will come: Implicit romantic fantasies and women’s avoidance of power. Unpublished manuscript. Rutgers University. Rudman, L. A., & Kilianski, S. E. (2000). Implicit and explicit attitudes toward female authority. Personality and Social Psychology Bulletin, 26, 1315–1328. Runciman, W. G. (1966). Relative deprivation and social justice: A study of attitudes to social inequality in twentieth century England. Berkeley, CA: Univ. of California Press. Sherif, M. (1966). In common predicament: Social psychology of intergroup conflict and cooperation. Boston: Houghton-Mifflin. Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. Cambridge, UK: Cambridge Univ. Press. Sidanius, J., Pratto, F., & Bobo, L. (1994). Social dominance orientation and the political psychology of gender: A case of invariance? Journal of Personality and Social Psychology, 67, 998–1011. Six, I., & Eckes, T. (1991). A closer look at the complex structure of gender stereotypes. Sex Roles, 24, 57–71. Smuts, B. (1996). Male aggression against women: An evolutionary perspective. In D. M. Buss & N. M. Malamuth (Eds.), Sex, power, and conflict (pp. 231–268). New York: Oxford Univ. Press. Spears, R., & Manstead, A. S. R. (1989). The social context of stereotyping and differentiation. European Journal of Social Psychology, 19, 101–121. Spence, J. T., & Buckner, C. E. (2000). Instrumental and expressive traits, trait stereotypes, and sexist attitudes. Psychology of Women Quarterly, 24, 44–62.
188
PETER GLICK AND SUSAN T. FISKE
Spence, J. T., & Helmreich, R. (1972). The attitudes toward women scale. JSAS Catalog of Selected Documents in Psychology, 2, 1–51. Staub, E. (1989). The roots of evil: The origins of genocide and other group violence. Cambridge, UK: Cambridge Univ. Press. Straus, M. A., Gelles, R. J., & Steinmetz, S. K. (1980). Behind closed doors: Violence in the American family. Garden City, NY: Doubleday/Anchor. Swim, J. K., Aikin, K. J., Hall, W. S., & Hunter, B. A. (1995). Sexism and racism: Oldfashioned and modern prejudices. Journal of Personality and Social Psychology, 68, 199–214. Swim, J. K., & Cohen, L. L. (1977). Overt, covert, and subtle sexism: A comparison between the attitudes toward women and modern sexism scales. Psychology of Women Quarterly, 21, 103–118. Swim, J. K., & Stangor, C. (Eds.). (1998). Prejudice: The target’s perspective. San Diego, CA: Academic Press. Tajfel, H. (1981). Social identity and intergroup relations. London: Cambridge Univ. Press. Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In J. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24). Chicago: Nelson. Tougas, F., Brown, R., Beaton, A. M., & Joly, S. (1995). Neo-sexism: Plus c¸ a change, plus c’est pariel. Personality and Social Psychology Bulletin, 21, 842–849. Thompson, M. M., Zanna, M. P., & Griffin, D. W. (1995). Let’s not be indifferent about (attitudinal) ambivalence. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 361–386). Mahwah, NJ: Erlbaum. Triandis, H. C., & Mar´ın, G. (1983). Etic plus emic versus pseudoetic: A test of the basic assumption of contemporary cross-cultural psychology. Journal of Cross-Cultural Psychology, 14, 489–500. Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell (Ed.), Sexual selection and the decent of man 1871–1971 (pp. 136–179). Chicago: Aldine. Twenge, J. M. (1997). Attitudes toward women 1970–1995: A meta-analysis. Psychology of Women Quarterly, 21, 35–51. United Nations Development Programme. (1998). Human development report 1998. New York: Oxford Univ. Press. Weiner, R. L., Hurt, L., Russell, B., Mannen, K., & Gasper, C. (1997). Perceptions of sexual harassment: The effects of gender, legal standard, and ambivalent sexism. Law and Human Behavior, 21, 71–93. Woolf, V. (1929/1981). A room of one’s own. New York: Harcourt Brace Jovanovich. Wright, S. C. (2001). Individual upward mobility and the perceived legitimacy of intergroup relations. In J. T. Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations. New York: Cambridge Univ. Press. Yee, M., & Brown, R. (1994). The development of gender differentiation in young children. British Journal of Social Psychology, 33, 183–196. Young-Breuhl, E. (1996). The anatomy of prejudices. Cambridge, MA: Harvard Univ. Press. Zanna, M. P., & Rempel, J. K. (1988). Attitudes: A new look at an old concept. In D. Bar-Tal & A. W. Kruglanski (Eds.), The social psychology of knowledge (pp. 315–334). Cambridge, UK: Cambridge Univ. Press.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
G. Daniel Lassiter Andrew L. Geers Patrick J. Munhall Ian M. Handley Melissa J. Beers
The medium is the message. (Marshall McLuhan)
The selectivity of any medium may lead to its use having influences of which the user may not always be conscious, and which may not have been part of the purpose in using it. (Daniel Chandler)
The camera . . . may presume, intrude, trespass, distort, exploit, and, at the farthest reach of metaphor, assassinate. (Susan Sontag)
I. Introduction There is a crisis within the American system of justice. According to the Death Penalty Infomation Center, since 1973, 87 death-row inmates have been exonerated—some only days prior to their scheduled executions—because of newly discovered evidence. In an opinion handed down by an Illinois Supreme Court Justice concerning a death-row inmate’s appeal, it was emphatically stated that: ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 33
189
C 2001 by Academic Press, Inc. Copyright All rights of reproduction in any form reserved. 0065-2601/01 $35.00
190
G. DANIEL LASSITER ET AL. The system is not working. Innocent people are being sentenced to death. If these men dodged the executioner [13 of the 87 released death-row inmates were in Illinois state prisons], it was only because of luck and the dedication of the attorneys, reporters, family members and volunteers who labored to win their release. They survived despite the criminal justice system, not because of it . . . . One must wonder how many others have not been so fortunate. (quote obtained from the Chicago Tribune Internet Edition, November, 14, 1999)
The problem, however, is not limited to one or two states, it appears pervasive. An unprecedented systematic examination of several thousand capital-sentence appeals (Liebman, Fagan, & West, 2000, p. i) reported that during a 23-year period (1973–1995) the overall rate of prejudicial error in the American capital punishment system was 68%. In other words, courts found serious, reversible error in nearly 7 of every 10 of the thousands of capital sentences that were fully reviewed during that period. Capital trials produce so many mistakes that it takes three judicial inspections to catch them—leaving grave doubt whether we do catch them all. (emphasis in original)
These chilling facts have recently captured national attention and have galvanized various members of the legal community to generate possible reforms to the system that could prevent innocent people from being imprisoned, or worse, executed. Citing a “shameful record of convicting innocent people and putting them on death row,” Governor George Ryan of Illinois took the bold step of suspending executions in his state until a special commission can thoroughly study the flaws in the system. The director of the American Civil Liberties Union’s Capital Punishment Project has called on other death-penalty states to impose similar moratoriums until solutions can be found. Some factors likely contributing to the current dismal state of affairs have already been identified—for example, egregiously incompetent defense lawyers, erroneous eyewitness accounts, scientifically unreliable evidence, and prosecutorial misconduct (Dwyer, Neufeld, & Scheck, 2000). In addition, many mistakes in the judicial process can be traced to the interrogation phase of criminal investigations, where coerced or false confessions are sometimes extracted from detained crime suspects (cf. Dwyer et al., 2000). A “simple” solution has been advanced to correct this particular problem: videotape all custodial interrogations. Supporting the widespread implementation of this seemingly easy fix, Barry Scheck, cofounder of the Innocence Project, which is dedicated to obtaining the release of wrongfully convicted prisoners with new DNA tests and evidence, stated that with videotapes “there’s no dispute later—what did the person say, what didn’t he say, was it coerced, was it not coerced?” (comments aired on “ABC World News Tonight,” May 11, 2000). Similarly, the Attorney General of Illinois has argued that the “use of videotape will materially advance the interests of all parties to the
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
191
criminal justice process by providing the most reliable evidence of what is said by and to a subject, and under what circumstances” (quoted in the Chicago Tribune Internet Edition, October 28, 1999). Although these remarks imply that videotaping police interrogations and confessions will remedy, at least, one cause of errors in the capital-punishment system, a substantial theoretical and empirical literature on how people process and evaluate information suggests there may be a fly in the ointment. More specifically, research derived largely from attribution theory points to the potential for a subtle, but nonetheless serious, bias associated with the use of videotaped confessions as evidence in courts of law. In this chapter, we briefly review the relevant body of work that gave rise to the above assertion and then describe findings from a programmatic series of studies that demonstrate that confession evidence presented in a videotaped format may indeed, in certain instances, introduce an undesirable bias (if not outright error) in the evaluation of such evidence by trial decision-makers. Results from other experiments that examined the basic processing mechanisms underlying this bias are also reported. Finally, the policy implications of the present research for our system of jurisprudence are discussed.
II. Confession Evidence—Background In criminal trials, fact finders (judges and jurors) make decisions based on an evaluation of the evidence presented. The kind of evidence that possibly has the greatest impact on the decision making of these trial fact finders is a defendant’s prior admission of guilt (Cohn & Udolf, 1979; Kassin & Neumann, 1997; Wigmore, 1970). In fact, according to McCormick (1972, p. 316) the probative value of a confession is so great that its introduction “makes the other aspects of a trial in court superfluous.” This sentiment, when considered together with estimates that admissions of guilt make their way into as many as 68% of criminal trials (see Kassin & Wrightsman, 1985), suggests that the outcome of the majority of such legal proceedings is largely determined by confession evidence. Given its clearly significant role in the administration of criminal justice, there is surprisingly little empirical research devoted to how confession evidence is actually evaluated by trial decision-makers (cf. Kassin, 1997). Kassin and Wrightsman and their colleagues (Kassin & McNall, 1991; Kassin & Sukel, 1997; Kassin & Wrightsman, 1980, 1981, 1985) have generated one of the few systematic programs of research investigating this important issue (see Kassin, 1997, and Wrightsman & Kassin, 1993 for reviews of this program of research). The factors that influence fact finders’ judgments concerning the voluntary status of a confession has been the focus of much of their work. The law requires that before a confession can be treated as evidence in the courtroom, the determination
192
G. DANIEL LASSITER ET AL.
must be made that it was voluntarily given rather than the result of some form of coercion—for example, a threat of punishment or a promise of leniency (Grano, 1993, and Kassin & Wrightsman, 1985, provide discussions of the law pertaining to the use of confessions). Depending on the jurisdiction, this issue of voluntariness is usually decided by the presiding judge or ultimately by the jury (see Kamisar, LaFave, & Israel, 1994, or Kassin & Wrightsman, 1985, for more detailed information concerning the procedures for determining voluntariness). In instances of the latter, the jurors have to be convinced that a confession was made freely and intentionally, otherwise they are instructed to disregard it entirely (Mathes & DeVitt, 1965). Although a U.S. Supreme Court decision ( Lego v. Twomey, 1972) is based on the assumption that jurors are readily capable of differentiating voluntary from involuntary confessions and thereby discounting the latter, the research evidence is far less optimistic. Kassin and Wrightsman (1980, 1981) had mock jurors read a detailed transcript of a criminal trial. In one version of the trial, the defendant was said to have confessed to the crime in response to a threat of punishment and in another version to a promise of leniency. As noted above, the law considers both of these strategies for eliciting confessions coercive. Yet, Kassin and Wrightsman’s studies demonstrated that mock jurors were not able to totally disregard confession evidence that resulted from a promise of leniency. More specifically, mock jurors who read that the confession followed a threat of punishment judged both the confession to be involuntary and the defendant to be not guilty, whereas mock jurors who read that the confession followed a promise of leniency judged the confession to be involuntary, but rendered a guilty verdict anyway. A more recent study by Kassin and McNall (1991) demonstrates that if a confession is elicited by an interrogator’s use of a minimization strategy—that is, “a ‘soft-sell’ technique in which the interrogator tries to lull the suspect into a false sense of security by offering sympathy, tolerance, face-saving excuses, and even moral justification, by blaming a victim or accomplice, by citing extenuating circumstances, or by playing down the seriousness of the charges” (p. 235)— mock jurors tend to react in the same manner as they do to admissions of guilt following promises of leniency, namely judging the confession to be less than voluntary, but still viewing the confessor as largely culpable for the crime. Kassin and Wrightsman (1980, 1981) have labeled this pattern of results the positive coercion bias and have noted that it is consistent with the literature on attribution which indicates that individuals tend to view behaviors enacted to secure a positive outcome as more freely and intentionally caused by an actor than equivalent behaviors enacted to avoid a negative outcome (Bramel, 1969; Kelley, 1971; Wells, 1980). How concerned should we be that trial fact finders may fall prey to judgment errors such as the positive coercion bias? Very. Techniques like minimization are in fact endorsed in police manuals that are used to train police detectives on how
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
193
to conduct an “effective” interrogation (e.g., Inbau, Reid, & Buckley, 1986), and observational studies confirm that detectives do indeed employ the minimization tactic, as well as a complementary strategy called maximization, quite frequently in their interrogations of crime suspects (Gudjonsson, 1992; Leo, 1992, 1996b; Wald, Ayres, Hess, Schantz, & Whitebread, 1967). Maximization is a “hard-sell” approach designed to elicit a confession by sheer intimidation. As described by Kassin (1997, p. 223): This intimidation is achieved by overstating the seriousness of the offense and the magnitude of the charges and even by making false or exaggerated claims about the evidence (e.g., by staging an eyewitness identification or a rigged lie-detector test, by claiming to have fingerprints or other types of forensic evidence, or by citing admissions that were supposedly made by an accomplice).
Whereas courts reject confessions that are obtained by direct threats of punishment or promises of leniency, confessions that result from threats or promises that are merely implied [as is the case with the minimization and maximization interrogation strategies (Kassin & McNall, 1991)] are often ruled voluntary and thus admissible as evidence at trial (Ayling, 1984; Heavner, 1984; Sasaki, 1988; Thomas, 1979; White, 1979). That confessions produced in this manner are being used to convict defendants seems unconscionable in light of the growing belief that such “subtle” or psychologically oriented interrogation approaches may actually elicit false confessions from the truly innocent. Dwyer et al. (2000, p. 89) in their book, Actual Innocence, argued that minimization and maximization are high-risk techniques, noting: A suspect is told he is being fitted for a first-degree murder charge, but the detective thinks there might be a better explanation for his actions. Perhaps the killing was an accident or was done in self-defense. This often leads guilty suspects to confess. And under circumstances of high pressure, with vulnerable people, the “maximization–minimization” technique also provides a strong incentive for innocent people to make false admissions.
Evidence is accumulating to support such claims (Ofshe, 1989; Ofshe & Leo, 1997). Leo and Ofshe (1998) reviewed 60 cases involving alleged police-induced false confessions and concluded that in 48% of these cases the false confession was instrumental in producing a wrongful conviction—which in one instance, they claim, led eventually to a wrongful execution! (Cassell, 1999, has challenged some of Leo and Ofshe’s conclusions.) These false confessions arose not from thirddegree torture tactics, but from the kind of psychological coercion personified in techniques like minimization and maximization (cf. Leo, 1992, 1996b). A recent experimental demonstration conducted by Kassin and Kiechel (1996) provides the most compelling proof to date of the power of psychological pressure to induce false confessions. Kassin and Kiechel (1996, p. 126) found that psychologically
194
G. DANIEL LASSITER ET AL.
based methods of influence commonly employed by interrogators led innocent people “to confess to an act they did not commit and, more important, to internalize the confession and perhaps confabulate details in memory consistent with that new belief.” These findings leave no doubt that the apparent inability of trial fact finders to adequately detect and/or adjust for these types of coercive pressures is reason for great concern.
III. Presentation Format of Confession Evidence: The Growing Emphasis on Videotape The type of interrogation pressure used to induce an admission of guilt is but one factor that may bias the evaluation of confession evidence. Another factor that surprisingly could have a systematic and pernicious influence on the evaluation of confession evidence is simply the manner in which that evidence is presented in the courtroom. Before elaborating on this possibility, we first briefly describe the rapid and dramatic change that has been taking place with respect to the typical presentation format of confession evidence. Until the 1980s, most confession evidence was recorded and presented in either a written or audiotaped format. However, as a result of the advances that have taken place in videotape technology during the past 2 decades—for example, improvements in the quality, portability, and cost of videotape equipment—lawenforcement agencies throughout the country have begun videotaping interrogation sessions and any admissions of guilt that such interrogations might yield (Cutler, 1988; Domash, 1985). For example, in 1983 the district attorney’s office in one borough of New York City alone estimated that it would use videotaped confessions on approximately 500 different occasions (“Smile, You’re on the D.A.’s Camera,” 1983). In a 1992 report to the National Institute of Justice, Geller presented national survey data indicating that a third of law enforcement agencies in the United States were videotaping some interrogations in the early 1990s. Furthermore, Geller (1992) reported that 97% of all departments in the nation that were videotaping either confessions or full interrogations found the procedure to be “very useful” or “somewhat useful.” The San Diego police expressed strong support for the practice, stating, “Not using video would be like not using state-of-the-art fingerprint analysis equipment. If better technology comes along, and its cost is reasonable, police should experiment with it if there is a reasonable chance that it can assist them in their work” (Geller, 1992, p. 153). Geller (1992, p. 154) concluded from his data that “the videotaping of suspect statements is a useful, affordable step on the road toward a more effective, efficient, and legitimate criminal justice system.” He also noted that “excluding the smallest agencies, the percentage of departments
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
195
videotaping confessional evidence will likely exceed 50 percent within a few years.” There are already two states—Alaska and Minnesota—in which videotaping interrogations is required. As of this writing, Illinois is considering a bill to make videotaping mandatory as well. The practice of videotaping police interrogations has many proponents in the legal community as well as in allied fields (Cassell, 1996; Dwyer et al., 2000; Gudjonsson, 1992; Johnson, 1997; Leo, 1996a), and it appears only a matter of time before the videotaped format becomes the norm for introducing confession evidence at trial. In fact, this growing emphasis on videotape technology within the criminal justice establishment is so pervasive that the Institute of Police Technology and Management has initiated courses to train police personnel on how to use videotaping to record and present lineups, crime scene descriptions, surveillance footage, and various other forms of evidence in addition to confessions (Cutler, 1988). Those who advocate videotaping interrogations usually argue that the presence of the camera will (1) deter the use of coercive methods to induce confessions and (2) provide a more complete and objective record of the interrogation so that judges and jurors can evaluate more thoroughly and accurately the voluntariness and veracity of any confession. At least one proponent is so sure of the soundness of the videotaping procedure that he has gone so far as to argue that legally required Miranda warnings to suspects concerning their rights to silence and counsel can be dispensed with if interrogations are routinely videotaped (Cassell, 1996). In the United States and many other countries, interrogations are typically recorded with the camera positioned behind the interrogator and focused squarely on the suspect (Geller, 1992; Kassin, 1997). Positioning the camera in this manner seems straightforward and logical because trial fact finders presumably need to see directly what the suspect is saying and doing to best assess the voluntariness and veracity of his or her statements. The rub, however, is that judgments of voluntariness may be influenced by the camera’s perspective.
IV. Summary of the Literature on Point-of-View/Salience Effects in Causal Attribution The basis for this disturbing suggestion lies in the extensive scientific literature concerning how people go about attributing causality to the behaviors and events that they observe in their environment. Research and theory on this attribution process and the factors that influence it has been accumulating for half a century. In his seminal theoretical work on the topic, Heider (1944, 1958) argued that people are motivated to determine the causes of the events happening around them. According to Heider, knowing the cause of some event or another person’s
196
G. DANIEL LASSITER ET AL.
behavior provides people with a sense of control and predictability—a necessity, if they are to interact effectively with their environment. Heider pointed out that “fundamental to the question of why someone behaves as he or she does . . . is whether the locus of causality for that behavior is in the person (internal) or in the environment (external), or both” (Fiske & Taylor, 1991, p. 25). In developing his theory of naive epistemology, Heider (1958, p. 54) was perhaps the first to note that “behavior . . . has such salient properties it tends to engulf the total field rather than be confined to its proper position as a local stimulus whose interpretation requires the additional data of a surrounding field.” This insight foreshadowed later empirical findings demonstrating that people often strongly favor internal attributions over external attributions for the behaviors they observe. Subsequent theoretical work derived from Heider’s initial formulation sought to spell out more clearly how people ought to proceed in order to rationally identify the cause or causes of a given event or behavior (e.g., Jones & Davis, 1965; Jones, Kanouse, Kelley, Nisbett, Valins, & Weiner, 1972; Kelley, 1967, 1972; Weiner, Frieze, Kukala, Reed, Rest, & Rosenbaum, 1972). Although empirical tests indicated that people do sometimes make attributions in the logical and reasonable ways prescribed by such rational models (e.g., McArthur, 1972), they also clearly showed that, in certain instances, people exhibit strong systematic biases in their attributional analyses—biases of which they are oftentimes seemingly unaware (e.g., Jones & Harris, 1967; Jones & Nisbett, 1972; Jones, Rock, Shaver, Goethals, & Ward, 1968; L. Ross, Amabile, & Steinmetz, 1977; L. Ross, Greene, & House, 1977; M. Ross & Sicoly, 1979). One such bias—consistent with Heider’s earlier intuition—is the pronounced inclination for people to overemphasize internal causes for another person’s behavior (e.g., intentions and dispositions) and to seemingly neglect aspects of the surrounding situation that alone could sufficiently account for the behavior (cf. Gilbert & Malone, 1995; Jones, 1979; Nisbett & L. Ross, 1980; L. Ross, 1977). The documentation of this so-called “fundamental attribution error” (L. Ross, 1977) and other biases like it led investigators to focus their theoretical and empirical efforts on trying to account for these various intriguing deviations from rationality. Jones and Nisbett’s (1972) stimulating work on actor–observer differences in attribution suggested one possible factor—a person’s perceptual perspective or point of view—that might explain or at least contribute to the fundamental attribution error. The idea that something as trivial and nondiagnostic as a person’s observational vantage point could affect his or her attributions for some event at first blush probably seemed implausible to many. Nonetheless, numerous studies have since been conducted which convincingly demonstrate that people’s attributions of causality are indeed strongly influenced, quite literally, by their point of view. This so-called “salience effect” specifically indicates that there is a pervasive tendency for people observing a social interaction to overestimate the causal role of the individual who is most visually prominent—that is, the one who can be
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
197
seen most clearly (see McArthur, 1981, and Taylor & Fiske, 1978 for extensive reviews of this literature). Considered together, the findings regarding the fundamental attribution error and salience effects lead to the conclusion that observers routinely fail to appreciate fully the causal influence of external factors or pressures on another individual’s behavior and that the problem is compounded when those situational forces are rendered even less visible or salient by virtue of observers’ visual perspective.
V. Implications of the Salience Effects Literature for the Use of Videotaped Confessions The developments in attribution theory and research we have just described were largely achieved prior to the actual existence of videotaped confessions. Nonetheless, it is our contention that they have alarming implications for the use of videotaped confessions as a tool to facilitate the administration of justice. That is, it seems reasonable to assume that in arriving at a judgment concerning possible coercive influences, observers of an interrogation might first try to determine who or what caused, or was the most responsible for, the act of confessing. To the extent that this assumption is valid, the above findings suggest that the use of videotaped confessions could produce judgments of coercion which vary systematically with the camera’s point of view. More specifically, observers might judge, all things being equal, that a relatively small degree of coercion was used when the camera focused primarily on the confessor (because the act of confessing would presumably be largely attributed to the confessor), that a relatively large degree of coercion was used when the camera focused primarily on the interrogator (because the act of confessing would presumably be largely attributed to the interrogator), and that a relatively moderate degree of coercion was used when the camera focused on both participants equally. This hypothesis, derived from the attribution literature, takes on even greater significance when one considers, as noted above, that most interrogations are videotaped with the camera focused primarily on the suspect or confessor (Geller, 1992; Kassin, 1997). That being the case, it is possible that the use of videotaped confessions is causing judges and/or jurors to be biased to perceive such confessions as voluntary, which in turn could have the detrimental effect of increasing the number of truly coerced or false confessions that are considered as reliable evidence in courts of law. Moreover, a recent U.S. Supreme Court ruling highlights the significance of this issue even further. Prior to 1991, if in the trial process an error was committed that allowed a coerced confession into evidence and the defendant was convicted, an appeal on behalf of the defendant would automatically nullify the verdict and produce a new trial. In a startling reversal of this
198
G. DANIEL LASSITER ET AL.
long-standing precedent, the U.S. Supreme Court ruled in Arizona v. Fulminante (1991) that the improper use of an involuntary confession in a trial resulting in a conviction is not in and of itself sufficient reason to invalidate the conviction. That is, if other evidence in a particular case was adequate to justify a conviction, then the admission of an involuntary confession could be viewed as “harmless error.” There is concern among some legal scholars that this ruling could increase the willingness of prosecutors to introduce as evidence confessions whose voluntary status is dubious (e.g., Kamisar, 1995). Such a possibility suggests that attention to factors that may potentially prejudice determinations of voluntariness is more critical than ever before.
VI. A Program of Research Investigating the Potential for Bias in Videotaped Confessions Because of its clear practical importance to the legal community, we have conducted a program of research aimed at testing the above hypothesis that the ostensibly trivial variable of camera perspective may actually bias people’s evaluations of videotaped confession evidence.1 Several aspects of this research are notable. 1. The research provides an excellent example of what Lassiter and Dudley (1991) call the a priori value of basic research. One value of basic research and theory is that it can be used to address real-world issues in an a posteriori fashion. That is, some threat to the general well-being of society (e.g., rampant aggression) clearly exists before any research is initiated to provide a scientific basis for its possible attenuation or elimination. The second way that basic research and theory can contribute to the maintenance of a healthy society is perhaps less obvious and, in our view, certainly underacknowledged. It is the notion that basic research is also capable of pointing out potential problems that could ultimately arise as a result of society’s ever-changing and increasingly complex nature. That is, the knowledge base provided by basic research (in this case the work on attribution processes) can actually help detect hidden or newly developing problems that might otherwise go unrecognized if such knowledge was unavailable. Thus basic research and theory can be used to address real-world issues in an a priori fashion as well. 2. Wells (1978) has distinguished between system-variable and estimatorvariable research. Wells (1978, p. 1552) defines the former as investigating 1
Scientific evaluation of the feasibility of new technologies and techniques as potential aids in criminal justice administration and operations has proven to be vitally important. In the early 1980s, for example, police investigators turned increasingly to hypnosis in an attempt to enhance the memories of victims and witnesses of crime (Reiser, 1980). Rigorous, systematic examination of this technique, however, subsequently revealed that the use of hypnosis as a forensic tool was fraught with serious problems (Dywan & Bowers, 1983; Laurence & Perry, 1983).
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
199
“variables that are manipulable by the criminal justice system, whereas the latter investigates variables whose influence can be estimated (but not controlled) by the criminal justice system.” Wells (1978, p. 1555) noted that “system-variable research . . . may, as a general rule, have greater applied utility for criminal justice than does estimator-variable research.” Thus another positive feature of the research to be described is that it is system-variable, rather than estimator-variable, research and therefore has the potential to yield information that could ultimately be used to alter the criminal justice system for the better. 3. Bray and Kerr (1982) have suggested that a reasonable approach to addressing the generality or external validity of some effect that has been previously shown to be internally sound or valid “is to conduct a series of carefully planned studies that collectively provide data that determine the limits of generalizability.” We adopted this strategy in our research program. Thus, although no single study will adequately address the external validity question, together they should provide a solid indication of whether the criminal justice system needs to be seriously concerned about how it acquires and utilizes videotaped confession evidence. 4. Diamond (1997) has argued that trial simulations at Stage One of a research program that involve relatively “easy” methods (e.g., using college-student participants and brief stimulus materials) should be followed up with Stage Two research that involves more elaborate, representative methods (e.g., using community adults as participants and extensive videotaped trials as stimuli). This two-stage approach advocated by Diamond was followed in our series of investigations. 5. The present research also includes what we are designating a Stage Three— that is, studies designed to identify the mediator(s) of the point-of-view/salience bias. Gaining a clearer understanding of the psychological mechanism(s) underlying this bias will better enable researchers to develop strategies for combating it. 6. Finally, the research to be described has already had real-world impact. Lani Takitimu, Police Officer in Charge of the National Electronic Interview Unit situated at the Police National Headquarters in Wellington, New Zealand, has informed us that a national policy in New Zealand regarding the videotaping of police interrogations was directly influenced by portions of the work described below (personal communication, November 3, 1993).
VII. Stage One: Establishing the Existence and Robustness of the Camera Perspective Bias in Videotaped Confessions Stage One of the research comprised eight studies that were, for the most part, relatively simple in their design and in the stimulus materials used. The mock confessions that we constructed for Studies 1–5 were designed to be composites of various elements that have been documented to occur in real interrogations or that
200
G. DANIEL LASSITER ET AL.
police manuals advise should occur. None of the stimulus tapes resulting from these staged interrogations and confessions lasted longer than 5 min. (Observational data by Leo, 1996b, suggest that interrogations of this length are not typical, but they do occur.) For Studies 6–8, we developed our confession stimulus from the transcript of an actual police interrogation and it was approximately 30 min in duration. With the exception of Study 7, all of the experiments in Stage One employed only continuous (rating scale) measures of participants’ judgments because they often exhibit greater sensitivity than dichotomous responses and because they are amenable to more powerful parametric analyses. (In a true courtroom, judgments concerning voluntariness and guilt would ultimately be rendered in a dichotomous fashion. However, at this point we were most concerned with being able to detect the bias, if it actually existed. Issues of mundane realism could always be addressed later—and were in Stage Two.) Finally, all participants in Studies 1–8 were college students recruited from psychology department subject pools.
A. STUDY 1: AN INITIAL DEMONSTRATION OF THE BIAS We (Lassiter & Irvine, 1986) began by using three cameras simultaneously to videotape a mock police interrogation. One camera was positioned so that the front of the “suspect” from the waist up and the back of the “detective” (part of his head and one shoulder) were visible. A second camera was positioned in a similar manner, but it was the detective’s front and suspect’s back that were visible. The third camera was positioned so that the sides of both the suspect and detective from the waist up could be seen equally well. All three cameras were set at an approximately 90◦ angle to the vertical plane. During the interrogation, the detective (a male) asks the suspect (a female) several questions about her recent activities. At one point he accuses her of stealing an article of clothing from a shopping center, which she denies. The detective continues his inquiry and ultimately the suspect confesses to the crime. Twenty-four participants were told their task in the experiment would be to assume the role of jurors in a courtroom and to evaluate individually the voluntariness and other aspects of a criminal confession obtained during a police interrogation. Participants were then shown one of the three videotapes of the mock police interrogation. (In this study, the videotapes were black and white; in all remaining studies they were in color.) Following the videotape presentation, participants were asked to indicate to what degree they thought the suspect was coerced into confessing. Participants responded on a 9-point scale, with higher numbers reflecting a judgment that greater coercion was involved. Participants also indicated how confident they were in their coercion judgments on a second 9-point scale (1 = not at all to 9 = completely). Finally, participants filled out several items from which an attribution index was formed. Higher values on this index
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
201
TABLE I MEANS FOR THE DEPENDENT MEASURES (STUDY 1) Camera point of view Measure
Suspect-focus
Equal-focus
Detective-focus
Coerciona Confidenceb Attribution indexc
3.13 6.88 3.50
4.75 6.88 −2.43
6.75 6.63 −4.63
Note. Data from Lassiter & Irvine (1986). a The higher the number, the greater the perceived coercion. b Confidence in coercion judgments with higher numbers indicating greater confidence. c Higher numbers indicate relatively more dispositional (or relatively less situational) attributions for the suspect’s behavior.
indicated that participants’ attributions for the suspect’s behavior (e.g., her level of nervousness and loquacity) were relatively more dispositional (or relatively less situational). Consistent with our hypothesis, it was found that subjects rated the interrogation the least coercive when the camera focused primarily on the suspect, rated it more coercive when the camera focused on the suspect and detective equally, and rated it the most coercive when the camera focused primarily on the detective (see Table I for means on all measures). This linear trend in coercion ratings was significant, thereby providing evidence that the point of view from which a confession is videotaped can have a considerable impact on observers’ judgments of whether that confession was voluntary or coerced. It might be argued that this pattern of results simply reflects the fact that participants were reluctant to say the confession was voluntarily given (i.e., not coerced) when they could not get a good look at the suspect, as was the case in the detective-focus condition and to a lesser extent in the equal-focus condition. This interpretation is rendered implausible, however, by the fact that participants indicated a high degree of confidence in their voluntariness judgments, with no significant differences across conditions. The attribution index revealed that subjects made the most dispositional attributions in the suspect-focus condition, less dispositional attributions in the equalfocus condition, and the least dispositional attributions in the detective-focus condition. Again this linear patterning of means was significant. These data, then, are supportive of the assumption that differences in judgments of coercion are mediated in part by causal attributions, with more dispositional attributions for the suspect’s behavior being associated with a judgment of less coercion or greater voluntariness.
202
G. DANIEL LASSITER ET AL.
B. STUDY 2: AN EXAMINATION OF ADDITIONAL ISSUES In our first study, participants viewed and evaluated only a single interrogation in which a suspect confessed to the crime of shoplifting. An important first step in building on this work, then, was to determine if the pattern of results found in Study 1 generalized to different interrogations and to different crimes. To examine this issue we (Lassiter, Slaw, Briggs, & Scanlan, 1992) devised three new mock interrogations (this time with a male suspect) each concerned with a different crime (i.e., rape, drug trafficking, or burglary). Replication of our Study 1 results across each of these interrogations and crimes would provide further evidence that the camera perspective bias in videotaped confessions is a real and pervasive phenomenon. Study 1 also lacked any nonvideotaped presentation formats that could serve as “control” conditions. Without such comparison groups we cannot know whether the camera focusing on the confessor increased perceptions of voluntariness or whether the camera focusing on the interrogator or on both individuals actually decreased perceptions of voluntariness. As noted in Section III, confession evidence has traditionally been presented in either a written or an audiotaped format. To our knowledge these two presentation formats have been used successfully for many years with no suggestion of any inherent prejudicial impact. For this reason we also included in the present experiment written and audiotaped versions of the confessions to provide a baseline so that the exact nature of any biasing effect of the videotaping procedure could be more clearly established. A third way in which we extended the previous research was to investigate the extent to which any bias in voluntariness judgments resulting from camera point of view would in turn prejudice likelihood-of-guilt assessments. That is, will individuals who perceive a videotaped confession as more voluntary simply because of a particular camera point of view also judge the confessor more likely to be guilty? Although it seems reasonable to assume that judgments of voluntariness and likelihood-of-guilt assessments would be positively correlated, some earlier findings indicate that this may not always be the case. As discussed in Section II, Kassin and Wrightsman (1980, 1981; Kassin & McNall, 1991) have shown that, in certain instances, individuals may recognize that a confession is involuntary, but nonetheless still judge the confessor to be guilty of the confessed crime. Such results clearly indicate that additional data are needed to clarify the relationship between judgments of voluntariness and assessments of guilt. A final focus of Study 2 was to explore a possible individual difference in the extent to which people are susceptible to the biasing effect of camera perspective. That is, individuals who are inclined to process information in a careful and thorough manner might be expected to be relatively unaffected by something as seemingly irrelevant and trivial as camera point of view. An instrument designed to measure this very tendency—the Need for Cognition Scale—has been developed by Cacioppo
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
203
and Petty (1982; Cacioppo, Petty, & Kao, 1984). According to these researchers, individuals who derive great enjoyment from engaging in effortful and extensive cognitive processing are said to be high in need for cognition, whereas individuals who derive little enjoyment from such activities are said to be low in need for cognition (cf. Cacioppo & Petty, 1982; Lassiter, Briggs, & Slaw, 1991). A sizable body of research provides support for this distinction (see Cacioppo, Petty, Feinstein, & Jarvis, 1996, for a review). On the basis of this literature, then, we examined the possibility that the judgments of individuals who exhibit a high (relative to low) need for cognition would be determined primarily by a thoughtful evaluation of the content of an interrogation and confession and therefore would be minimally influenced by the specific format in which such information was presented. The three interrogations begin with the detective questioning the suspect about his whereabouts at a given date and time. Although the suspect initially denies any wrongdoing, the detective uses various ploys (e.g., minimization and maximization) to induce self-incriminating statements from the suspect. The suspect is informed that there is evidence linking him to the crime in question, but the suspect repeatedly denies the accusation. The detective continues to confront the suspect with reasons why he should admit his guilt (e.g., in one interrogation the detective says, “. . . any confession you make now will certainly be held in your favor”). All three mock interrogations end with the suspect finally confessing to the crime under investigation. Audiotapes and transcripts of the interrogations were generated from the videotaped versions (only two camera perspectives—suspect-focus and equal-focus—were used in this study) resulting in four different presentation formats for each of the three mock interrogations. A single voluntariness index was formed by summing three items assessing participants’ (N = 172) perceptions of the extent to which the confession was given freely and intentionally (Cronbach’s ␣ = .74). This index was submitted to a 3 (nature of crime) × 4 (confession-presentation format) analysis of variance (ANOVA). (A preliminary analysis indicated no significant effects of need for cognition.) The results of this two-way ANOVA revealed a significant main effect for the crime factor—participants perceived the drug-trafficking confession to be less voluntary than either the rape or burglary confessions. More important, the ANOVA also revealed a significant main effect for confession-presentation format. The means corresponding to this effect are presented in Table II. As can be seen, the suspect-focus videotape produced judgments of voluntariness that were greater than those produced by the other confession-presentation formats. A planned contrast showed this pattern was reliable. A comparison of the equal-focus videotape, audiotape, and transcript formats indicated no significant differences among these groups in terms of the judgments of voluntariness they produced. Finally, the ANOVA revealed no significant two-way interaction, thereby indicating that the effect of confession-presentation format on voluntariness judgments was consistent across the different crimes and interrogations.
204
G. DANIEL LASSITER ET AL. TABLE II MEANS FOR VOLUNTARINESS INDEX AS A FUNCTION OF CONFESSION PRESENTATION FORMAT (STUDY 2) Suspect-focus videotape 17.32
Equal-focus videotape
Audiotape
Transcript
14.27
14.75
15.42
Note. Higher scores on the voluntariness index indicate judgments of greater voluntariness (possible scores = 3 to 27). (Data from Lassiter et al., 1992.)
Participants indicated their assessments of the suspect’s probable guilt on a single 9-point scale, with higher numbers denoting a greater perceived likelihood of guilt. A three-way ANOVA (Nature of Crime × Confession-Presentation Format × Need for Cognition) performed on these ratings revealed a significant main effect of crime—the suspect was judged more likely to be guilty when he confessed to a rape, as opposed to a drug-trafficking or burglary, charge. The only other reliable effect to emerge from this analysis was a main effect of need for cognition. Participants high in need for cognition (based on a median split) judged the suspect to have a greater likelihood of being guilty than did participants low in need for cognition. The fact that, at the group level, confession-presentation format influenced voluntariness judgments but not likelihood-of-guilt assessments would appear to suggest that the two variables were unrelated in the present study. Further correlational analyses of the data, however, indicated that such a conclusion is not entirely correct. Across all subjects, voluntariness judgments were found to be weakly, but significantly, related to likelihood-of-guilt assessments, r = .18. On the individual level, then, subjects who judged the confession to be more voluntary also assessed that the suspect was more likely to be guilty. The results of Study 2 increased our confidence that, relative to other confessionpresentation formats, suspect-focus videotapes tend to enhance voluntariness judgments. This tendency is robust enough that it affects even those individuals who are presumably predisposed to process information in an effortful and a systematic manner (i.e., those high in need for cognition). Our failure to find that a high need for cognition did not eliminate, or at least attenuate, the camera perspective bias is consistent with two other reported studies (Briggs & Lassiter, 1994) showing that, when viewing a “getting-acquainted” conversation, high-need-for-cognition individuals’ judgments of causality were no more resistant to a salience manipulation than were those of low-need-for-cognition individuals. From a theoretical standpoint these demonstrations of a lack of any effect of need for cognition on susceptibility to the salience bias would seem to imply that the mechanism underlying this bias is not related very strongly to the degree of high-level cognitive effort generated during information processing. From a practical standpoint
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
205
such results suggest that we cannot take comfort in the assumption that a real courtroom situation will engender a high motivation to engage in effortful cognitive processing among trial participants. With respect to determining the voluntariness of videotaped confessions, it appears that a high level of motivation or involvement is no protection against potential bias. (We return to this issue below.) Our Study 2 results, however, do not paint an entirely negative picture with regard to the use of videotaped confessions in the courtroom. Videotaped confessions that focused on both the suspect and the interrogator equally were found to generate voluntariness judgments that were comparable to those based on more traditional presentation formats—that is, audiotapes and transcripts. Thus, it is clear that the videotaping procedure per se is not inherently prejudicial. Rather, it is the manner in which the videotaping procedure is implemented that holds the potential for bias. It appears, then, that the advantages associated with the videotape method—for example, a more detailed record of the interrogation is provided to trial participants—can be maintained without introducing bias if an equal-focus perspective is taken by the video camera. Unfortunately, as noted in Section III, at this point in time this is not the perspective that is typically taken when the videotape method is employed.
C. STUDY 3: DOES DELIBERATION MAKE A DIFFERENCE? A possible safeguard against the camera perspective bias in videotaped confessions that is already present in our current system of justice is the requirement that jurors must deliberate before rendering their judgments. Kaplan (1982; Kaplan & Miller, 1978) has argued, based on an information integration perspective (Anderson, 1974; Kaplan, 1975), that juror biases can be reduced by increasing the weight jurors give to evidential information. The process of deliberation is one way of achieving this goal. That is, according to Kaplan (1982, p. 213), the “advantage of a deliberating jury over a single juror . . . is that among the jurors more legal facts are noticed, remembered, and taken into account. If these facts are then shared in deliberation, more facts will be available to the single juror to counteract the preexisting disposition and/or extralegal information” (cf. Ellsworth, 1989; McCoy, Nunez, & Dammeyer, 1999). An important question addressed in our third study, then, was whether the pointof-view bias in videotaped confessions still persists even after individuals deliberate (Lassiter, Beers, Geers, Handley, Munhall, & Weiland, in press, Study 1). Also, in our last study, the results did not demonstrate conclusively that the bias also affects assessments of guilt. Obviously, the overall applied significance of this line of work can be called into question if camera perspective is not shown ultimately to influence decisions concerning guilt and innocence. Therefore participants’ guilt
206
G. DANIEL LASSITER ET AL.
judgments were collected again in Study 3 and in most of the subsequent studies in Stages One and Two. Three hundred sixty-two participants were run in groups of five or six at a time. After examining the confession evidence (regarding either the burglary or rape crimes used previously), participants were informed that, similar to real jurors, they would now have the opportunity to discuss the issue of the confession’s voluntariness. More specifically, they were asked to determine whether the confession was given freely by the suspect and therefore should be considered valid evidence in court. Participants were told they could discuss whatever they thought would help them decide the voluntariness question. Participants were given as much time as they needed for deliberation. Following the deliberation, participants received separate questionnaires that they were instructed to complete individually. Using the three camera angles employed in Study 1, plus audiotape and transcript formats as in Study 2, we again found the bias in voluntariness judgments observed previously. (Because participant’s individual judgments could no longer be considered independent after they had deliberated, analyses were performed on the mean voluntariness index of the separate groups.) Most important, a significant main effect of camera point of view on likelihood-of-guilt assessments was obtained, with the suspect-focus videotape resulting in higher estimates of guilt than either the equal- or detective-focus videotapes (see Table III for means on both measures). As was the case with the voluntariness judgments, the audiotape and transcript formats yielded guilt assessments comparable to that of the equal-focus videotape. It should be noted that none of the above results was qualified by the nature-of-crime factor. We assumed that, to a large extent, the impact of confession-presentation format on likelihood of guilt assessments was mediated by judgments of voluntariness. To directly test this assumption, we conducted a path analysis following TABLE III MEANS FOR DEPENDENT MEASURES (STUDY 3) Confession-presentation format Suspect-focus videotape
Equal-focus videotape
Detective-focus videotape
Audiotape
Voluntariness index
20.97
19.65
18.48
19.31
19.30
Likelihood of guilt
8.22
8.13
7.55
8.07
7.95
Measure
Transcript
Note. Higher scores on the voluntariness index and guilt measure indicate judgments of greater voluntariness and more probable guilt, respectively. (Data from Lassiter et al., in press, Study 1.)
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
207
Fig. 1. Path diagram and coefficients (standardized beta weights) for Study 3. Solid paths are significant, p < .05. (Data from Lassiter et al., in press, Study 1.)
procedures outlined by Kenny (1979; Kenny, Kashy, & Bolger, 1998). Regression analyses were performed to estimate the magnitude and significance of the path coefficients (standardized beta weights). The resulting values are presented in Fig. 1. Consistent with our assumption, the paths from confession-presentation format to voluntariness judgments and from voluntariness judgments to likelihood-of-guilt assessments were both significant. The direct path from confession-presentation format to likelihood-of-guilt assessments (after partialing out the effect of voluntariness judgments) was not significant. Overall, this analysis suggests that with regard to guilt assessments, the biasing effect of confession-presentation format occurs primarily via its influence on voluntariness judgments. The fact that the point-of-view bias was still obtained after deliberation suggests that the process of exchanging information and discussing one’s views about the evidence with others is not an effective antidote to the prejudicial effect of camera perspective. It is also important to note that, for the second time, the camera perspective that produced judgments that were the most comparable to the written and audiotaped versions of the confessions (i.e., the more traditional presentation formats) was the equal focus.
D. STUDY 4: CAN FOREWARNING ELIMINATE THE BIASING EFFECT OF CAMERA PERSPECTIVE? Informal examinations of the content of the group discussions in Study 3 revealed that the issue of camera perspective never came up in deliberations. If no one thought to bring up this issue, then a lack of awareness of it altogether might explain why individuals are not able to correct or eliminate the influence camera point of view is having on their judgments (cf. Wilson & Brekke, 1994). This insight led us to consider the straightforward question, can people obviate the biasing effect of camera point of view when they are explicitly alerted to its possible prejudicial impact? A fourth study was conducted to provide an answer to this question (Lassiter et al., in press, Study 2).
208
G. DANIEL LASSITER ET AL.
To make the experimental sessions seem more like an actual trial, testimony from two witnesses for the prosecution and two witnesses for the defense was provided in transcript form. (The testimony of the prosecution and defense witnesses was presented before and after the presentation of the confession, respectively.) The strength of the evidence presented by both the defense and the prosecution witnesses was designed to be approximately equivalent. The inconclusive nature of the testimony made the confession the central piece of evidence in the case. Before viewing the videotaped confession, half of the participants were warned about potential effects of watching the videotaped confession from a particular perspective. More specifically, the experimenter said: “Because the confession was videotaped you should be aware that your judgments could be affected by the angle of the camera. In thinking about the videotape, you should focus on what the detective and the defendant actually said and how they behaved. Do not allow the angle of the camera to influence your decision about whether the confession was voluntary or coerced.” This warning was omitted for the remaining participants. All participants (N = 104) then viewed either the suspect-focus or detective-focus version of the videotaped confession (regarding the crime of burglary used previously). As in Studies 1 and 2, no deliberation occurred. Separate 2 (suspect-focus vs detective-focus videotape) × 2 (forewarning vs no forewarning) ANOVAs were performed on participants’ judgments of voluntariness and guilt. Both analyses revealed a significant main effect of camera perspective, indicating that participants who viewed the suspect-focus videotape rated the confession as more voluntary and the defendant more likely to be culpable than did participants who viewed the detective-focus videotape (see Table IV). The attempt to eliminate or at least attenuate the point-of-view bias was unsuccessful as neither the main effect of forewarning nor the two-way interaction attained significance. TABLE IV MEANS FOR THE DEPENDENT MEASURES (STUDY 4) No Forewarning Measure
Forewarning
Suspect-focus
Detective-focus
Suspect-focus
Detective-focus
Voluntariness index
20.52
16.96
18.63
16.48
Likelihood of guilt
8.08
7.24
8.07
7.41
Note. Higher scores on the voluntariness index and guilt measure indicate judgments of greater voluntariness and more probable guilt, respectively. (Data from Lassiter et al., in press, Study 2.)
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
209
Fig. 2. Path diagram and coefficients (standardized beta weights) for Study 4. Solid paths are significant, p < .05. (Data from Lassiter et al., in press, Study 2.)
To determine if voluntariness judgments were mediating the effect of camera perspective on guilt assessments, a path analysis was conducted as in Study 3. The pattern of results was once again consistent with the notion that camera perspective has an indirect effect on guilt assessments (see Fig. 2). That is, the paths from camera focus to voluntariness judgments and from voluntariness judgments to likelihood-of-guilt assessments were both significant, whereas the direct (nonmediated) path from camera focus to likelihood-of-guilt assessments was not.
E. STUDY 5: WILL DIRECTING ATTENTION TO CONTENT DIMINISH THE BIAS? Our very direct and straightforward attempt to eliminate the biasing effect of camera perspective failed in Study 4. This lack of success led us to try a diametrically opposite strategy in our next study. That is, instead of calling attention to the camera perspective and hoping people can minimize its effect on their judgments, we decided in a fifth experiment (Lassiter et al., in press, Study 3) to induce individuals to pay even greater attention to the content of the interrogation and confession. If more of their focus and concentration is on the content and the information revealed therein, people’s judgments may be less swayed by the tangential factor of camera perspective. Prior to presenting the videotaped confession, roughly half of the 87 participants were told to direct their full attention to the content of the interrogation. More precisely, the attention-to-content participants were instructed to identify the important aspects of the interrogation by pressing a hand-held button. Each time the suspect or the detective said or did something significant or informative, participants were to press the button, which tallied their judgments. The experimenter emphasized that it was the participant’s important task to determine what aspects of the interrogation were to be considered significant. The remaining participants did not receive these instructions nor did they engage in the button-pressing task. All participants then viewed either the suspect-focus or
210
G. DANIEL LASSITER ET AL. TABLE V MEANS FOR DEPENDENT MEASURES (STUDY 5) No attention-on-content task Measure
Attention-on-content task
Suspect-focus
Detective-focus
Suspect-focus
Detective-focus
Voluntariness index
20.60
16.85
18.08
15.91
Likelihood of guilt
8.25
7.45
8.29
8.13
Note. Higher scores on the voluntariness index and guilt measure indicate judgments of greater voluntariness and more probable guilt, respectively. (Data from Lassiter et al., in press, Study 3.)
detective-focus videotaped confession (of the burglary crime) and completed the dependent measures. Participants’ responses to the voluntariness and guilt measures were entered into separate 2 (suspect-focus vs detective-focus videotape) × 2 (attention-on-content vs no attention-on-content) ANOVAs. From these analyses, only a significant main effect of the camera perspective emerged such that the suspect-focus participants judged the confession as more voluntary and the suspect as more likely guilty than did detective-focus participants (see Table V). The same path analysis was conducted as before and the results were comparable to those in Studies 3 and 4 (see Fig. 3). That is, the two paths involving voluntariness judgments were both significant, but the remaining direct path was not. Although an examination of the button-pressing responses of attention-oncontent participants indicated they understood the task and that they were taking it seriously, their concentration on identifying the most important and meaningful aspects of the interrogation and confession was not sufficient to prevent the point-of-view bias from affecting their judgments of voluntariness and guilt.
Fig. 3. Path diagram and coefficients (standardized beta weights) for Study 5. Solid paths are significant, p < .05. (Data from Lassiter et al., in press, Study 3.)
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
211
F. STUDY 6: ARE LONGER/CASE-BASED CONFESSIONS BIAS-PROOF? As noted in Section VII, none of the mock confessions constructed for Studies 1–5 were derived entirely from a specific, actual police interrogation and overall they were relatively short in duration. It is possible that the content of a single, case-based police interrogation and confession could contain certain kinds of information or more impactful information that could cause people to give more weight to the content and thus be less likely to be influenced by the camera’s point of view (cf. Kaplan, 1982). We also speculated that the biasing effect of camera perspective may be most likely to occur when the amount of content information available for people to consider is limited—that is, when the confession is brief. If that were indeed the case, then confessions of greater length may be less likely to produce a camera perspective bias. The primary purpose of Study 6 (Lassiter et al., in press, Study 4), then, was to examine both these possibilities by presenting mock jurors with a videotaped confession that was based closely on an actual police interrogation and that was significantly longer in duration (approximately 30 min). The videotaped confession was a recreation of portions of the interrogation and confession of Bradley Page, a college student, who was convicted of the manslaughter of his romantic partner, Bibi Lee, based largely on his disputed confession. (We are very grateful to Richard Leo for providing us with a transcript of the Page interrogation.) Many psychological and legal experts view Page’s confession as an instance of a coerced-compliant confession (cf. Kassin & Wrightsman, 1985) and his ensuing conviction as a miscarriage of justice (e.g., Leo & Ofshe, 1998; Pratkanis & Aronson, 1991; Wrightsman & Kassin, 1993). Elliot Aronson, who testified at Page’s trial as an expert on “noncoercive” persuasion, was given access to audiotapes of the interrogation and provided the following brief account of what essentially transpired while Page was in custody. After inducing Brad to waive his rights to an attorney (“we’re all friends, here, aren’t we?”), the police interrogators had him go over his story several times. During the interrogation, they kept asking him how he could possibly have left his girlfriend alone in the park and driven back home. Brad felt terribly guilty about it, saying several times, “It was the biggest mistake of my life!” Each time they asked the question, his guilt seemed to grow. Finally, the interrogators told Brad that late on the night that Bibi had disappeared he had been seen near the site of the shallow grave [where Lee’s body was recovered] and that his fingerprints had been found on a rock that had been used as the murder weapon. Neither of these statements was true [maximization ploy]. Brad said that he had no recollection of having left his apartment that night and had no idea how his fingerprints could have gotten on the murder weapon (he didn’t even know what the weapon was). But he had no reason to distrust the interrogators, so, understandably, he became terribly confused and asked them if it is possible for a person to “blank it out.” The interrogators informed him that such things were common occurrences and that it might help him relieve his guilty conscience if he closed his eyes and tried to imagine how he might have killed Bibi if he had killed her
212
G. DANIEL LASSITER ET AL. TABLE VI MEANS FOR DEPENDENT MEASURES (STUDY 6) Camera point of view Measure
Suspect-focus
Equal-focus
Detective-focus
Voluntariness index
20.32
18.14
16.29
Likelihood of guilt
8.62
8.34
8.21
Recommended sentence
7.72
7.31
6.57
Note. Higher scores on the voluntariness index and guilt measure indicate judgments of greater voluntariness and more probable guilt, respectively. Higher scores on the sentence recommendation measure signify a more severe sentence. (Data from Lassiter et al., in press, Study 4.)
[minimization ploy]. Brad proceeded to do as he was told, inventing what he later described as an imaginative scenario. Two hours after his alleged confession, when he was told that the police considered it to be a confession, he seemed genuinely astonished and immediately recanted. (Pratkanis & Aronson, 1991, pp. 175–176, emphasis in original)
Our partial reenactment of the Page interrogation and “confession” was recorded simultaneously by three video cameras that yielded a suspect-focus, equal-focus, and detective-focus version of the confession. Eighty-six participants viewed one of the versions of the confession and then completed mostly the same measures as used in the prior studies. One addition of note was the inclusion of a 9-point scale item that asked participants, “if the suspect were convicted, how severe should his sentence be?” (1 = minimum sentence and 9 = maximum sentence). This question was added to see if the influence of camera perspective would extend to judgments beyond voluntariness and guilt and would actually taint other related decisions such as sentence recommendations. All three measures produced the same significant linear trend in means (see Table VI). The suspect-focus version of the confession led to the highest voluntariness, likelihood-of-guilt, and severity-of-sentence ratings, whereas the detectivefocus version led to the lowest ratings across the various measures. Another path analysis was conducted with sentence recommendations included as an additional variable. The resulting path diagram is depicted in Fig. 4 (only the significant paths are shown). Consistent with the first three studies, the impact of camera point of view on guilt assessments was mediated by voluntariness judgments. Interestingly,
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
213
Fig. 4. Path diagram and coefficients (standardized beta weights) for Study 4. Only significant paths are depicted, p < .05. (Data from Lassiter et al., in press, Study 4.)
camera perspective influenced sentence recommendations in a direct manner, but also indirectly via its effect on voluntariness and guilt judgments. At this point it was becoming increasingly clear that the potential for some degree of prejudicial impact associated with videotaped confessions was disquietingly real. Having participants witness simulated events that occurred in an actual police interrogation and exposing them to nearly 30 min of content information did not diminish the biasing effect of camera perspective one iota. In addition, for the first time the point-of-view bias in videotaped confessions was shown to have the potential to also influence decisions regarding the severity of sentence that might be imposed.
G. STUDY 7: IS ACCOUNTABILITY THE KEY? When the first author of this chapter presented some of the above findings to colleagues at another university, a question arose about the accountability of our participants. That is, the questioner felt that the evidence of bias might evaporate if our research participants were made to feel truly accountable for, or had to justify, the judgments they were producing. Research on the effects of accountability (see Lerner & Tetlock, 1999, for a review) on judgments does suggest that, in certain instances, increased accountability can attenuate bias. However, this literature also provides empirical examples of accountability amplifying bias or having no effect at all on people’s judgments. Lerner and Tetlock (1999, p. 263) argued that “predecisional accountability to an unknown audience will improve judgment to the extent that a given bias results from lack of effort, self-critical awareness of one’s judgment processes, or both.” The fact that, in Study 2, we found no reduction in the biasing effect of camera perspective for individuals who are naturally motivated to be effortful and critical thinkers (i.e., those high in need for cognition) led us to believe an accountability manipulation might not be effective in this case. Nonetheless, Study 7 (Lassiter, Munhall, Geers, Weiland, & Handley, 2000) was conducted to find a data-based answer to the questioner’s query.
214
G. DANIEL LASSITER ET AL.
After being told that they would be assessing the voluntariness of a videotaped confession, participants (N = 63) were assigned to either a low-accountability or high-accountability condition. In the high-accountability condition, participants were told the following: “We are also interested in whether the basis of your judgments about the confession are consistent with the way judges believe jurors make decisions. A local judge, [name], has been helping us with this project and has agreed to meet with you to review your judgments about the confession and to determine if the manner in which you arrived at your judgments is correct.” The experimenter then scheduled a time for the participants to meet individually with the judge so that they could “explain your decisions concerning the confession to him.” Participants were told they would be compensated monetarily for their time (up to $50). (Participants really believed the meeting was going to take place as in all cases they spontaneously made a point to write down all the details of the scheduled appointment). In the low-accountability condition, these instructions were omitted. All participants then saw a brief videotape of the above-named judge (an actual retired judge from the community) providing some guidelines concerning the determination of voluntariness. The videotape depicts the judge, dressed in his judicial robe, sitting at the bench in the local court house. Seeing the judge like this further conveyed a sense of reality about having to justify their judgments for those in the high-accountability condition. Participants then viewed either the suspect-focus or equal-focus version of the Page confession used in Study 6. Because the judge’s remarks directed participants’ attention specifically to the question of the confession’s voluntary status, we made this the sole judgment they would have to justify. However, unlike previous studies that relied exclusively on rating scale measures, participants also had to declare their judgments of voluntariness in a dichotomous fashion. We did this to preclude the possibility of them hedging their bets by simply choosing the midpoint on the rating scale items. We asked participants to also provide an indication of their confidence in their dichotomous judgments of voluntariness on a 9-point rating scale similar to that used in Study 1. To assess whether the accountability manipulation was effective in inducing a more thoughtful consideration of the confession evidence, we additionally had participants write down which aspects of the videotaped confession were most important to them and why. Participants could write as much as they wanted. We used participants’ responses to this open-ended question as a gauge of the extent of cognitive elaboration they engaged in while evaluating the confession evidence. The number of lines of text produced by participants was interpreted as an indication of the amount of thought they devoted to the judgment task, with more lines assumed to reflect greater thought. The cognitive-elaboration measure was subjected to a 2 (accountability) × 2 (camera perspective) ANOVA. This analysis yielded a highly significant main
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
215
TABLE VII RESULTS ON THE VOLUNTARINESS MEASURES (CONTINUOUS AND DICHOTOMOUS) AND ON THE COGNITIVE ELABORATION MEASURE (STUDY 7) Low accountability Measure Continuousa Dichotomousb Cognitive elaborationc
High accountability
Suspect-focus
Equal-focus
Suspect-focus
Equal-focus
16.63 .76 5.53
12.50 .43 7.93
17.87 .73 8.27
15.20 .41 10.88
Note. Data from Lassiter, Munhall, Geers, Weiland, and Handley (2000). a Higher scores indicate judgments of greater voluntariness. b The proportion of participants judging the confession voluntary. c Amount of thought about the confession as measured by the number of lines of written text generated by participants.
effect of the accountability manipulation. High-accountability participants wrote down more thoughts (i.e., lines of text) about the confession than did their lowaccountability counterparts (see Table VII for means). Interestingly, there was also a significant effect of camera perspective. Participants who viewed the equalfocus version of the confession engaged in more cognitive elaboration than did those who viewed the suspect-focus version. Finally, the two-way interaction was not significant. As noted above, all participants in the high-accountability condition wrote in their datebooks the details of their scheduled appointment with the judge. In addition, these same participants were genuinely surprised when they were informed that no such meeting would take place. These behaviors, combined with the results of the elaboration measure, led us to conclude that the accountability manipulation was successful. An Accountability Condition × Camera Perspective ANOVA performed on the continuous measure of voluntariness revealed a significant main effect of camera perspective only. As can been seen in Table VII, the suspect-focus version of the confession once again produced higher judgments of voluntariness than did the equal-focus version. An analysis of the dichotomous measure of voluntariness produced an identical pattern of results. That is, the suspect-focus version of the confession produced significantly more voluntary (vs involuntary) judgments than did the equal-focus version in both the low- and high-accountability conditions. Finally, there were no significant effects of camera perspective or the accountability manipulation on participants’ confidence in their voluntariness judgments (overall M = 6.81). The unexpected finding that the equal-focus version of the confession elicited more cognitive elaboration from participants than did the suspect-focus version
216
G. DANIEL LASSITER ET AL.
Fig. 5. Path diagram and coefficients (standardized beta weights) for Study 7. Solid paths are significant, p < .05. (Data from Lassiter, Munhall, Geers, Weiland, & Handley, 2000.)
suggests that the effect of camera focus on voluntariness judgments might have been mediated by the quantity of thought about the confession. To examine this possibility, we again conducted a path analysis. As is shown in Fig. 5, the direct paths from camera perspective to cognitive elaboration and to voluntariness judgments were both significant. The path from cognitive elaboration to voluntariness judgments, however, was not significant. Overall, this analysis indicates that with regard to judgments of voluntariness, the biasing effect of camera perspective is not mediated by the amount of thoughtful consideration given to the confession. The results of Study 7 indicate that an increased sense of accountability (that, in turn, induces greater processing effort) is not enough to minimize the biasing effect of camera perspective. This finding, taken in combination with Study 4’s demonstration that forewarning is also an insufficient corrective, may be consistent with the following point made by Wilson and Brekke (1994, p. 131): “[Sometimes] even when people are aware that information can bias them and are motivated to resist the bias, they either adjust responses too little or too much. The reason, we suggest, is that they are unaware of how much they have been biased and thus do not know how much to alter their responses.” Thus, as Lerner and Tetlock (1999, p. 263) concluded, “bias correction hinges not only on the motivation to correct, but also on the ability to correct one’s mental processes.” It may be the case that the camera perspective bias in videotaped confessions, and salience effects more generally, may be especially hard to undo on account of “a lack of awareness of mental processes, the limitations of mental control, and the difficulty of detecting bias” (Wilson & Brekke, 1994, p. 117).
H. STUDY 8: DOES CAMERA PERSPECTIVE AFFECT HOW BELIEVABLE THE SUSPECT APPEARS? It is possible that the effects reported in the preceding studies may have something to do with how believable/credible the statements made by the suspect are judged to be. More specifically, when the camera is focused on the face of the
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
217
suspect, he or she may be perceived to be more truthful, which in turn leads observers to see the statements as more voluntary and hence the suspect more likely to be guilty. This idea (first suggested to us by Rich Petty) can be derived from the vast literature on deceiving and detecting deceit (Ekman, 1992; Kraut, 1980; Zuckerman, DePaulo, & Rosenthal, 1981), which indicates that people at least perceive another’s eyes (are they shifty? avoidant?) to be an important cue as to whether the person is lying or telling the truth (cf. DePaulo, Stone, & Lassiter, 1985). Because observers may have a harder time determining the gaze of the suspect in the equal-focus videotapes, and certainly so in the detective-focus videotapes, they may come to distrust the confession more which could perhaps explain the previous findings. Study 8 (Lassiter, Munhall, Geers, Handley, & Weiland, 2000) examined this possibility. One hundred thirty-eight participants viewed either the suspect-focus or equal-focus version of the Page confession. A three-item voluntariness index and three-item credibility index were generated from participants’ responses to the dependent measures and entered into a 2 (camera perspective: suspect-focus or equal-focus) × 2 (index: voluntariness or credibility) between–within ANOVA. There was no overall effect of camera perspective, but this was due to the significant Perspective × Index interaction (see Table VIII). Simple effects tests indicated that voluntariness, but not credibility, judgments were significantly affected by the camera perspective. Given this divergence in the way the two judgments were affected by camera perspective, it is not surprising that the voluntariness–credibility correlation turned out to be nonsignificant, r = .12. It appears from these data, then, that the biasing effect of camera perspective on voluntariness, guilt, and severity-of-sentence judgments is not due simply to a focus-induced change in how believable the confessor comes across.
TABLE VIII MEANS FOR THE VOLUNTARINESS AND CREDIBILITY INDEXES (STUDY 8) Camera perspective Measure
Suspect-focus
Equal-focus
Voluntariness index
19.01
17.14
Credibility index
14.18
14.55
Note. Higher numbers indicate that the confession was judged more voluntary and more credible for the voluntariness and credibility indexes, respectively. (Data from Lassiter, Munhall, Geers, Handley, & Weiland, 2000.)
218
G. DANIEL LASSITER ET AL.
VIII. Stage Two: Examining the Camera Perspective Bias in Videotaped Confessions in the Context of Elaborate Trial Simulations As with any research of this kind, there are limitations of the preceding investigations that need to be acknowledged. First, our experiments did not involve actual confession evidence, an actual trial, or actual jurors. Therefore, the extent to which our findings generalize to real situations can be questioned. However, concern about this issue should be diminished to some extent by MacCoun’s (1989, p. 1046) review of a large body of mock juror research in which he concluded that “mock jurors do not appear to reach decisions by a fundamentally different process than actual jurors.” Because Studies 1–8 were part of our Stage One research (cf. Diamond, 1997), we used relatively simple stimulus materials and excluded many other trial components. For example, other than in Study 4, there was no additional evidence for participants to consider other than the confession itself. Obviously, in real trials, fact finders are almost always presented with other evidence in addition to the confession. Although the research presented so far provides strong proof that the camera perspective bias in videotaped confessions is a robust phenomenon, it is not inconceivable that the presence of other kinds of evidence could cause a dilution of this prejudicial effect. Study 4 did not provide an adequate test of this dilution possibility because the witness testimony was presented in a written format and therefore was likely experienced by participants as less vivid and “real” than the confession, which was in a videotaped format (cf. Taylor & Thompson, 1982). Also, for reasons of convenience, in Stage One we used only college students as our mock trial participants. Some investigators (e.g., Feild & Barnett, 1978; Foss, 1976) have questioned the use of students as participants in jury-simulation studies. The responses of students, it is argued, may be quite different from those of jury-eligible adults, in which case the generalizability of the findings of studies using student mock jurors is likely to be severely limited. Recent reviews of the mock juror/jury literature (Bornstein, 1999; MacCoun, 1989), however, indicate that the judgments of student and adult mock jurors are comparable. Despite such reassuring findings, the impact of the present program of research on the criminal justice establishment will no doubt be increased if it is demonstrated that the pointof-view bias in videotaped confessions is manifested not only by students but by older, nonstudent adults as well. Another drawback of the Stage One studies has to do with the fact that participants made their judgments only on continuous rating scales (with the exception of Study 7). This was done to ensure that our measures were as sensitive as possible in detecting any evidence of a biasing effect of camera point of view. However, verdicts in actual courtrooms are made in an either/or manner, and we cannot be certain that the bias observed with rating scales will still obtain with cruder, but
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
219
more ecologically valid, dichotomous measures (cf. Kerr, 1978). (Study 7 did not require participants to render a verdict of guilt or innocence, so this all-important question still needs to be answered.) If the present program of research is going to have an impact on the legal community, it is incumbent upon us to deal with these issues as best we can. As noted by Bornstein (1999, p. 88), “courts have not welcomed psycholegal research findings with open arms, especially when derived from methods that are neither very realistic nor representative of actual legal processes.” [The tide may be turning somewhat. Recently, the U.S. Department of Justice released the first national guide for collecting and preserving eyewitness evidence. Psychological science contributed significantly to the development of this document (see Wells, Malpass, Lindsay, Fisher, Turtle, & Fulero, 2000).] Our Stage Two research, then, comprises three studies that are notable for their, in Bornstein’s (1999, p. 88) words, “harder, more representative methods.” All three studies involved extensive videotaped trial simulations that required from 3 to 5 h of participants’ time. [As noted by MacCoun (1989, p. 1046), “manipulations of many variables (that may affect juror decision making) . . . are not ethically or legally feasible in actual trial settings.” Therefore, mock jurors—formulating judgments regarding a simulated legal trial—are used for most experimental tests of juror decision making and the variables that influence it.] In Studies 9 and 11, all participants were nonstudent, jury-eligible adults recruited from both rural and urban communities in Ohio. In Study 10, both nonstudent and student participants were used so that a systematic comparison of their responses could be made. In all Stage Two studies, dichotomous measures of participants’ judgments were obtained. In addition to addressing these concerns of mundane realism, we continued to explore in Stage Two other possible limits on, and maybe even a possible benefit of, the influence of camera perspective on the judgment process of observers.
A. STUDY 9: IS THERE ANY EFFECT OF MULTIPLE VIEWINGS OF THE CONFESSION ON THE CAMERA PERSPECTIVE BIAS? In real courtroom trials, fact finders most likely have the opportunity to evaluate confession evidence more than once if they feel the need to do so before rendering any decisions. (This was indeed the case in the actual trial that will serve as the basis for our stimulus materials in this study.) Could such repeated observation of a videotaped confession affect the magnitude of the bias associated with camera point of view? The existing body of literature on salience effects does not provide any insights on this point. However, the argument could be made that the opportunity to examine a videotaped confession more than once should work to reduce the influence of camera point of view on observers’ judgments. That is, drawing once
220
G. DANIEL LASSITER ET AL.
again on the work of Kaplan and Miller (1978), additional viewings of a videotaped confession could cause the contents to become more salient or to be given more weight in observers’ integration of the information (cf. Anderson, 1974; Kaplan, 1975), which in turn could increase the effect of the contents on judgments while concurrently inhibiting the manifestation of the bias. The possibility also exists that repeated observation of a videotaped confession could further intensify the biasing effect of camera point of view. Such a result would be consistent with some findings from the literature on mere exposure effects (Harrison, 1977; Zajonc, 1968). Studies by Cacioppo and Petty (1989), Grush (1976), Perlman and Oskamp (1971), and Swap (1977) indicate that initial evaluations of a stimulus can become more polarized with increasing exposure. Thus the bias (found with suspect-focus videotapes) toward judging the defendant as having confessed freely and to be guilty, rather than being minimized, may become even more pronounced with additional observation. Study 9, then, was conducted to empirically determine which, if either, of these two possibilities might occur. The case of Peter Reilly was used to generate the stimulus materials. Gudjonsson (1992, p. 252) provides the following synopsis of the particulars of this case: In 1973, Reilly was an easy-going and well-liked 18-year-old-youth. He lived in Canaan, Connecticut, with his 51-year-old mother. At 8 p.m. on the 28th of September he went to a Methodist Church for a Youth Centre meeting. He returned home around 9:50 p.m. to discover his mother’s mutilated body. She had been brutally murdered minutes before Reilly arrived home. He immediately called for an ambulance and was clearly in a distressed state. Within hours he became the prime suspect for the murder and after intensive interrogation he made a self-incriminating confession, which resulted in his arrest and conviction for manslaughter.
Reilly’s confession is considered by many to be a prime example of a coercedinternalized confession (cf. Kassin & Wrightsman, 1985)—that is, Reilly eventually came to accept as true, at least for a short time, the assertions of his interrogators that he was indeed a perpetrator of matricide! Key aspects of Reilly’s interrogation and confession are summarized in this unnerving account by Kenrick, Neuberg, and Cialdini (1999, p. 152): At the scene and even when taken in for questioning, Reilly waived his right to an attorney, thinking that if he told the truth, he would be believed and released in short order. This was a serious miscalculation. [Recent research by Kassin and Fong (1999), however, suggests that most of us share Reilly’s belief that our innocence will surely be discerned by others.] Over a period of 16 hours, he was interrogated by a rotating team of four officers, including a polygraph operator who confidently informed Reilly that, according to the lie detector, he had killed his mother. The chief interrogator told Reilly, falsely, that additional evidence proving his guilt had been obtained [maximization]. He also suggested to the boy how he
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
221
could have done the crime without remembering any such thing: Reilly had become furious with his mother [employing minimization, the interrogators further suggested that Reilly’s reaction was justifiable because of his mother’s constant antagonisms], had erupted into a murderous fit during which he slaughtered her, and now had repressed the horrible memory. It was their job, Reilly’s and the interrogator’s, to “dig, dig, dig” at the boy’s subconscious until the memory was recovered. Dig, dig, dig they did, exploring every way to bring that memory to the surface, until Reilly did begin to recall—dimly at first but then more vividly—slashing his mother’s throat and stomping on her body. Analyzing, reanalyzing, and reviewing these images convinced him that they betrayed his guilt. Along with his interrogators, who pressed him relentlessly to break through his “mental block,” Reilly pieced together from the scenes in his head an account of his actions that fit the details of the murder. Finally, a little more than 24 hours after the grisly crime, though still uncertain of many specifics, Peter Reilly formally confessed in a signed, written statement. This statement conformed closely to the explanation that had been proposed by his interrogators and that he had come to accept as accurate—even though he believed at the outset of his questioning and even though, as later events demonstrated, none of it was true. [Two years after his conviction it was determined that the prosecution had suppressed exonerating evidence, leading to a repeal of Reilly’s conviction and the dismissal of all charges.]
Detailed accounts of the case provided by Barthel (1976) and Connery (1977) were used to recreate portions of the actual interrogation (of which there is an audiotaped record). As was done in Study 1, this partial recreation of the interrogation and confession of Peter Reilly was videotaped simultaneously by three cameras: one taking a suspect-focus position, another a detective-focus position, and the last an equal-focus position. A reenactment of key events occurring in Reilly’s actual trial was also staged based on the Barthel and Connery accounts. The trial simulation was elaborate and was professionally videotaped in an actual courtroom with the assistance of the telecommunications department at Ohio University. A local retired judge portrayed the role of the presiding judge in the trial and two practicing attorneys assumed the roles of prosecutor and defense counsel. Individuals who were mostly recruited from local theater groups enacted the roles of the other trial principals. All actors received some monetary remuneration for their participation. The trial reenactment was video recorded from the vantage point of the jury box. The camera remained stationary throughout the recording. Some zooming and panning of the camera occurred; for example, during witness testimony the camera would at points focus more closely on the witnesses’ faces. The total cost of making the videotaped trial simulation was approximately $10,000. Included in the videotaped simulation were the testimony of two prosecution and three defense witnesses (one of which was “Reilly” himself), Reilly’s confession, the introduction of other items of evidence, opening and closing arguments of the prosecution and defense, and the judge’s rulings on points of law as well as his explication of the requirements of proof to the jurors. The videotaped trial lasted approximately 2.5 h, with the confession accounting for just over 40 min of that time.
222
G. DANIEL LASSITER ET AL.
The 52 participants were residents of the town of Lancaster, Ohio that is located in the southeastern portion of the state approximately 40 miles from Columbus, the state capital. They were recruited via an ad placed in the local newspaper that offered $15 dollars to volunteers wishing to participate in a federally funded mock jury research project. (In subsequent Stage Two studies, the amount of money paid to nonstudent participants ranged from $15 up to $60, depending on the particular study and its location.) Two-thirds of these volunteers were female and their mean age was 51 years. Based on their responses to a preliminary “Juror Profile” questionnaire, 57% had no more than a high school education, 40% were Democrats, 45% were Republicans, and 13% were Independents. The group was 90% White and 62% indicated they were currently married. Thirty percent reported that they had served as actual jurors on at least one occasion. Participants viewed the trial with one of the three versions of the confession. At the conclusion of the trial, but prior to rendering any decisions, half the participants were shown the confession a second time (because, it was explained, that was what happened in the actual trial). Finally, all participants individually provided their verdicts (guilty or not guilty) and their assessments of the voluntary status of the confession (voluntary or involuntary). Participants were asked to make these same judgments on rating scales as well. The dichotomous measures yielded no significant effects. Scale ratings were thus examined in subsequent analyses. A two-item voluntariness index was created and entered into a 3 (camera perspective) × 2 (number of confession viewings: one vs two) ANOVA. No significant effects were associated with the number of viewings. However, the same linear pattern of voluntariness ratings as a function of camera perspective we found in previous studies was once again significant (see Table IX). A second two-way ANOVA performed on a continuous measure of participants’ assessments of Reilly’s likelihood of guilt also yielded a reliable linear effect of
TABLE IX MEANS FOR DEPENDENT MEASURES (STUDY 9) Camera point of view Measure
Suspect-focus
Equal-focus
Detective-focus
Voluntariness index
12.66
11.60
7.46
Likelihood of guilt
5.89
5.11
3.09
Note. Higher scores on the voluntariness index and guilt measure indicate judgments of greater voluntariness and more probable guilt, respectively. (Data from Lassiter, Beers, Geers, & Munhall, 2000.)
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
223
Fig. 6. Path Diagram and coefficients (standardized beta weights) for Study 9. Solid paths are significant, p < .05. (Data from Lassiter, Beers, Geers, & Munhall, 2000.)
camera perspective. The more of Reilly (or the less of the detective) they could see, the more likely they thought he was guilty of the killing of his mother. Finally, a path analysis conducted on the data revealed, as before, that the effect of camera perspective on likelihood of guilt assessments was mediated, at least in part, by voluntariness judgments (see Fig. 6). Even in the context of a realistic, videotaped simulation of an actual trial that included the direct testimony and cross-examination of several witnesses, the presentation of physical evidence, prosecution and defense arguments, judicial rulings on points of law, and most of the other trappings associated with such legal proceedings, camera perspective still biased mock jurors’ judgments. However, this effect did not come out on the critical dichotomous measures of voluntariness and guilt. We had intended to run more participants in this study, but it turned out that many of our participants were not sanguine about sitting through a second viewing of the 40-min Reilly confession. Therefore, we made the decision to cut the experiment short. The resulting smaller number of participants than anticipated may be a factor that weakened our chances of finding clear results on the dichotomous measures. It also appears from this study that viewing a videotaped confession for a second time does not diminish the biasing effect of camera perspective. However, given the difficulties described above, we view this finding with caution.
B. STUDY 10: DOES JUDICIAL INSTRUCTION CURB THE BIASING EFFECT OF CAMERA PERSPECTIVE? One common courtroom procedure that could possibly prevent the occurrence of the camera perspective bias is the judge’s instruction to the jury. Generally speaking, evidence supporting the effectiveness of judicial instruction designed specifically to counteract certain juror biases is sparse (e.g., Mitchell & Byrne, 1973; Wolf & Montgomery, 1977). However, Kassin and Wrightsman (1981) reported on some data that indicate that judicial instruction may hold some potential as a corrective influence on jurors’ biased evaluations of confession evidence.
224
G. DANIEL LASSITER ET AL.
As described in Section II, these researchers had previously found (Kassin & Wrightsman, 1980) that mock jurors exhibited a positive coercion bias—that is, in rendering their verdicts, mock jurors failed to discount fully a confession that was induced by a promise of leniency. In their 1981 experiments, Kassin and Wrightsman attempted to reduce this bias by presenting mock jurors with various forms of judicial instruction: One form directed jurors to ignore a coerced confession, a second form additionally defined the legal concept of coercion and emphasized the unreliability of such confessions, a third form defined coercion and emphasized the unconstitutionality and unfairness of such confessions, and a fourth form defined coercion and emphasized both the unreliability and unfairness of such confessions. Results showed that the first three forms of instruction had no effect either on subjects’ judgments of voluntariness or on their verdicts. The instructions that emphasized both the unreliability and unfairness of coerced confessions, however, did successfully lower judgments of voluntariness for illegally coerced confessions. Still, the effectiveness of this form of instruction was not complete as verdicts remained unaltered. The possibility exists, then, that some form of judicial instruction may help curb the camera perspective bias in videotaped confessions. Drawing on Kassin and Wrightsman’s (1981) earlier work, Study 10 tested the effectiveness of judicial instruction as a means of reducing the influence of camera point of view on mock jurors’ voluntariness and guilt judgments (Lassiter, Geers, Munhall, Handley, & Weiland, 2000, Study 1). More specifically, two forms of judicial instruction were examined. One form—similar to the version used by Kassin and Wrightsman (1981)—emphasized the need for mock jurors to be cognizant of both issues of reliability and fairness in evaluating confession evidence. This form of judicial instruction was included because, as noted above, it has been shown to reduce, to a certain degree, the biased evaluation of some kinds of confession evidence. The second form of judicial instruction was the same as the first; however, it further emphasized to mock jurors that they should not allow the perspective from which the confession was videotaped to influence in any way their evaluation of the confession. This form of judicial instruction was included because it more specifically directs mock jurors’ attention to the source of the bias and thus provides a strong (second) test of their ability to override the bias when alerted to its existence. Previous research (Feldman, 1978, cited in Horowitz & Willging, 1984; Kassin & Wrightsman, 1979) also suggests that the timing of judicial instruction may determine its effectiveness. Kassin and Wrightsman (1979), for example, found that a judge’s requirement-of-proof instruction had more impact on mock jurors’ verdicts when made prior to the presentation of evidence than when made after the presentation of evidence. Therefore, another purpose of Study 10 was to examine whether the timing of judicial instruction (before vs after the presentation
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
225
of the confession evidence) moderates to any extent its possible effectiveness in minimizing the effect of camera point of view on judgments. Participating in small groups, 73 jury-eligible adults (mean age = 41 years) from relatively small (Athens, population approximately 21,000), medium (Lima, population approximately 69,000), and large (Cincinnati, population approximately 1,200,000) localities in Ohio and 132 undergraduates from Ohio University were asked to assume the role of jurors in a new trial simulation. In the adult sample, 53% were female, 71% were White (85% if missing data are excluded), 23% had obtained a college degree, 37% were married, and approximately 65% had one or more children. The modal reported income range of this group was $15,000 to $30,000; however, 15% chose not to respond to this item. Ten adults indicated an income exceeding $45,000. Finally, 30% considered themselves political independents and approximately half identified Christianity as their religious affiliation. The new trial simulation was based on the case of Bradley Page described in Study 6. The Page trial simulation was developed, staged, and recorded in a manner similar to that used to produce the Reilly simulation (also at a comparable cost). The Page simulation included an additional prosecution witness and was longer than the Reilly simulation by approximately 45 min. The experiment comprised a 2 (camera perspective) × 5 (judicial instruction) design. Approximately half of the participants viewed the trial simulation with the suspect-focus version of the confession, and the rest viewed the simulation with the equal-focus version of the confession. Orthogonal to the camera perspective manipulation, five different judicial instruction groups were run. Participants heard the judge either state that issues of reliability and fairness should be foremost in their minds when drawing conclusions about the confession (taken from approved instruction manuals used by the judiciary) or additionally warn them specifically not to let the camera perspective influence their judgments regarding the confession. Participants received one of these two forms of judicial instruction either just before they viewed the confession or at the end of the trial when the judge gave his charge to the jurors to consider all the evidence that had been presented (the usual timing of such instruction). The fifth and final judicial instruction group was a no-instruction control. After the judge gave his charge to the mock jurors to consider all the evidence that had been presented, they individually completed the dependent measures. Before discussing the results we want to point out that participants in this study showed clear signs of being very engaged in the proceedings. For example, even after 4 h of participation, many of them chose to stay after the simulation was completed to ask thoughtful questions, gather more information about the actual Page case, and discuss their concerns about bias creeping into real jurors’ decisions. We take this behavior as an indication that participants were highly involved and treated the trial simulation very seriously.
226
G. DANIEL LASSITER ET AL. TABLE X PROPORTION OF GUILTY VERDICTS, MEAN CONFIDENCE IN VERDICT, AND PROPORTION ASSESSING CONFESSION WAS VOLUNTARY AS A FUNCTION OF CAMERA PERSPECTIVE AND PARTICIPANT STATUS (STUDY 10) Camera perspective Measure
Suspect-focus
Equal-focus
Nonstudents Guilty verdicts Mean confidencea Voluntariness assessments
.28 7.80 .33
Guilty verdicts Mean confidencea Voluntariness assessments
.33 7.52 .30
.12 7.88 .12 Students .17 7.55 .18
a Rated on a 9-point scale, with higher numbers denoting greater confidence. (Data from Lassiter, Geers, Munhall, Handley, & Weiland, 2000, Study 1.)
Preliminary analyses revealed no significant effects of the judicial instruction manipulation on any of the dependent measures; therefore the data were collapsed across this variable. As can be seen in Table X, the proportion of guilty verdicts rendered in the suspect-focus condition was twice as great as that observed in the equal-focus condition and this difference was significant. Participants were highly confident of their verdicts, regardless of camera perspective. The pattern of results on judgments of voluntariness was very similar to that obtained for verdicts. The proportion of participants judging the confession to be voluntary in the suspect-focus condition reliably exceeded that found in the equal-focus condition. The correlation between voluntariness assessments and verdicts was substantial, = .80. Finally, it is important to note that both student and nonstudent participants evinced very similar patterns of responses (i.e., no significant effects related to this factor). This realistic, fact-based trial simulation demonstrates that a suspect-focus camera perspective can cause triers of fact to judge a videotaped confession to be more voluntary and, more important, can increase their tendency to convict a defendant on the basis of such evidence. Furthermore, this effect is not easily eliminated. Judicial instruction emphasizing the need to be cognizant of reliability and fairness concerns in evaluating the confession and, in some cases, directly alerting mock jurors to the potentially prejudicial effect of camera perspective did not mitigate the bias. This was true whether the judicial instruction preceded or followed the presentation of the confession.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
227
C. STUDY 11: CAN VIDEOTAPED CONFESSIONS IMPROVE OBSERVERS’ ABILITY TO DETECT COERCIVE INFLUENCES? As noted in Section III, one reason supporters of the videotaping practice recommend its widespread acceptance is because it provides an objective record of the interrogation process that all parties can examine (cf. Gudjonnson, 1992; Leo & Ofshe, 1998). Urging more states to require such electronic recordings, Leo and Ofshe (1998, p. 494) argued, “The existence of an exact record of the interrogation is crucial for determining the voluntariness and reliability of any confession statement, especially if the confession is internally inconsistent, is contradicted by some case facts, or was elicited by coercive methods or from highly suggestible individuals.” This statement implies that more accurate assessments of the voluntariness and reliability of confessions can be obtained via the videotape method. Certainly, if interrogators use obvious, assaultive, coercion, any reasonable observer will recognize the illegitimacy of the confession. However, as discussed above (Section II), such third-degree intimidation has been replaced by nonassaultive psychological manipulation that is not always recognized as coercive but, as research has shown, can nonetheless lead to false admissions of guilt. Our previous 10 studies suggest that videotaping confessions—rather than improving accuracy—might actually lead to less accurate assessments of voluntariness and reliability, at least when the camera’s eye is pointed directly at the suspect—currently the norm. Remember, however, that in Studies 2 and 3, we found that videotaped interrogations with the camera positioned so that both the confessor and interrogator were equally visible produced judgments that were comparable to those based on either reading a transcript or listening to an audiotape of the interrogations. Thus, if implemented judiciously—that is, if an equal-focus camera perspective is taken—the videotape method probably will yield voluntariness and reliability assessments that are no less “accurate” than those based on the more traditional presentation formats. But can we do better? Maybe. As we noted in our first published article on this topic, perhaps the best way to videotape custodial interrogations is to position the camera so that it records the visual perspective of the accused. “This would allow those charged with evaluating the status of a confession the maximum opportunity to spot coercive influences should they be at work” (Lassiter & Irvine, 1986, p. 275). Although most criminal justice practitioners, and even the average person on the street, might condemn this approach as cockeyed, its logic is borne out in the empirical literature. Storms (1973) demonstrated that the tendency to overattribute another person’s behavior to internal, dispositional causes (i.e., the fundamental attribution error) could be corrected to some degree by having observers view a videotape that depicted exactly what the other person saw. Having the opportunity to literally “put yourself in another’s place” enabled observers to better
228
G. DANIEL LASSITER ET AL.
appreciate the external forces experienced by that person because those forces were now more “exposed” and thus more likely to be detected by observers. Consistent with this result, a number of other studies found, using a variety of methods, that when situational factors are made especially salient or obvious, those factors are much more likely to be taken into account in the shaping of observers’ causal impressions (e.g., Arkin & Duval, 1975; Duval & Wicklund, 1972; Lassiter, 1986; Regan & Totten, 1975; M. Ross, 1975; Wilson & Lassiter, 1982). In Study 11, then, we attempted to determine if a videotaped confession recorded from the perspective of the accused (i.e., focused on the interrogator) can facilitate observers’ capacity to detect coercive influences, internal inconsistencies, or contradictions with other known case facts, thus leading them to conclude the confession is unreliable (Lassiter, Geers, Munhall, Handley, & Weiland, 2000, Study 2). We again had participants view the videotaped simulation of the trial of Bradley Page. Leo and Ofshe (1998) categorized Page’s self-incriminating statements obtained during his interrogation as a highly probable false confession. These authors arrived at this conclusion primarily because there was no other evidence to corroborate Page’s account of what supposedly happened. Leo and Ofshe (1998, p. 456) give the following as examples: Page . . . stated that he made love to the dead body on a blanket taken from his vehicle; in fact, the blanket contained no evidence of sexual activity, no blood stains from Lee’s massive head wounds, no signs of having been washed, and the hairs found on the blanket were not Lee’s. Page guessed that he used a spare hubcap that was in his vehicle in an attempt to bury Lee, but the fibers and soil from the hubcap did not match either the fibers of Lee’s clothing or the soil where her body was found. Page also stated that he dragged Lee’s body more than 100 yards before burying it. Had this happened there would have been a trail of blood that surely would have been found by the various search and rescue and dog tracking teams that, beginning the day after her disappearance, spent hundreds of hours combing the area where Lee’s body was eventually found.
Leo and Ofshe’s conclusion is further buttressed by Pratkanis and Aronson’s (1991) analysis of the social influence factors that likely induced Page to falsely incriminate himself. We chose the Page stimulus materials because, as mentioned earlier, we had a transcript of Page’s actual interrogation, at least the portions that were audiotaped (less than 4 h), so these materials allowed us to come as close as possible to retaining the critical aspects of his confession and the circumstances surrounding its elicitation. The numerous discrepancies between Page’s statement and other facts that came out in the trial were largely reproduced in our simulation. Page’s own testimony about how the interrogators implored him to help them solve the
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
229
case by imagining how the crime might have occurred, if he were to have done it, was also presented and was based on trial excerpts that we obtained from various legal, academic, and journalistic research materials. Participants viewing the trial simulation, then, found themselves in a quandary: How to explain Page’s, as Leo and Ofshe (1998, p. 455) described it, “vague, confused, and speculative” admission of guilt when nothing else in the trial conformed very well to the specifics of his narrative. Mock jurors essentially had two choices. They could decide that the prosecution’s argument that the interrogators skillfully extracted the truth from Page was in fact what happened despite the evidence to the contrary. Or they could accept the defense’s position that during a protracted interrogation, Page was persistently manipulated until the detectives were able to finally convince him to tell a “what if” story about how the killing might have unfolded. We thought that the version of the videotaped confession participants viewed might tip the balance in their decision to convict or acquit Page. More specifically, when the camera focused equally on Page and his interrogator, we anticipated— due to the prevailing proclivity of observers to view the causes of a person’s actions as emanating from that person—that participants would be more inclined to decide that Page’s confession was voluntary and thus find Page guilty as charged. However, when the camera focused on the interrogator—the source of the pressure that was impinging on Page—we anticipated, consistent with the literature described above, that participants would pay greater heed to the influence of the interrogator and consequently be more inclined to decide the confession was involuntary and thus find Page not guilty. Because the overall rate of conviction was low in Study 10, we made some adjustments to try to increase the number of guilty verdicts. From discussions with participants in the preceding study we learned that several of them felt the prosecution should have presented DNA evidence to support their case. Such evidence was not widely obtained or used at the time of Page’s actual trial. The simulation did not specify a date, so participants were assuming it was more current than it actually was. In this study, then, we informed them that the murder of Bibi Lee happened in the mid-1980s, which is the case. Some participants also wondered whether the confession should have been admitted into evidence at all. We clarified this issue by explicitly noting that the judge determined in a pretrial hearing that the confession could be admitted into evidence, which is also the case. Finally, some participants were uncertain as to the legal distinction between voluntary and involuntary manslaughter and this was noted as a reason for their hesitancy to convict. We thus added a more detailed definition of these legal concepts that we extracted from various legal resources. Piloting showed these adjustments to be effective. One other important difference between this study and the last is that mock jurors were directed to deliberate before rendering their verdicts. Up to 45 min
230
G. DANIEL LASSITER ET AL. TABLE XI PROPORTION OF GUILTY VERDICTS AND PROPORTION ASSESSING CONFESSION WAS VOLUNTARY AS A FUNCTION OF CAMERA PERSPECTIVE (STUDY 11) Camera perspective Measure Guilty verdicts Voluntariness assessments
Equal-focus
Detective-focus
.40 .40
.05 .09
Note. Data from Lassiter, Geers, Munhall, Handley, & Weiland (2000, Study 2).
was allowed for deliberation. On average, jurors required 25 min to conclude their deliberations. Volunteers were 42 jury-eligible adults recruited from the Youngstown, Ohio, area who received $60 in return for up to approximately 5 h of participation. This group was the youngest of the three adult samples we examined, with a mean age of 37.5 years. The racial mix was also more diverse than that of preceding samples. Sixty-three percent listed White, 27% listed Black, and 10% listed Other. Finally, there were many more declared democrats in this sample (69%). Participants viewed the trial simulation with either the equal-focus or detectivefocus version of the confession included. Unlike Study 10, judicial instruction was not manipulated. Instead mock jurors received only the reliability/fairness instruction just prior to their deliberations. After their deliberations, participants rendered their verdicts individually. As was the case in Study 10, all indications pointed to the participants being very involved in the proceedings. The direction of the differences obtained was as predicted, but the magnitude of those differences was startling. As can be seen in Table XI, the conviction rate dropped 35 percentage points by simply changing the camera angle from an equal focus to a detective focus. This difference was highly significant. This effect is quite remarkable, but even more so because it came about after mock jurors had discussed the evidence intently, in some cases for the full 45 min that had been allotted. [Participants were highly confident in their verdicts (M = 7.68), with no significant differences due to camera perspective.] The pattern of results for judgments of voluntariness was similar and also statistically significant. The correlation between the two judgments was again very large, = .80. We discuss the implications of these results and all of the other Stage One and Two findings in more detail, but first we turn our attention to the issue of the basic mechanisms or processes that underlie the potent point-of-view/salience effects that we have observed in the present studies and that have been found more generally in the attribution literature.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
231
IX. Stage Three: Exploring Possible Mediators of the Biasing Effect of Camera Perspective Stage Three comprised four experiments whose main purpose was to elucidate the judgmental impact of one’s mere visual perspective. Like Stage One, these studies used rather simple stimulus materials and participants were drawn from college populations. Because the focus of this stage of the research was on pinning down the basic mechanisms underlying the point-of-view/salience bias, the stimuli we used in three of the experiments were similar to those in the original studies conducted by Taylor and Fiske (1975) and their colleagues (Taylor, Crocker, Fiske, Sprinzen, & Winkler, 1979)—that is, 5-min “getting-acquainted” conversations between two college students. In the remaining experiment, however, we demonstrated that the same basic effects found with these stimuli do generalize to the kinds of videotaped confession materials we used in Stage One. Although salience effects in causal attribution and related judgments have proven to be robust and generalizable, the process or processes underlying them have remained elusive. Early attempts to identify a mediator of salience effects focused on memory processes (e.g., Fiske, Kenny, & Taylor, 1982; Smith & Miller, 1979). Generally, the argument is that salient information tends to be more memorable than nonsalient information, and this difference in memory is responsible for the greater causality ascribed to salient information. The evidence consistent with a solely memory-mediated explanation of salience effects, however, is quite limited (cf. McArthur, 1980). Newtson and his colleagues (Newtson, Rindner, Miller, & LaCross, 1978) and McArthur (1980) have suggested that salience effects may have more to do with how people initially pick up or register information from an observed interaction than with how they subsequently remember that information. That is, the point of view from which individuals observe an interaction is argued to influence the initial registration or perceptual organization of information from the ongoing interaction, which in turn directly influences causal attributions and related judgments. Although an intriguing hypothesis, the position that salience effects are perceptually mediated has not, to date, been empirically evaluated. Studies 12–15, then, were designed to determine if the manner in which people subjectively register or segment information from an observed interaction does indeed play a mediating role in the production of the point-of-view/salience bias. All four studies employed the behavior segmentation technique developed by Newtson (1973, 1976, 1980; Newtson, Hairfield, Bloomingdale, & Cutino, 1987). Briefly, this technique involves having participants view an ongoing behavior sequence and instructing them to identify the most informative segments in the sequence by pressing a button (that activates a recording device) each time, in their
232
G. DANIEL LASSITER ET AL.
judgment, a meaningful action occurs. Participants’ button-pressing responses provide an indication of the number of actions discriminated or registered (i.e., segmentation rate), with more actions, or a higher segmentation rate, associated with the extraction of a greater amount of information from the observed behavior. Considerable data are available attesting to the reliability, validity, and nonreactivity of this technique (Newtson & Engquist, 1976; Newtson, Engquist, & Bois, 1976; Newtson, Engquist, & Bois, 1977; Newtson et al., 1978). Moreover, the segmentation procedure has been used successfully to investigate the possible role of perceptual processing in a variety of social judgment effects (Cohen & Ebbesen, 1979; Geers & Lassiter, 1999; Graziano, Moore, & Collins, 1988; Hanson & Hirst, 1989; Hogue & Atkinson, 1989; Lassiter, 1988; Lassiter, Briggs, & Bowman, 1991; Lassiter, Geers, Apple, & Beers, 2000; Lassiter, Koenig, & Apple, 1996; Lassiter & Slaw, 1991; Lassiter & Stone, 1984; Lassiter, Stone, & Rogers, 1988; Markus, Smith, & Moreland, 1985; Massad, Hubbard, & Newtson, 1979; Newtson, 1973; Newtson & Rindner, 1979; Wilder, 1978a, 1978b).
A. STUDY 12: AN INITIAL TEST OF THE PERCEPTUAL SEGMENTATION HYPOTHESIS In his seminal article on behavior segmentation processes, Newtson (1973) reported that a higher segmentation rate produced more dispositional attributions for an actor’s behavior. Newtson noted that this result was consistent with Kelley’s (1967) suggestion that observers in a higher information state regarding an actor are more likely to interpret his or her behavior in dispositional terms. This link between segmentation rate and causal attribution appears robust, as it has been replicated on several occasions (Deaux & Major, 1977; Lassiter et al., 1988; Wilder, 1978a, 1978b). Newtson and his colleagues (Newtson et al., 1978) later discovered that the availability of feature changes in an ongoing behavior sequence influenced the number of meaningful actions observers could differentiate in the sequence. That is, when certain feature changes in the sequence were not visible to observers, they were unable to use those hidden changes as the basis for organizing the behavior into meaningful actions, even though they could readily infer that such changes were occurring. Combining their results with the finding that a higher segmentation rate is associated with more dispositional attributions for an observed other’s behavior, Newtson et al. (1978) suggested that point-of-view effects in causal attribution might be, in part, a consequence of differences in segmentation, since not all the same feature changes would be available to observers who are viewing an event from different perspectives. According to this argument, then, observers’ divergent perspectives would cause them to segment an interaction differently, which in turn would lead them to arrive at disparate causal attributions. Study 12 provided an initial test of
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
233
this possibility (Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 1). The stimulus conversation between two male undergraduates was recorded simultaneously by three video cameras each taking a different perspective. These differing camera perspectives (Person A focus, Person B focus, and equal focus) served as our manipulation of perceptual salience (cf. Taylor et al., 1979, Experiment 3) and were essentially identical to the camera angles employed in the studies involving the confession stimuli. In an attempt to make the actual contribution of both interactants objectively equal, each of them initiated the same number of questions during the conversation. Recruited for an experiment on “social observation,” 104 volunteers (participating individually) were asked to view one version of the videotaped conversation. Participants were given two buttons (one labeled “A” and one labeled “B”) and were instructed to press the “A” button whenever Person A did anything meaningful in the videotape and to press the “B” button whenever Person B did anything meaningful. Cards with the letters “A” and “B” on top of the video monitor reminded participants of who was who. After viewing and segmenting the conversation, all participants completed a short questionnaire. Included in this questionnaire were three items taken from previous research (e.g., Taylor & Fiske, 1975) designed to assess participants’ perceptions of causality. Participants were asked to indicate on separate 9-point scales how much a particular interactant “set the tone of the conversation,” “determine[d] the kind of information exchanged,” and “cause[d] his partner to behave as he did.” Participants responded to each of these items twice; once for Person A and once for Person B, with the order counterbalanced. The results (depicted in Fig. 7) were consistent with the perceptual-mediation hypothesis outlined above. Participants identified relatively more meaningful actions in the behavior of the interactant who, by virtue of the camera perspective, was most visually salient to them. This interaction of camera perspective and target person (A vs B) was highly significant. Following Taylor and Fiske (1975) and others, the three attribution items were summed to create a single causality index. A second 3 (camera perspective) × 2 (target person) ANOVA performed on this measure revealed the typical salience-effect pattern (i.e., a significant two-way interaction). Participants ascribed relatively greater causality to the interactant who was most clearly visible to them, and this visibility was determined wholly by the camera perspective. As was done in several of the previous studies, we conducted a path analysis to determine if segmentation rate was a viable mediator of participants’ causality judgments. The results of this analysis are presented in Fig. 8. Again, consistent with the perceptual-mediation hypothesis, the paths from camera perspective to segmentation rate and from segmentation rate to causality index were both significant. The direct path from camera perspective to causality index (after partialing out the effect of segmentation rate) was not significant.
234
G. DANIEL LASSITER ET AL.
Fig. 7. Difference scores (Person A − Person B in SD units) as a function of camera focus for the segmentation and causality measures (Study 12). (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 1.)
B. STUDY 13: DOES THE SEGMENTATION TASK ALTER OBSERVERS’ CAUSALITY JUDGMENTS? Study 12 provided initial evidence for the perceptual mediation of point-ofview/salience effects. One possible problem with this study, however, is that engaging in the segmentation task itself may have altered or influenced participants’ causality judgments in some manner. Although, as noted above, the segmentation procedure has been shown to be nonreactive, we wanted to be certain that the pattern of causality judgments obtained was not simply an artifact of participants’ having had explicitly segmented the observed interaction. To address this concern, in Study 13 (Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000,
Fig. 8. Path diagram and coefficients (standardized beta weights) for Study 12. Solid paths are significant, p < .05. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 1.)
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
235
Study 2) only half of the participants engaged in the segmentation task. All participants, however, provided causality judgments. In this experiment we also used the videotape of the burglary confession from Stage One as our stimulus. This was done to ensure that the effects found with the conversation stimulus are generalizable to the more unusual kind of interaction that occurs in police interrogations. Eighty-three participants were assigned to segment or only view either the suspect-focus or detective-focus version of the videotaped confession. Those having to segment were once again given two buttons to press, only this time one was to record the meaningful actions of the suspect and the other was to record the meaningful actions of the detective. Following the presentation of the confession evidence, participants provided their judgments of the confession’s voluntariness on rating scales like those used in Stage One. Similar to the results of Study 12, participants perceived the suspect’s behavior as relatively more meaningful when the camera was focused on him. The pattern was just the opposite when the camera was instead focused on the detective. In this condition, the behavior of the detective was perceived to be relatively more meaningful. The interaction testing this overall pattern was significant (see Fig. 9 for means on both measures). Participants’ voluntariness judgments were once again significantly altered by the camera’s perspective. Those viewing the suspect-focus version of the confession judged it to be more voluntary than did
Fig. 9. Dependent measures in SD units (Study 13). For the segmentation measure only, a difference score (Suspect − Detective) is displayed. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 2.)
236
G. DANIEL LASSITER ET AL.
Fig. 10. Path diagram and coefficients (standardized beta weights) for Study 13. Solid paths are significant, p < .05. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 2.)
those viewing the detective-focus version. Importantly, this difference was obtained whether participants were segmenting the interrogation or not (i.e., no significant two-way interaction). Finally, a path analysis indicated for the second time that the effect of camera perspective on judgments was mediated, at least in part, by the manner in which the observed information was initially segmented. That is, the paths from camera perspective to segmentation rate and from segmentation rate to voluntariness index were both significant, whereas the direct (nonmediated) path from camera perspective to voluntariness index was not (see Fig. 10).
C. STUDY 14: IS SEGMENTATION RATE CONFOUNDED WITH THE AMOUNT OF INFORMATION RECALLED? Studies 12 and 13 indicated that variation in perceptual segmentation is a possible mediator of the point-of-view/salience bias. However, these studies did not permit a clear demonstration of a direct effect of segmentation rate on causality/voluntariness judgments. That is, several previous studies (Geers & Lassiter, 1999; Hanson & Hirst, 1989; Lassiter, 1988; Lassiter & Slaw, 1991; Lassiter et al., 1988) have shown that segmentation rate influences the later recall of information from an observed event, with a higher rate of segmentation leading to greater subsequent recall. Thus consistent with a memory-mediation account, it could be that the greater recall which typically results from a higher rate of segmentation is ultimately responsible for the bias in judgments (see Fig. 11). The goal of Study 14, then, was to establish whether segmentation rate still contributes significant variance to the biasing effect of camera perspective once the effect of recall is partialed out (see Fig. 12).
Fig. 11. Memory-mediated model of point-of-view/salience effects in causal attribution. (Adapted from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000.)
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
237
Fig. 12. Perception-mediated model of point-of-view/salience effects in causal attribution. (Adapted from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000.)
The experiment (Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 3) was identical to Study 12 except for the following changes: The equal-focus version of the conversation was not used, half of the participants (total N = 66) did not perform the segmentation task, and a measure of the number of meaningful actions recalled was obtained after participants completed the causality items. Results on all three measures are displayed in Fig. 13. First, it is important to note that the act of segmenting the observed interaction did not significantly alter responses on any of the dependent measures. However, consistent with the two preceding studies, significant two-way interactions (Camera Focus × Target Person) were found for both the segmentation and causality measures, indicating that more meaningful information was extracted from, and greater causality attributed to, the most visually salient interactant. This same pattern was also evident on the recall measure. That is, participants remembered significantly more about the interactant who was most directly in the camera’s focus. This latter
Fig. 13. Difference scores (Person A − Person B in SD units) for Study 14. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 3.)
238
G. DANIEL LASSITER ET AL.
Fig. 14. Path diagram and coefficients (standardized beta weights) for Study 14. Solid paths are significant, p < .05. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 3.)
finding lends credence to the possibility that recall is perhaps a more proximate mediator of salience effects than segmentation. A path analysis was conducted to isolate the effects of segmentation and recall and determine if one, both, or neither mediates the influence of camera perspective on causality judgments. The results of this analysis can be seen in Fig. 14. Segmentation and recall independently accounted for significant portions of variance. Once segmentation and recall were partialed out, the remaining direct effect of camera perspective on causality judgments was rendered insignificant.
D. STUDY 15: AN EXPERIMENTAL DEMONSTRATION OF A DIRECT EFFECT OF SEGMENTATION ON CAUSALITY JUDGMENTS Study 14 showed that segmentation rate contributed unique variance to the effect of camera perspective on causality judgments. However, because the data in support of this conclusion were correlational in nature, the evidence favoring the kind of direct effect of perceptual segmentation on causality judgments depicted in Fig. 12 is not as compelling as it could be. The purpose of Study 15, then, was to tease apart experimentally the independent effects of segmentation from recall in order to establish more conclusively whether there is indeed a direct effect of segmentation on the point-of-view/salience bias (Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 4). The experiment replicated the design of Study 14 except for the following critical changes: All participants performed the segmentation task, but half additionally counted backward aloud as they viewed one of the two versions of the interaction sequence. This type of counting task has been found to reduce reliably the amount of information that
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
239
can be encoded into long-term memory (Reitman, 1971; Shiffrin, 1973) without, however, hindering the segmentation process (Lassiter, 1988; Lassiter, Apple, & Beers, 2000; Lassiter, Geers, Flannery, & Ploutz-Snyder, 1999). If memory truly mediates the effect of segmentation on the point-of-view/salience bias, then the counting manipulation should affect not just recall but the measures of perceived causality as well. However, if only the recall measure is affected by the counting manipulation, it would suggest that any effect of segmentation on the point-ofview/salience bias is a relatively direct (nonmemory-mediated) one. The counting task used in Study 15 should not only interfere with memory encoding processes, it should also disrupt subjects’ability to engage in higher-order reasoning processes (cf. Fleming & Arrowood, 1979; Lassiter, Apple, & Beers, 2000). That being the case, Study 15 will also allow for a test of an explanation of the point-of-view/salience bias that can be derived from Gilbert’s (1989, 1995) three-stage model of social inference. According to Gilbert, upon observing another’s behavior, perceivers first categorize the behavior or identify what the person is doing (e.g., acting anxiously). Next, perceivers characterize the behavior or identify a dispositional cause for it (e.g., the person is the anxious type). Finally, perceivers correct their dispositional characterization of the behavior by taking into account situational constraints on the behavior (e.g., the person was waiting to hear the results of a biopsy). Gilbert further contends that the first two stages of this process are largely spontaneous or perception-based and therefore require minimal cognitive resources. The correction stage, however, is said to be relatively resource dependent and thus more susceptible to disruption than the categorization or characterization stages. Several studies (e.g., Gilbert, Krull, & Pelham, 1988; Gilbert, Pelham, & Krull, 1988) have shown this model to be tenable. It might be the case, then, that the point-of-view/salience bias is the result of a combination of perceptual and higher-level cognitive processing. More specifically, the bias might initially arise as a result of a spontaneous, perceptual process; yet with sufficient cognitive resources available to the perceiver, it could be corrected, at least partially, by a higher-order reasoning or deliberation process. This line of explanation leads to the prediction that the bias is most likely to occur when perceivers do not have adequate cognitive resources available to engage in the relatively effortful correction process. In terms of Study 15, then, there is the possibility that the point-of-view/salience bias will be more pronounced in the counting condition than in the no-counting condition. Such a result would imply that some postperception inference process is perhaps necessary to fully account for the biasing effect of visual perspective. Two hundred twenty-one volunteers participated. After all participants were given the segmentation-task instructions, half of the them were additionally told to simultaneously perform the counting task. To further assess the generalizability of our findings, a second conversation tape was presented to half the participants.
240
G. DANIEL LASSITER ET AL.
Fig. 15. Difference scores (Person A − Person B in SD units) for Study 15. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 4.)
Counting and noncounting participants then viewed the stimulus conversation with either Person A salient or Person B salient. The main pattern of results for the segmentation, recall, and causality data are presented in Fig. 15. (Note that as displayed—in terms of difference scores— no main effect of the counting task would be evident.) First of all, the counting manipulation was successful in significantly impairing participants’ recall for the observed interaction. Whereas noncounters recalled a mean of 12.61 distinct pieces of information from the conversation, counters could only come up with a mean of 9.51. Importantly, neither the segmentation or causality data revealed any reliable effects of the counting manipulation. As can be seen in Fig. 15, however, all three dependent measures produced the characteristic salience-effect interaction pattern (which was significant in each case). The same path analysis was performed as in the previous study. Unlike in Study 14, recall was not a reliable mediator of the influence of camera perspective (see Fig. 16). Nonetheless, segmentation continued to account for a significant portion of variance in causality judgments. Finally, all results held true for both of the stimulus conversations. Overall, the findings of Study 15 provide the strongest results yet that differences in the initial perceptual registration of information likely contribute to the point-of-view/salience bias in causal attribution. Even when memory processes (and presumably other higher-order reasoning processes) were considerably degraded, the manner in which participants perceptually segmented the
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
241
Fig. 16. Path diagram and coefficients (standardized beta weights) for Study 15. Solid paths are significant, p < .05. (Data from Lassiter, Geers, Munhall, Ploutz-Snyder, & Breitenbecher, 2000, Study 4.)
interaction into its meaningful components—which was determined largely by their visual perspective—continued to influence impressions of the relative causal agency of the two interactants.
X. Theoretical and Practical Implications of the Present Program of Research Taken as a whole, the present investigations dramatically confirm what film theorists have long surmised: “The camera angle can force the viewer to observe a subject from a particular point of view, concentrate his attention on some aspect or characteristic of the subject, [or] shift his attention. . . . The manipulation of camera angles can be . . . significant in distorting and accenting content . . . thereby adding . . . psychological import not inherent in the material” (Jacobs, 1970, p. 26). As noted in Sections I and III, many legal scholars, criminal justice practitioners, political leaders, and social scientists have called for the universal adoption of videotaping as a “quick fix” for the problem of some innocent people being induced to incriminate themselves when confronted by standard police interrogation tactics. Our research indicates that the indiscriminate application of the videotaping procedure to solve the problem of coerced or false confessions slipping through the system could potentially exacerbate an already deplorable situation. As pointed out at the beginning of this chapter, in the United States and in many other countries (such as Canada, Australia, and the United Kingdom) videotaped interrogations and confessions are customarily recorded with the camera lens zeroed in on the suspect. One reason for this particular positioning of the camera is likely the belief that a careful examination of not only the suspects’ words, but
242
G. DANIEL LASSITER ET AL.
also their less conspicuous actions or expressions, will ultimately reveal the truth of the matter. As stated by Geller (1992, p. 44): The opportunity to assess a defendant’s veracity based on nonverbal cues is considered a very substantial benefit of videotaping—indeed, it is the principal reason many urge that criminal justice systems incur the expense of shifting from audio to video recording. As the New South Wales Police point out, Sigmund Freud in 1905 observed the way gestures and expressions provide a window into the psyche: “He that has eyes to see and ears to hear may convince himself that no mortal can keep a secret. If his lips are silent, he chatters with his fingertips, betrayal oozes out of him at every pore.”
The empirical validity of such beliefs aside,2 we have shown that focusing the video camera primarily on the suspect in an interrogation has the effect of impressing upon viewers the notion that his or her statements are more likely freely and intentionally given and not the result of some form of coercion. Moreover, a comparison of judgments derived from suspect-focus videotapes with judgments based on “control” media—transcripts and audiotapes—leads to the conclusion that the greater perceptions of voluntariness associated with suspect-focus videotapes is an unmistakable bias of the most serious kind—one that runs contrary to the cornerstone of our system of justice, the presumption of innocence. The camera may “never blink,” but that doesn’t mean what it “sees” can be considered an unadulterated view of reality. As the celebrated communications theorist Marshall McLuhan (1962, 1964, 1968) maintained, the information being conveyed is not entirely independent of the method of conveyance.
A. THE CAMERA PERSPECTIVE BIAS—A CASE OF MENTAL CONTAMINATION? Both for appreciating the tenacious nature of the camera perspective bias and for generating possible strategies for combating it, we believe it is useful to think about the bias in terms of the notion of mental contamination (Wilson & Brekke, 1994). Wilson and Brekke (1994, p. 117) defined mental contamination as “the process whereby a person has an unwanted judgment, emotion, or behavior because of mental processing that is unconscious or uncontrollable. By unwanted, we mean 2 Generally, people (with no special training) are not especially good at detecting deception and reading leaked cues (DePaulo et al., 1985; Kraut, 1980; Zuckerman et al., 1981). Interestingly, a recent study (Kassin & Fong, 1999) demonstrated that individuals who were taught to distinguish truth from deception by viewing videotapes used to train police interrogators (John E. Reid and Associates, 1991) were actually worse at accurately assessing the veracity of a “suspect’s” statements than untrained individuals. In addition, trained individuals—despite their lower accuracy—were more confident that their judgments were correct!
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
243
that the person making the judgment would prefer not to be influenced in the way he or she was.” These authors reviewed several factors that, to varying degrees, appear to contaminate, or “render impure or unsuitable,” people’s judgments and decisions in a variety of situations. Based on this review, Wilson and Brekke abstracted what they considered to be the four elements most responsible for instances of mental contamination. These are (1) a lack of awareness of the unwanted bias, (2) an absence of motivation to correct the bias, (3) an inability to accurately gauge the direction and magnitude of the bias, and (4) insufficient control over one’s mental processes or responses that potentially could correct the bias. According to Wilson and Brekke, all four of these impediments must be surmounted to avoid mental contamination or biased judgments. The present findings, when combined with the mental contamination framework, suggest that the prospects for overcoming the camera perspective bias, or salience effects in general, are bleak indeed. In none of our experiments was there even a scintilla of evidence to indicate that participants spontaneously, and on their own, became aware that their judgments were being affected by the camera angle. Thus, from the outset the deck is stacked against viewers—they are not the least bit suspicious that their judgments could be influenced by the camera’s point of view. Even when the potential prejudicial effects of camera positioning are explicitly brought to their attention, however, viewers appear to be no better off. Studies 4 and 10 showed that participants who received clear warnings about the biasing effect of camera perspective nonetheless manifested levels of mental contamination similar to those who did not receive the warning. Why would such straightforward attempts to debias the prejudicial effects of camera perspective fail so completely? Wilson and Brekke (1994) offered a couple of possibilities. First, they suggested that people may underestimate their own susceptibility to a given bias. Consistent with this notion, Wilson and Brekke described data showing that people believe others are more likely to be biased by various factors than they would be. Related to this, Wilson and Brekke noted that people were less likely to heed debiasing information if it was incongruent with their causal theories about what factors might reasonably be expected to sway their judgments. It is not hard to imagine that most people would not seriously entertain the possibility that their judgments about the content of a videotape could be altered simply by the positioning of the camera used to record it. The second reason why calling participants’ attention to the camera perspective bias did not immunize them against it may have to do with the fact that, according to Wilson and Brekke, people also overestimate their own mental control. Again, data discussed by Wilson and Brekke indicate that people do indeed seem to exhibit unwarranted faith in their ability to control their own thoughts and feelings and, as a result, speciously believe they can resist the influence of potentially biasing factors.
244
G. DANIEL LASSITER ET AL.
Perhaps the evidence we obtained for a biasing effect of camera perspective is due to the fact that our participants were not sufficiently motivated to form accurate judgments and therefore failed to correct the bias—the second hurdle in the mental contamination framework. We believe, however, that this is very unlikely given a number of diverse findings that lead to the conclusion that, at least in several of our studies, the motivation to be accurate and avoid bias was quite high. First, in Study 2 we measured participants’ level of need for cognition, which is associated with individual differences in the extent to which people desire accuracy and are willing and able to cognitively exert themselves to achieve it (cf. Cacioppo et al., 1996). Those individuals with the relatively highest level of cognitive motivation fared no better in correcting the bias than did those whose cognitive motivation was lower. This result is not a fluke, as Briggs and Lassiter (1994) replicated it twice using two different conversation stimuli (similar to those used in Stage Three). Second, the findings of Study 7, in which we directly manipulated the degree of accountability participants experienced regarding their judgments, also belie an interpretation of our results as being due to inadequate motivation for accuracy. Finally, we believe in all of our Stage Two experiments participants were taking the mock trial quite seriously and were doing their utmost to reach a fair and just decision based on the evidence that was presented. Many participants in the Stage Two studies who found the defendant not guilty seemed genuinely surprised and concerned upon learning their verdict was different than the one rendered in the actual trial. Remember also that in Study 11, after nearly 4 h of participation, our volunteers chose to deliberate about the evidence for on average 25 min, with on more than one occasion that being stretched to the full 45-min maximum. These are not the reactions we would expect from participants if they were truly blas´e about their role in the whole proceedings. Given that, at least in some of our studies, participants were likely both cognizant of the point-of-view bias and motivated to prevent it from intruding on their judgment processes (for example, in the Study 10 trial simulation, especially those who received the explicit warning of the camera’s possible prejudicial impact), why were they still unsuccessful at nullifying the bias? The third stage of the mental contamination framework suggests one possible explanation. That is, participants likely would have a difficult, if not impossible, task on their hands trying to ascertain exactly how much their judgments had been biased—a necessary precondition if they are to effect the appropriate correction. Participants would have to decide, for example, if their initial inclination to convict was swayed by the camera perspective and to what extent. In this instance, participants might ask themselves: “Could I have been affected so much by the camera angle that I should reconsider my decision and actually vote to acquit?” In light of this conundrum, it is little wonder no systematic evidence that participants could override the camera perspective bias was obtained in our experiments.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
245
A final factor, suggested by the mental contamination framework, that could have played a part in producing our results is that people just are not very efficacious with regard to controlling the various mental processes that can contribute to the decisions and judgments they make (cf. Wegner & Pennebaker, 1993). Consistent with this possibility, our Stage Three experiments do indicate that the camera perspective bias is, to some degree, tied to the earliest stages of information processing over which people may have somewhat less control. That is, confirming the earlier data of Newtson et al. (1978, p. 380), we found in Studies 12– 15 that “behavior that differs systematically in the availability of action-defining changes . . . [was] segmented into perceived actions in a systematically different way.” Although it is true that people can typically vary their perceptual processing of events to a considerable degree (e.g., Newtson, 1973), our Stage Three studies and Newtson et al.’s experiment clearly demonstrate that our visual perspective places constraints on how much and what kind of information we can extract from observed behavior sequences.
B. EXPOSURE CONTROL AS A REMEDY FOR THE BIASING EFFECT OF CAMERA PERSPECTIVE In light of our pessimistic characterization of the camera perspective bias as an instance of mental contamination that in all likelihood cannot be readily undone, is there any recommendation that we can suggest for preventing this bias from finding its way into real courtrooms? Wilson and Brekke (1994, p. 134) argued that when all else fails, “a final strategy for avoiding mental contamination is to make sure that it never has the opportunity to occur by avoiding contaminants that might bias one’s judgments.” Applying this strategy to the case of videotaped confessions would mean not allowing suspect-focus videotaped confessions to ever be introduced at trial. Are we thus recommending that videotaped interrogation and confession evidence not be used at all in courts of law? No, because our data indicate that when the camera perspective allows for the suspect and detective to be viewed equally well, there appears to be no discernible bias associated with the videotaping procedure. Interestingly, this very approach to preventing the point-of-view bias in videotaped confessions has already been established in New Zealand. In the early 1990s, the Police Executive Committee of New Zealand approved the videotaping of police interviews/interrogations on a national basis. In implementing this policy, various procedural guidelines were established. One critical issue that had to be dealt with was where to point the camera. In a letter we received from one of the authors of “The New Zealand Video Interview Project” (Lani W. Takitimu, personal communication, November 3, 1993), we were informed that “After reading your earlier literature on camera angle, we opted for showing side
246
G. DANIEL LASSITER ET AL.
profiles of both the Police Officer and the suspect, although we knew at the time, this was different to how they were recording interviews in parts of Australia, Canada and the United Kingdom.” Thus, New Zealand made it a national policy that police interrogations be videotaped from an equal-focus perspective based only on the first study conducted in this research program (Lassiter & Irvine, 1986). With the greater wealth of data that we now have on this topic, we do not hesitate to recommend that a similar policy be adopted in the United States as well as in the other above-mentioned countries. However, those who must make policy decisions regarding the implementation of the videotape method should not rule out the possibility of directing the camera primarily at the interrogator(s) whom a detained suspect must face. As the vast literature on attribution processes, and our Study 11 especially, indicate, this particular camera perspective may hold the greatest potential for facilitating judges and jurors’ all-important evaluations concerning the reliability of a given videotaped confession.
C. CONCLUDING REMARKS In its landmark Miranda v. Arizona (1966) ruling, the U. S. Supreme Court stated that individuals held for interrogation must be advised of their constitutional rights to silence and counsel, otherwise any statements they make—even if highly incriminating—would be considered inadmissible in a court of law. Prior to the Miranda ruling, a confession would be suppressed only if the determination was made that it resulted from some actual coercion—which in this age of psychological interrogation (cf. Leo & Ofshe, 1998) is no easy task. In an article appearing in the FBI Law Enforcement Bulletin, Hendrie (1997) reviewed a number of important developments affecting custodial interrogations since Miranda. From his review, Hendrie drew the following conclusion: The Supreme Court has implicitly abandoned the underlying principle of the Miranda decision—that custodial police interrogation is inherently coercive—and has carved out many exceptions to the Miranda exclusionary rule. Consequently, a violation of the Miranda ruling does not necessarily mean that a statement will be inadmissible. The Supreme Court has made it clear that the Miranda warnings are not constitutionally required but are only prophylactic rules designed to protect a suspect’s rights against compelled selfincrimination. Voluntariness remains the constitutional standard that must be met when obtaining a statement from a suspect. (p. 30, emphasis added)
To the extent that Hendrie is correct, we believe it would be, in the words of William James (1897, p. 19), an “awfully solemn” error to continue to permit suspect-focus videotaped confessions to be introduced as evidence in actual courts of law.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
247
Acknowledgments Aspects of this research were supported by funds from the Ohio Board of Regents and the National Science Foundation (BNS-8911067 and SBR-9514966). We thank Chaz Adam, Kevin Apple, Joe Balding, Marvin Bowman, Dave Breitenbecher, Mike Briggs, Kim Dudley, Stewart Dudley, Lisa Eliason, Michelle Gill, Andrea Hadley, Blis Hanousek-DeVault, Tom Hodson, Audrey Irvine. Birgit Jensen, Kelly Kinnison, Linda Koenig, Melanie LaForce, Bill Lassiter, Matthew Leafgren, Richard Leo, Frank LoSchiavo, Lindsay Munson, Richelle Newvahner, Laurie Olsen, Rob PloutzSnyder, Lisa Poerio, Michael Prisley, John Ray, Alicia Santuzzi, Jason Secondi, Carla Scanlan, Dave Slaw, Julie Stone, Katie Stricker, Linda Watkins, Paul Weiland, and Amy Wells for insightful discussions on the research/manuscript, for help with developing the trial simulations, or for assistance with data collection and analysis. If we inadvertently failed to acknowledge anyone’s contribution to this 17-year program of research, please forgive us.
References Anderson, N. H. (1974). Information integration theory: A brief survey. In D. Krantz, R. Atkinson, R. D. Luce, & P. Suppes (Eds.), Contemporary developments in mathematical psychology (Vol.2). San Francisco: Freeman. Arizona v. Fulminante, 111 S. Ct. 1246 (1991). Arkin, R., & Duval, S. (1975). Focus of attention and causal attributions of actor and observers. Journal of Experimental Social Psychology, 11, 427–438. Ayling, C. J. (1984). Corroborating false confessions: An empirical analysis of legal safeguards against false confessions. Wisconsin Law Review, 4, 1121–1204. Barthel, J. (1976). A death in Canaan. New York: Thomas Congdon Books. Bornstein, B. H. (1999). The ecological validity of jury simulations: Is the jury still out? Law and Human Behavior, 23 , 75–91. Bramel, D. (1969). Determinants of beliefs about other people. In J. Mills (Ed.), Experimental social psychology. New York: Macmillan. Bray, R. M., & Kerr, N. L. (1982). Methodological considerations in the study of the psychology of the courtroom. In N. L. Kerr & R. M. Bray (Eds.), The psychology of the courtroom (pp. 287–324). New York: Academic Press. Briggs, M. A., & Lassiter, G. D. (1994). More evidence for the robustness of salience effects. Journal of Social Behavior and Personality, 9, 171–180. Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42, 116–131. Cacioppo, J. T., & Petty, R. E. (1989). Effects of message repetition on argument processing, recall, and persuasion. Basic and Applied Social Psychology, 10, 3–12. Cacioppo, J. T., Petty, R. E., Feinstein, J., & Jarvis, B. (1996). Dispositional differences in cognitive motivation: The life and times of individuals low versus high in need for cognition. Psychological Bulletin, 119, 197–253. Cacioppo, J. T., Petty, R. E., & Kao, C. F. (1984). The efficient assessment of need for cognition. Journal of Personality Assessment, 48, 306–307. Cassell, P. G. (1996). All benefits, no costs: The grand illusion of Miranda’s defenders. Northwestern University Law Review, 90, 1084–1124. Cassell, P. G. (1999). The guilty and the “incocent”: An examination of alleged cases of wrongful conviction from false confessions. Harvard Journal of Law and Public Policy, 22, 523–603.
248
G. DANIEL LASSITER ET AL.
Cohen, C. E., & Ebbesen, E. B. (1979). Observational goals and schema activation: A theoretical framework for behavior perception. Journal of Experimental Social Psychology, 15, 305– 329. Cohn, A., & Udolf, R. (1979). The criminal justice system and its psychology. New York: Van Nostrand Reinhold. Connery, D. S. (1977). Guilty until proven innocent. New York: G. P. Putnam’s Sons. Cutler, B. L. (1988, December). Videotaped evidence in court. The American Psychological Association Monitor, 19, p. 32. Deaux, K., & Major, B. (1977). Sex-related patterns in the unit of perception. Personality and Social Psychology Bulletin, 3, 297–300. DePaulo, B. M., Stone, J. I., & Lassiter, G. D. (1985). Deceiving and detecting deceit. In B. R. Schenkler (Ed.), The self and social life (pp. 323–370). New York: McGraw-Hill. Diamond, S. S. (1997). Illuminations and shadows from jury simulations. Law and Human Behavior, 21, 561–571. Domash, S. F. (1985, October 6). Videotaped confessions grow. New York Times, Section 21, pp. 1 and 8. Duval, S., & Wicklund, R. A. (1972). A theory of objective self awareness. New York: Academic Press. Dwyer, J., Neufeld, P., & Scheck, B. (2000). Actual innocence: Five days to execution and other dispatches from the wrongly convicted. New York: Doubleday. Dywan, J., & Bowers, K. (1983). The use of hypnosis to enhance recall. Science, 222, 184–185. Ekman, P. (1992). Telling Lies: Clues to deceit in the marketplace, politics, and marriage. New York: Norton. Ellsworth, P. (1989). Are twelve heads better than one? Law and Contemporary Problems, 52, 205–224. Feild, H. S., & Barnett, N. J. (1978). Simulated jury trials: Students vs. “real” people as jurors. Journal of Social Psychology, 104, 287–293. Fiske, S. T., Kenny, D. A., & Taylor, S. E. (1982). Structural models for the mediation of salience effects on attribution. Journal of Experimental Social Psychology, 18, 105–127. Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York: McGraw-Hill. Fleming, J., & Arrowood, A. J. (1979). Information processing and the perseverance of discredited self-perceptions. Personality and Social Psychology Bulletin, 5, 201–205. Foss, R. D. (1976). Group decision processes in the simulated trial jury. Sociometry, 39, 305–316. Geers, A. L., & Lassiter, G. D. (1999). Affective expectations and information gain: Evidence for assimilation and contrast effects in affective experience. Journal of Experimental Social Psychology, 35, 394–413. Geller, W. A. (1992). Police videotaping of suspect interrogations and confessions: A preliminary examination of issues and practices (A report to the National Institute of Justice). Washington, DC: U.S. Department of Justice. Gilbert, D. T. (1989). Thinking lightly about others: Automatic components of the social inference process. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought: Limits of awareness, intention, and control (pp. 189–211). New York: Guilford. Gilbert, D. T. (1995). Attribution and interpersonal perception. In A. Tesser (Ed.), Advanced social psychology (pp. 99–147). New York: McGraw-Hill. Gilbert, D. T., Krull, D. S., & Pelham, B. W. (1988). Of thoughts unspoken: Social inference and the self-regulation of behavior. Journal of Personality and Social Psychology, 55, 685–694. Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 117, 21–38. Gilbert, D. T., Pelham, B. W., & Krull, D. S. (1988). On cognitive busyness: When person perceivers meet persons perceived. Journal of Personality and Social Psychology, 54, 733–740. Grano, J. D. (1993). Confessions, truth, and the law. Ann Arbor: Univ. of Michigan Press. Graziano, W. G., Moore, J. S., & Collins, J. E., II (1988). Social cognition as segmentation of the stream of behavior. Developmental Psychology, 24, 568–573.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
249
Grush, J. E. (1976). Attitude formation and mere exposure phenomena: A nonartificial explanation of empirical findings. Journal of Personality and Social Psychology, 33, 281–290. Gudjonsson, G. (1992). The psychology of interrogations, confessions and testimony. Chichester, UK: Wiley. Hanson, C., & Hirst, W. (1989). On the representation of events: A study of orientation, recall, and recognition. Journal of Experimental Psychology: General, 118, 136–147. Harrison, A. A. (1977). Mere exposure. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 39–83). New York: Academic Press. Heider, F. (1944). Social perception and phenomenal causality. Psychological Review, 51, 358–374. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hendrie, E. M. (1997). Beyond Miranda. FBI Law Enforcement Bulletin, 66, 25–32. Heavner, J. M. (1984). Admissibility of confessions: The voluntariness requirement and police trickery in North Carolina. Wake Forest Law Review, 20, 251–275. Hogue, T. E., & Atkinson, M. L. (1989). Approach and avoidance of social information. Canadian Journal of Behavioural Science, 21, 310–322. Horowitz, I. A., & Willging, T. E. (1984). The psychology of law: Integrations and applications. Boston: Little,Brown. Inbau, F. E., Reid, J. E., & Buckley, J. P. (1986). Criminal interrogation and confessions (3rd ed.). Baltimore: Williams & Wilkins. Jacobs, L. (1970). The movies as medium. New York: Farrar, Straus & Giroux. James, W. (1897). The will to believe and other essays in popular philosophy. New York: Longmans, Green. John E. Reid and Associates. (1991). The Reid Technique: Interviewing and interrogation. [Videotape] Johnson, G. (1997). False confessions and fundamental fairness: The need for electronic recording of custodial interrogations. Boston University Public Interest Law Journal, 6, 719–751. Jones, E. E. (1979). The rocky road from acts to dispositions. American Psychologist, 34, 107– 117. Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process in person perception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 219–266). New York: Academic Press. Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental Social Psychology, 3, 2–24. Jones, E. E., Kanouse, D. E., Kelley, H. H., Nisbett, R. E., Valins, S., & Weiner, B. (1972). Attribution: Perceiving the causes of behavior. Morristown, NJ: General Learning Press. Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: Divergent perceptions of the causes of behavior. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the causes of behavior (pp. 79–94). Morristown, NY: General Learning Press. Jones, E. E., Rock, L., Shaver, K. G., Goethals, G. R., & Ward, L. M. (1968). Pattern of performance and ability attribution: An unexpected primacy effect. Journal of Personality and Social Psychology, 10, 317–340. Kamisar, Y., LaFave, W., & Israel, J. (1994). Modern criminal procedure (8th ed.). St. Paul, MN: West. Kaplan, M. F. (1975). Information integration in social judgment: Interaction of the judge and informational components. In M. F. Kaplan & Schwartz (Eds.), Human judgment and decision processes. New York: Academic Press. Kaplan, M. F. (1982). Cognitive processes in the individual juror. In N. L. Kerr & R. M. Bray (Eds.), The psychology of the courtroom (pp. 197–220). New York: Academic Press. Kaplan, M. F., & Miller, L. E. (1978). Reducing the effects of juror bias. Journal of Personality and Social Psychology, 36, 1443–1455. Kassin, S. M. (1997). The psychology of confession evidence. American Psychologist, 52, 221–233.
250
G. DANIEL LASSITER ET AL.
Kassin, S. M., & Fong, C. T. (1999). “I’m innocent!”: Effects of training on judgments of truth and deception in the interrogation room. Law and Human Behavior, 23, 499–516. Kassin, S. M., & Kiechel, K. L. (1996). The social psychology of false confessions: Compliance, internalization, and confabulation. Psychological Science, 7, 125–128. Kassin, S. M., & McNall, K. (1991). Police interrogations and confessions. Law and Human Behavior, 15, 231–251. Kassin, S. M., & Neumann, K. (1997). On the power of confession evidence: An experimental test of the fundamental difference hypothesis. Law and Human Behavior, 21, 469–484. Kassin, S. M., & Sukel, H. (1997). Coerced confessions and the jury: An experimental test of the “harmless error” rule. Law and Human Behavior, 21, 27–46. Kassin, S. M., & Wrightsman, L. S. (1979). On the requirements of proof: The timing of judicial instruction and mock juror verdicts. Journal of Personality and Social Psychology, 37, 1877– 1887. Kassin, S. M., & Wrightsman, L. S. (1980). Prior confessions and mock juror verdicts. Journal of Applied Social Psychology, 10, 133–146. Kassin, S. M., & Wrightsman, L. S. (1981). Coerced confessions, judicial instruction, and mock juror verdicts. Journal of Applied Social Psychology, 11, 489–506. Kassin, S. M., & Wrightsman, L. S. (1985). Confession evidence. In S. Kassin & L. Wrightsman (Eds.), The Psychology of evidence and trial procedure. Beverly Hills, CA: Sage. Kassin, S. M., & Wrightsman, L. S. (1988). The American jury on trial: Psychological perspectives. New York: Hemisphere. Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska Symposium on Motivation (Vol. 15, pp. 192–238). Lincoln: Nebraska Univ. Press. Kelley, H. H. (1972). Attribution in social interaction. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the causes of behavior (pp. 1–26). Morristown, NY: General Learning Press. Kenny, D. A. (1979). Correlation and causality. New York: Wiley–Interscience. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T., Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 1, pp. 233–265). New York: McGraw-Hill. Kenrick, D. T., Neuberg, S. L., & Cialdini, R. B. (1999). Social psychology: Unraveling the mystery. Needham Heights, MA: Allyn & Bacon. Kerr, N. L. (1978). Severity of prescribed penalty and mock jurors’ verdicts. Journal of Personality and Social Psychology, 36, 1431–1442. Kraut, R. E. (1980). Humans as lie detectors: Some second thoughts. Journal of Communication, 39, 209–216. Lassiter, G. D. (1986). Effect of superfluous deterrence on the perception of others. Journal of Experimental Social Psychology, 22, 163–175. Lassiter, G. D. (1988). Behavior perception, affect, and memory. Social Cognition, 6, 150–176. Lassiter, G. D., Apple, K. J., & Beers, M. J. (2000). Distinguishing between perceptual differentiation and cognitive elaboration: A dual-process model of action memory. Unpublished manuscript, Ohio University, Athens. Lassiter, G. D., Beers, M. J., Geers, A. L., Handley, I. M., Munhall, P. J., & Weiland, P. E. (in press). Further evidence for a robust point-of-view bias in videotaped confessions. Current Psychology. Lassiter, G. D., Briggs, M. A., & Bowman, R. E. (1991). Need for cognition and the perception of ongoing behavior. Personality and Social Psychology Bulletin, 17, 156–160. Lassiter, G. D., Briggs, M. A., & Slaw, R. D. (1991). Need for cognition, causal processing, and memory for behavior. Personality and Social Psychology Bulletin, 17, 694–700. Lassiter, G. D., & Dudley, K. A. (1991). The a priori value of basic research: The case of videotaped confessions. Journal of Social Behavior and Personality, 6, 7–16.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
251
Lassiter, G. D., Geers, A. L., Apple, K. J., & Beers, M. J. (2000) Observational goals and behavior unitization: A reexamination. Journal of Experimental Social Psychology, 36, 649–659. Lassiter, G. D., Geers, A. L., Flannery, B. C., & Ploutz-Snyder, R. J. (1999). Attributional consequences of variation in behavior perception: It’s what you unitize and how you unitize it. Unpublished manuscript, Ohio University, Athens. Lassiter, G. D., Geers, A. L., Munhall, P. J., Handley, I. M., & Weiland, P. E. (2000). Videotaped confessions: Camera perspective biases verdicts in a simulated trial. Manuscript in preparation. Lassiter, G. D., Geers, A. L., Munhall, P. J., Ploutz-Snyder, R. J., & Breitenbecher, D. L. (2000). Evidence for the perceptual mediation of salience effects. Manuscript in preparation. Lassiter, G. D., Koenig, L. J., & Apple, K. J. (1996). Mood and behavior perception: Dysphoria can increase and decrease effortful processing of information. Personality and Social Psychology Bulletin, 22, 794–810. Lassiter, G. D., & Irvine, A. A. (1986). Videotaped confessions: The impact of camera point of view on judgments of coercion. Journal of Applied Social Psychology, 16, 268–276. Lassiter, G. D., Munhall, P. J., Geers, A. L., Weiland, P. E., & Handley, I. M. (2000). [Accountability and the camera perspective bias in videotaped confessions]. Unpublished raw data. Lassiter, G. D., Munhall, P. J., Geers, A. L., Handley, I. M., & Weiland, P. E. (2000). [Does camera perspective affect the perceived veracity of videotaped confessions?]. Unpublished raw data. Lassiter, G. D., & Slaw, R. D. (1991). The unitization and memory of events. Journal of Experimental Psychology: General, 120, 80–82. Lassiter, G. D., Slaw, R. D., Briggs, M. A., & Scanlan, C. R. (1992). The potential for bias in videotaped confessions. Journal of Applied Social Psychology, 22, 1838–1851. Lassiter, G. D., & Stone, J. I. (1984). Affective consequences of variation in behavior perception: When liking is in the level of analysis. Personality and Social Psychology Bulletin, 10, 253– 259. Lassiter, G. D., Stone, J. I., & Rogers, S. L. (1988). Memorial consequences of variation in behavior perception. Journal of Experimental Social Psychology, 24, 222–239. Laurence, J.-R., & Perry, C. (1983). Hypnotically created memory among highly hypnotizable subjects. Science, 222, 523–524. Lego v. Twomey, 404 U.S. 477 (1972). Leo, R. A. (1992). From coercion to deception: The changing face of police interrogation in America. Crime, Law, and Social Change, 18, 35–39. Leo, R. A. (1996a). The impact of Miranda revisited. The Journal of Criminal Law and Criminology, 86, 621–692. Leo, R. A. (1996b). Inside the interrogation room. The Journal of Criminal Law and Criminology, 86, 266–303. Leo, R. A., & Ofshe, R. J. (1998). The consequences of false confessions: Deprivations of liberty and miscarriages of justice in the age of psychological interrogation. The Journal of Criminal Law and Criminology, 88, 429–496. Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects of accountability. Psychological Bulletin, 125, 255–275. Liebman, J. S., Fagan, J., & West, V. (2000). A broken system: Error rates in capital cases, 1973–1995. New York: Columbia Univ. School of Law. MacCoun, R. J. (1989). Experimental research on jury decision-making. Science, 244, 1046–1050. MacCoun, R. J., & Kerr, N. L. (1988). Asymmetric influence in mock jury deliberation: Jurors’ bias for leniency. Journal of Personality and Social Psychology, 54, 21–33. Markus, H., Smith, J., & Moreland, R. L. (1985). Role of the self-concept in the perception of others. Journal of Personality and Social Psychology, 49, 1494–1512. Massad, C. M., Hubbard, M., & Newtson, D. (1979). Selective perception of events. Journal of Experimental Social Psychology, 15, 513–532.
252
G. DANIEL LASSITER ET AL.
Mathes, W. C., & DeVitt, E. J. (1965). Federal jury practice and instructions. St. Paul, MN: West. McArthur, L. Z. (1972). The how and what of why: Some determinants and consequences of causal attribution. Journal of Personality and Social Psychology, 22, 171–193. McArthur, L. Z. (1980). Illusory causation and illusory correlation: Two epistemological accounts. Personality and Social Psychology Bulletin, 6, 507–519. McArthur, L. Z. (1981). What grabs you? The role of attention in impression formation and causal attribution. In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition: The Ontario symposium (Vol. 1, pp. 201–241). Hillsdale, NJ: Erlbaum. McCormick, C. T. (1972). Handbook of the law of evidence (2nd ed.). St. Paul, MN: West. McCoy, M. L., Nunez, N., & Dammeyer, M. M. (1999). The effect of jury deliberations on jurors’ reasoning skills. Law and Human Behavior, 23, 557–575. McLuhan, H. M. (1962). The Gutenberg galaxy. Toronto: Univ. of Toronto Press. McLuhan, H. M. (1964). Understanding media. New York: McGraw-Hill. McLuhan, H. M., & Fiore, Q. (1968). War and peace in the global village. New York: McGraw-Hill. Miranda v. Arizona, 384 U.S. 336 (1996). Mitchell, H. E., & Byrne, D. (1973). The defendant’s dilemma: Effects of jurors’ attitudes and authoritarianism on judicial decisions. Journal of Personality and Social Psychology, 25, 123–129. Myers, D. G., & Kaplan, M. F. (1976). Group induced polarization in simulated juries. Personality and Social Psychology Bulletin, 2, 63–66. Myers, D. G., & Lamm, H. (1976). The group polarization phenomenon. Psychological Bulletin, 83, 602–627. Newtson, D. (1973). Attribution and the unit of perception of ongoing behavior. Journal of Personality and Social Psychology, 28, 28–38. Newtson, D. (1976). Foundations of attribution: The perception of ongoing behavior. In J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attribution research (Vol. 1, pp. 223–247). Hillsdale, NJ: Erlbaum. Newtson, D. (1980). An interactionist perspective on social knowing. Personality & Social Psychology Bulletin, 6, 520–531. Newtson, D., & Engquist, G. (1976). The perceptual organization of ongoing behavior. Journal of Experimental Social Psychology, 12, 436–450. Newtson, D., Engquist, G., & Bois, J. (1976). The reliability of a measure of behavior perception. JSAS Catalog of Selected Documents in Psychology, 6, 5. Newtson, D., Engquist, G., & Bois, J. (1977). The objective basis of behavior units. Journal of Personality and Social Psychology, 12, 847–862. Newtson, D., Hairfield, J., Bloomingdale, J., & Cutino, S. (1987). The structure of action and interaction. Social Cognition, 5, 191–237. Newtson, D., & Rindner, R. (1979). Variation in behavior perception and ability attribution. Journal of Personality and Social Psychology, 37, 1847–1858. Newtson, D., Rindner, R. J., Miller, R., & LaCross, (1978). Effects of availability of feature changes on behavior segmentation. Journal of Experimental Social Psychology, 14, 379–388. Nisbett, R., & Ross, L. (1971). Human inference: Strategies and shortcomings of social judgment. Englewood Cliffs, NJ: Prentice-Hall. Ofshe, R. J. (1989). Coerced confessions: The logic of seemingly irrational action. Cultic Studies Journal, 6, 6–15. Ofshe, R. J., & Leo, R. A. (1997). The social psychology of police interrogation: The theory and classification of true and false confessions. Studies in Law, Politics and Society, 16, 189–251. Perlman, D., & Oskamp, S. (1971). The effects of picture context and exposure frequency on evaluations of Negroes and Whites. Journal of Experimental Social Psychology, 7, 503–514. Pratkanis, A. R., & Aronson, E. (1991). Age of propaganda: The everyday use and abuse of persuasion. New York: Freeman.
VIDEOTAPED CONFESSIONS: IS GUILT IN THE EYE OF THE CAMERA?
253
Regan, D. T., & Totten, J. (1975). Empathy and attribution: Turning observers into actors. Journal of Personality & Social Psychology, 32, 850–856. Reiser, M. (1980). Handbook of investigative hypnosis. Los Angeles, CA: Law Enforcement Hypnosis Institute. Reitman, J. S. (1971). Mechanisms of forgetting in short-term memory. Cognitive Psychology, 2, 185–195. Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 174–220). New York: Academic Press. Ross, L., Amabile, T. M., & Steinmetz, J. L. (1977). Social roles, social control, and biases in social perception processes. Journal of Personality and Social Psychology, 35, 485–494. Ross, L., Greene, D., & House, P. (1977). The “false consensus effect”: An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13, 279– 301. Ross, M. (1975). Salience of reward and intrinsic motivation. Journal of Personality and Social Psychology, 32, 245–254. Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of Personality and Social Psychology, 37, 322–336. Sasaki, D. W. (1988). Guarding the guardians: Police trickery and confessions. Stanford Law Review, 40, 1593–1616. Shiffrin, R. M. (1973). Information persistence in short-term memory. Journal of Experimental Psychology, 100, 39–49. “Smile, You’re on the D.A.’s Camera.” (1983, June 27). Time, p. 61. Smith, E. R., & Miller, F. D. (1979). Salience and the cognitive mediation of attribution. Journal of Personality and Social Psychology, 37, 2240–2252. Storms, M. D. (1973). Videotape and the attribution process: Reversing actors’ and observers’ points of view. Journal of Personality and Social Psychology, 27, 165–175. Swap, W. C. (1977). Interpersonal attraction and repeated exposure to rewarders and punishers. Personality and Social Psychology Bulletin, 3, 248–251. Taylor, S. E., & Fiske, S. T. (1975). Point of view and perceptions of causality. Jounal of Personality and Social Psychology, 32, 439–445. Taylor, S. E., & Fiske, S. T. (1978). Salience, attention, and attribution: Top of the head phenomena. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 11, pp. 249–288). New York: Academic Press. Taylor, S. E., & Thompson, S. C. (1982). Stalking the elusive “vividness” effect. Psychological Review, 89, 155–181. Taylor, S. E., Crocker, J., Fiske, S. T., Sprinzen, M., & Winkler, J. D. (1979). The generalizability of salience effects. Journal of Personality and Social Psychology, 37, 357–368. Thomas, J. G. (1979). Police use of trickery as an interrogation technique. Vanderbilt Law Review, 32, 1167–1213. Wald, M., Ayres, R., Hess, D. W., Schantz, M., & Whitebread, C. H. (1967). Interrogations in New Haven: The impact of Miranda. The Yale Law Journal, 76, 1519–1648. Wegner, D. M., & Pennebaker, J. W. (1993). The handbook of mental control. Englewood, NJ: PrenticeHall. Weiner, B., Frieze, I., Kukla, A., Reed, L., Rest, S., & Rosenbaum, R. M. (1972). Perceiving the causes of success and failure. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the causes of behavior (pp. 95–120). Morristown, NY: General Learning Press. Wells, G. L. (1978). Applied eyewitness-testimony research: System-variables and estimator variables. Journal of Personality and Social Psychology, 36, 1546–1557.
254
G. DANIEL LASSITER ET AL.
Wells, G. L. (1980). Asymmetric attributions for compliance: Reward vs. punishment. Journal of Experimental Social Psychology, 16, 47–60. Wells, G. L., Malpass, R. S., Lindsay, R. C. L., Fisher, R. P., Turtle, J. W., & Fulero, S. M. (2000). From the lab to the police station: A successful application of eyewitness research. American Psychologist, 55, 581–598. White, W. S. (1979). Police trickery in inducing confessions. University of Pennsylvania Law Review, 127, 581–629. White, W. S. (1997). False confessions and the Constitution: Safeguards against untrustworthy confessions. Harvard Civil Rights–Civil Liberties Law Review, 32, 105–157. Wigmore, J. H. (1970). Evidence (Vol. 3) (revised by J. H. Chadbourn). Boston: Little, Brown. Wilder, D. (1978a). Effects of predictability on units of perception and attribution. Personality and Social Psychology Bulletin, 4, 281–284. Wilder, D. (1978b). Predictability of behaviors, goals, and unit of perception. Personality and Social Psychology Bulletin, 4, 604–607. Wilson, T. D., & Brekke, N. (1994). Mental contamination and mental correction: Unwanted influences on judgments and evaluations. Psychological Bulletin, 116, 117–142. Wilson, T. D., & Lassiter, G. D. (1982). Increasing intrinsic interest with superfluous extrinsic constraints. Journal of Personality & Social Psychology, 42, 811–819. Wolf, S., & Montgomery, D. A. (1977). Effects of inadmissible evidence and level of judicial admonishment to disregard on the judgments of mock jurors. Journal of Applied Social Psychology, 7, 205–219. Wrightsman, L. S., & Kassin, S. M. (1993). Confessions in the courtroom. Newbury Park, CA: Sage. Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9, 1–27. Zuckerman, M., DePaulo, B. M., & Rosenthal, R. (1981). Verbal and nonverbal communication of deception. In L. Berkowitz (Ed.), Advances in experimental social psychology. (Vol. 14, pp. 1–59). New York: Academic Press.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE: AN INTEGRATIVE ANALYSIS WITH APPLICATIONS IN SOCIAL PSYCHOLOGY Rex A. Wright Leslie D. Kirby
For a number of years now, studies in our laboratory have explored implications of an integrative analysis concerned with the determinants of cardiovascular (CV) responses (elevations above baseline) in people confronted with behavioral challenges (Wright, 1996, 1998). The analysis takes as a working hypothesis the proposition by Obrist (1976, 1981) that sympathetic nervous system influence on the CV system is proportional to effort or task engagement (what Obrist termed “active coping”). It then uses a motivational intensity theory by Brehm (e.g., Brehm & Self, 1989) and an extension from it to specify when people should manifest greater and lesser degrees of effort-related CV response. Much of the research inspired by the integrative analysis is psychophysiological in nature. Furthermore, relevant publications have appeared largely in books and journals outside the social psychological mainstream. However, these considerations should not be taken to suggest that the integrative analysis has no relevance to social psychology. To the contrary, it has considerable relevance for at least two reasons. First, the analysis constitutes a conceptual framework for analyzing the CV influence of variables in which social psychologists have long been interested. Perhaps foremost among these variables are different emotional states (Ax, 1953; Schachter, 1964; Zillmann, 1983). Also among them are situational and personal factors such as social evaluation, optimism, and efficacy. Research examining the CV influence of social variables has attracted special attention recently because of the possibility that chronically exaggerated CV responses may lead to negative health endpoints, such as CV disease and hypertension (Blascovich & Katkin, 1993; Contrada, Cather, & O’Leary, 1999). If CV responsivity increases health risk, then so may emotions, social situations, and personal characteristics associated with elevated CV response levels (T. Smith & Gerin, 1998). ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY, VOL. 33
255
C 2001 by Academic Press, Inc. Copyright All rights of reproduction in any form reserved. 0065-2601/01 $35.00
256
REX A. WRIGHT AND LESLIE D. KIRBY
Second, by identifying CV outcomes that should vary with effort, the analysis suggests a technique for documenting effort assumptions in social psychological research. This is significant because the construct of effort is a part of countless social analyses, related to topics ranging from social loafing (Latan´e, Williams, & Harkins, 1979; Shepperd, 1993) to learned helplessness (Abramson, Seligman, & Teasdale, 1978). Although social effort assumptions are ubiquitous, they have proved notoriously difficult to verify. Most often, investigators evaluate them by way of self-report and performance. However, such measures are suspect when used for this purpose. Effort reports are suspect because they are subject to social desirability and defensiveness influences (Pyszczynski & Greenberg, 1983; Rhodewalt & Fairfield, 1991). Performance is suspect because its relation to effort is known to vary, depending on such things as task complexity and performer competence (Heckhausen, 1991). Inclusion of CV measures in relevant social research protocols could provide convergent effort information and improve markedly investigators’ understanding of processes being studied. Our main goal in preparing this chapter is to bring to the attention of the broader social psychological research community the integrative CV response analysis and the body of evidence supporting it. As most other presentations of the analysis have appeared outside the social mainstream, they may well have fallen outside many social psychologists’ scope of awareness. Additional goals are to (1) update previous reviews of evidence relevant to the integrative view, (2) consider in greater detail than has been done previously the motivational process in circumstances involving challenges that have been characterized as “unfixed” (Brehm & Self, 1989), and (3) contrast the integrative CV response perspective with some popular alternative CV response views. Once we have met these aims, we comment on the research findings, point out some unresolved issues, analyze the CV influence of some specific social variables in the present integrative terms, and identify some social psychological studies that might benefit by including CV measures in their research protocols. We begin by describing the alternative CV response views with which the integrative CV response perspective eventually will be contrasted. The alternative view descriptions will place the integrative formulation in context and set the stage for the later contrasts.
I. Psychological Determinants of Cardiovascular Response: Three Conceptions Notions concerning the psychological determination of CV response have varied over time. However, two themes have been dominant in the past half-century, and one contemporary hypothesis has received particular attention in the past
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
257
decade, especially in social psychological publications. Therefore, these will be the alternative views on which we focus. One of the dominant CV response themes is that where behavioral contingencies are positive (i.e., where a good performance will secure something pleasant), CV responsivity is proportional to the perceived value of incentives driving behavior (Elliott, 1969; Fowles, 1980, 1982). Thus, for example, CV responses should be greater in people paid a lot for success on a task than in people paid a little for success on the task. Some research scientists have contended that the impact of incentive value appraisals on CV responses is direct (e.g., Fowles, Fisher, & Tranel, 1982). Others have assumed that the impact is indirect, mediated by the degree of effort, or engagement, that is engendered during task performance (T. Smith, Allred, Morrison, & Carlson, 1989; T. Smith, Nealey, Kircher, & Limon, 1997; T. Smith, Ruiz, & Uchino, 2000). The other dominant theme is that where at least some behavioral contingencies are negative (i.e., where a good performance will avert, terminate, or temper something unpleasant) CV responsivity is proportional to subjective threat (Cox, 1978; Lazarus, 1968). The threat/CV-response connection generally has been assumed to be direct, and its presumed existence has led not only to the expectation that CV responses will be greater where potential negative outcomes are highly aversive than where they are mildly aversive, but also to the suggestion that CV responses should be related to appraisals of control (Bandura, 1991; Williams, 1995). Theoretically, those who view themselves as having little control over negative outcomes should experience more threat and thus CV arousal than those who view themselves as having great control over the outcomes. The contemporary hypothesis that has received special attention recently is that CV responses in (attractive and aversive) motivated performance situations are determined by the relation between primary and secondary appraisals of those situations. This view derives from the biopsychosocial arousal regulation (BAR) model proposed by Blascovich and his colleagues (Blascovich & Mendes, 1999; Blascovich, Mendes, Hunter, & Salomon, 1999; Blascovich & Tomaka, 1996; Tomaka, Blascovich, Kibler, & Ernst, 1997; Tomaka, Palacious, Schneider, Colotla, Concha, & Herrald, 1999, see also Dienstbier, 1989; Lazarus & Folkman, 1994). In BAR terms, primary appraisal refers to perceived demand, but not demand in the usual “requirement” sense of the word. Instead, demand is conceived as an evaluation resulting from a calculus involving three perceptual elements: required effort, uncertainty, and danger. Secondary appraisal, in BAR terms, refers to coping resources (ability or efficacy). If resources meet or exceed “demand,” then the overall appraisal is thought to be one of challenge. By contrast, if demand exceeds resources, then the overall appraisal is thought to be one of threat.1 1
This characterizes the theory as it has most commonly been described. However, Blascovich and Mendes (1999) recently speculated that challenge appraisals also may occur where resources fall just short of demand.
258
REX A. WRIGHT AND LESLIE D. KIRBY
A central BAR tenet is that challenge appraisals induce marked cardiac increases paired with a decline in total peripheral resistance (TPR—resistance to flow in the circulatory network), whereas threat appraisals induce modest cardiac and resistance increases. Because of its partial focus on threat appraisals, the BAR view can be seen as a variation on the more traditional threat CV response theme.
II. The Integrative Analytic Perspective As we noted at the outset, the integrative effort analysis of CV response is made up of two components. The first is Obrist’s active coping proposal; the second is Brehm’s intensity theory.
A. ENGAGEMENT, SYMPATHETIC OUTFLOW, AND CARDIOVASCULAR RESPONSE The Obrist proposal—that sympathetic influence on the CV system is proportional to engagement—is similar, of course, to the effort idea that underlies one version of the traditional incentive hypothesis. However, it differs from the incentive effort idea in two important respects. First, it assumes that an engagement/CVresponse link will be present irrespective of whether behavioral contingencies are favorable or unfavorable. Second, it identifies a specific mechanism underlying effort-related CV adjustment: activity in the sympathetic branch of the autonomic nervous system.2 Although it is not necessary for present purposes to understand the full physiological ramifications of Obrist’s proposal (for relevant discussions, see Berntson et al., 1993; Hassett & Danforth, 1982; Obrist, 1981; Papillo & Shapiro, 1990), it is important to understand that the proposal implies that the most common CV parameters studied—heart rate (HR), systolic blood pressure (SBP), and diastolic blood pressure (DBP)—should not be equally likely show effort effects. In theory, the most sensitive of these three CV variables to sympathetic nervous system influence should be SBP, which is defined as blood pressure at the peak of a pulse following a heartbeat. Systolic blood pressure is a function of (1) resistance in the vascular network as a whole (i.e., TPR) and (2) the force with which the 2 The autonomic nervous system has two branches, the sympathetic branch and the parasympathetic branch. Although the integrative analysis takes as a working hypothesis the idea that engagement effects are sympathetically mediated, it recognizes that alternative CV response mediation is possible. For example, heart rate is partially under parasympathetic control, and effort influence on that parameter could sometimes involve no sympathetic contribution (Berntson, Cacioppo, & Quigley, 1993).
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
259
heart contracts on a given beat (myocardial contractility). Total peripheral resistance is thought not to be systematically related to sympathetic arousal because a sympathetic discharge has the effect of enlarging the caliber of vessels in some vascular beds while diminishing the caliber of vessels in others. It is the balance of constrictive and dilatory effects that determines TPR; hence, that parameter can rise, fall, or show no change in response to increased sympathetic activation (Obrist, 1981; Papillo & Shapiro, 1990). Myocardial contractility, by contrast, is directly enhanced by sympathetic outflow (e.g., Papillo & Shapiro, 1990). Because sympathetic excitation potentiates contractile force, and that force has an elevating influence on SBP, sympathetic excitation generally is expected to raise SBP. One condition under which SBP would not be expected to index sympathetic activity is where a sympathetic discharge yielded a TPR decrease that was substantial enough to offset a positive contractility effect. Another condition might be where sympathetic activity was so low that its impact on contractility was masked by some countervailing nonautonomic influence, such as that of baroreceptors, which sense vessel wall extension and reflexively reduce contractility in response to a pressure increase (Papillo & Shapiro, 1993). Even in this latter circumstance, a SBP effect should be in evidence if TPR increased to a sufficient degree. Heart rate should be somewhat sensitive to sympathetic influence, but not as sensitive as SBP to that influence, for two reasons (Obrist, 1981). First, HR is controlled not only by the (rate-accelerating) sympathetic branch of the autonomic nervous system, but also by the (rate-decelerating) parasympathetic branch of that system. Second, parasympathetic influence on HR tends to dominate where behavioral demands are in a low-to-moderate range. Thus, where sympathetic activity is modest, its influence on HR could be weakened or masked by countervailing parasympathetic activity. Generally speaking, DBP—or blood pressure in the trough between pulses— should be a poor sympathetic index because this parameter is determined chiefly by TPR.3 As noted above, TPR has potential for rising, falling, or not changing in the presence of increased sympathetic activity. Therefore, although DBP can parallel the level of sympathetic excitation, it by no means will necessarily. 3 Diastolic blood pressure also can be affected by HR in that the pace of pumping determines the time available for pressure to descend to a low point following a heartbeat. However, the effect of HR on DBP is likely to be limited for a least two reasons. First, up to a point, there can be HR acceleration and still an adequate interval between beats to allow pressure to reach the nadir set by the level of TPR. Hence, holding constant the rate of postbeat pressure descent, modest HR increases would be expected to have little or no diastolic effect. Second, even substantial HR increases should not necessarily impact DBP because the rate of postbeat pressure descent in fact varies with TPR (Obrist, 1981). The lower the TPR, the more quickly pressure diminishes following a beat. If pressure drops quickly enough, it can attain the low value set by TPR even with a very brief beat-to-beat interval.
260
REX A. WRIGHT AND LESLIE D. KIRBY
B. BREHM’S INTENSITY THEORY 1. Basic Propositions Brehm’s motivational intensity theory was developed to account for momentary effort—effort expended at a particular point in time. Unlike most relevant motivational analyses (see e.g., Geen, 1995; Heckhausen, 1991), it proposes that effort is not determined directly by variables associated with the benefit that may be accrued by success, but rather by what can, will, and must be done to succeed. So long as success is viewed as possible and worthwhile, effort should correspond to the difficulty of activity currently required. Where success is viewed as impossible or excessively difficult, effort should be low. In the former case, effort is understood to be futile; in the latter, anticipated cost outweighs potential benefit. Theoretically, the role of benefit variables should be to determine success importance and thus the upper bound of what people would be willing to do to succeed (termed the level of potential motivation). Included among benefit variables thought to determine success importance are need, incentive value, and the expectancy that desired outcomes will be attained once instrumental behavior has been carried out (i.e., outcome expectancy, Maddux, 1995). In considering motivational intensity theory, it is useful to think in economic terms, identifying any outcome or set of outcomes contingent on behavior as a commodity that may be purchased at a price and effort as the currency that may be expended to make the purchase. How much people would be willing to expend in a given circumstance should depend on their need for the commodity, the value of the commodity, and the likelihood that payment will result in the commodity being secured. But how much people will expend should depend on the price of the commodity. Low prices draw low currency regardless of people’s upper price limit. Higher prices draw greater currency up to the point that price exceeds either what people are willing to “pay” or what people are capable of paying. The graphs in Fig. 1 illustrate the hypothesized relation between task demand (price) and effort (payment) at three levels of potential motivation (upper price limit). Figure 1A depicts the case in which the upper bound, or price limit, is relatively low. Thus, moving from left to right in the panel, as task difficulty increases so should effort, until the low price limit is reached, at which point effort is withheld because expending effort on the task is no longer worthwhile. Figure 1B depicts the case in which the upper bound is higher. In this case, effort is proportional to difficulty over a broader difficulty range because success is seen as worth more effort. Figure 1C depicts the case in which people exert low effort not because they are unwilling to pay, but rather because they are unable to meet the price (i.e., because people see themselves as lacking the requisite ability). In this case, effort will be proportional to difficulty up to the point that success is viewed as impossible.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
261
Fig. 1. Effort–difficulty relation where potential motivation is low, moderate, and high.
2. Primary Implications Together, Obrist’s effort proposal and Brehm’s motivation intensity ideas have several primary implications. One is that moderately difficult challenges should tend to evoke greater sympathetically mediated CV responses than easy and extremely difficult challenges. This is because effort should be greater where success is viewed as difficult, but possible and worthwhile, than where success is viewed as easy, “too difficult,” or impossible. A second primary implication is that sympathetic CV responses to a possible challenge should be moderated by the importance of success. If importance is high enough to make a possible challenge worthwhile to meet, sympathetic CV effects should be proportional to challenge difficulty, i.e., high if the challenge is substantial and low if the challenge is mild. However, if importance falls short of this level (i.e., if difficulty surpasses potential motivation), sympathetic CV effects should be minimal irrespective of demand. A third primary implication is that success importance should have no effort-related CV impact where a challenge is impossible to meet. If nothing can be done to succeed, then engagement and associated CV responses should invariably be low. Although the preceding implications are straightforward and significant, they apply only to challenges that Brehm and Self (1989) have described as fixed. Brehm and Self conceive of fixed challenges as ones that call for a particular level of performance. They distinguish these from unfixed challenges, which they conceive as ones that allow people to perform as well as they please. The distinction is comparable to that made by some (e.g., Locke & Latham, 1984) between “all-ornone” challenges and “piece-rate” challenges. It is illustrated by two weightlifters, one who must lift a certain weight to earn an amount of money (fixed case) and the other who can lift as much weight as she chooses and will earn in accordance with the amount she lifts (unfixed case). Brehm and Self assume that people presented with an unfixed challenge will strive for the highest possible performance level that they are willing to attain, given the significance of the motive driving behavior. The more important the motive (e.g., the greater the need for money), the higher is
262
REX A. WRIGHT AND LESLIE D. KIRBY
expected to be the performance level for which people strive (e.g., weight targeted), up to the point that people reach their performance peak. Thus, a fourth primary integrative implication is that, where performance is unconstrained, effort-related CV responses should correspond to success importance until a maximum effort has been achieved. We should emphasize that whether a challenge is fixed or unfixed, the proximate determinant of effort is thought to be perceived task difficulty. The difference between fixed and unfixed circumstances can be seen as lying mainly in the manner in which difficulty is established. Where a challenge is fixed, difficulty is established by a force outside performers’ control. Where a challenge is unfixed, it is established by performers themselves when they target a performance goal. 3. Ability Extension Obrist’s proposal and Brehm’s intensity ideas also have a set of secondary implications that derive from an incorporation of ability, or efficacy, perception into the intensity theoretical framework (e.g., Wright, 1998). Ability perception can be incorporated by supposing that people who view themselves as incapable with respect to a given challenge will view that challenge as more difficult than will those who view themselves as capable with respect to the challenge. In other words, holding constant task characteristics (e.g., the nature of a math problem), demand appraisal should be greater in those who feel unable in the relevant performance realm (math) than in those who feel able in the realm. This supposition has intuitive appeal and indeed has been posited by various writers in the past quartercentury (Bandura, 1986; Ford & Brehm, 1987; Kukla, 1972; Locke & Latham, 1994; Meyer, 1987; Sch¨onpflug, 1983; C. Smith & Pope, 1992). Furthermore, the supposition can be conceived in the same economic terms that may be used to clarify the basic intensity model. In these terms, low- and high-ability people have available the same number of currency (effort) units with which to “pay.” However, low-ability people view their units as having less value, or purchase power, than high-ability people do. Because low-ability units are devalued, it takes more of them to meet a given commodity price (i.e., demand level) than it takes high-ability units to meet the price. If the perceived-ability/difficulty suggestion is taken as a given, then it is a simple matter to map separately effort functions for low- and high-ability people onto the Fig. 1 graphs. One merely moves the effort line (holding slope constant) to a relatively high point on the y axis for low-ability people and to a relatively low point on the axis for high-ability people (Fig. 2). This shift reflects the view that in order to succeed at any level of objective difficulty, low-ability people will need to expend more effort than high-ability people. Inspection of the modified effort functions makes apparent five of the secondary (i.e., ability) implications. First, so long as both low- and high-ability people view
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
263
Fig. 2. Effort–difficulty relation at two levels of potential motivation for people with low and high perceived ability.
a challenge as possible and worthwhile to meet, sympathetic CV responses should be greater for the low-ability group than for the high-ability group because the low-ability people should try harder. Figure 2 illustrates this by depicting a higher effort line for low-ability people than for high-ability people at lower difficulty levels. Second, low-ability people should withhold effort at a lower level of objective demand than high-ability people and, hence, should sometimes show less sympathetic CV responsiveness than high-ability people when presented with a challenge. Figure 2 illustrates this by depicting earlier effort “cliffs” (abrupt shifts downward) for low-ability people than for high-ability people as difficulty rises. Third, even people with high perceived ability should withhold effort if a challenge is difficult enough. At and beyond this degree of challenge, sympathetic CV responses should be minimal irrespective of the perceived capacity to perform. Figure 2 illustrates this by depicting later effort cliffs for high-ability people and effort lines that are equivalent for the ability groups beyond the difficulty level at which high-ability people first withhold effort. Fourth, the importance of success should moderate the relation between ability percepts and sympathetic CV responsiveness to a possible challenge. Because success importance should determine the difficulty level at which each ability group withholds effort, it should determine where sympathetic influence will be (1) greater for low- than high-ability people (both groups engaged), (2) lower for low- than high-ability people (low-ability withholding, and high-ability engaged), and (3) low for both groups (both groups withholding). Figure 2 illustrates this by depicting earlier effort cliffs in Fig. 2A than in Fig. 2B. And, fifth, success importance should not influence the perceivedability/CV-response relation where a challenge is impossible to meet. If nothing can be done to succeed, then sympathetic CV responses should be low irrespective of the level of perceived ability. Figure 2 illustrates this by depicting low effort for both ability groups at the upper end of the X axis in Fig. 2C. All of the implications above, of course, pertain to fixed challenge contexts. To date, there has been little consideration of how ability perception should impact
264
REX A. WRIGHT AND LESLIE D. KIRBY
effort where people are free to perform as well as they please. But a reasonable conclusion, based on Brehm and Self’s unfixed reasoning, is that it should have no impact. In theory, unfixed challenges should lead both low- and high-ability groups to strive for the highest possible performance level that they view as worthwhile. Although low-ability people may target levels that are less difficult, objectively, than the ones set by their high-ability counterparts, this does not mean that they should exert less effort in pursuing them, as easier performance targets should seem as difficult to low-ability people as harder performance targets do to highability people. Therefore, a sixth secondary implication is that in unfixed settings low- and high-ability groups should evince sympathetic CV responses that are equivalent and correspondent in magnitude to perceived success importance until greater effort cannot be exerted. 4. Caveats For purpose of clarity, we might note several points about the ability extension. One is that it uses the term “ability” to refer to performance capacity that is not limited in origin or longevity. The extension assumes that ability at a given moment is determined by multiple factors, including genetics, training, and energy level in the relevant performance system (Muraven & Baumeister, 2000; Muraven, Tice, & Baumeister, 1998). Some abilities, such as those inherited physically or acquired through extended learning, are likely to be relatively enduring; others, such as those associated with energy resources, should tend to be short-lived. Another point is that the ability extension assumes that ability perceptions can be accurate or inaccurate, and influenced by secondary factors such as mood (Gendolla, 2000) and cognitive availability (Kahneman, Slovic, & Tversky, 1982). The perceptions are most likely to be accurate when they are based on veridical feedback from multiple performance sessions; they are less likely to be so when they are based on false feedback and veridical feedback from limited performance sessions. The influence of secondary factors on ability perception accuracy is not clear, although there is evidence that less flattering self-assessments commonly may be more correct ones (Taylor & Brown, 1998). A third point is that the extension assumes that groups differing in perceived ability do not differ broadly in their perception of how important it is to satisfy motives, but allows that they may sometimes differ in this perception when an ability conception is associated with a self-definition, or identity (Gollwitzer, Wicklund, & Hilton, 1982; Wicklund & Braun, 1987; Wicklund & Gollwitzer, 1982). Regarding the latter, people who view themselves as capable in a particular realm (e.g., carpentry) may be more likely to adopt an identity associated with the realm (“carpenter”) than people who view themselves as incapable in the realm (Hackett & Betz, 1995). If high-ability people do adopt an ability-related identity, then they should consider it especially important to meet challenges instrumental
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
265
to being “complete” with respect to the identity (i.e., accomplished in the realm, Gendolla, 1998, 1999). A final point is that the ability extension assumes that the implied effects are performance-domain-specific (Bandura, 1986; Ewart, 1995; Maddux, 1995); that is, that the effects will be found only where an ability conception pertains to a current challenge. Where a conception (e.g., regarding math) does not pertain to a current challenge (to sink a putt), it should not predict effort-related CV outcomes.
III. Contrast with the Alternative Views Differences between the integrative perspective on CV response determination and the traditional incentive value theme are plain. The incentive theme suggests that CV responsivity should increase with the value of attractive incentives. The integrative analysis, by contrast, suggests that attractive incentive value determines only potential motivation. Where a challenge is fixed, such value should moderate the relation between difficulty and effort-related CV responses so long as success is possible and have no impact on those responses where success is impossible. Where a challenge is unfixed, such value should mediate effort-related CV responses up to the point that greater effort cannot be exerted. Differences between the integrative perspective and the traditional threat theme are plain as well, but call for elaboration. The theme in this case intimates that CV responses should (1) increase with the degree of harm or loss associated with failure and (2) be inversely proportional to the perception of control. Regarding the first intimation, the integrative analysis conceives of harm and loss not in terms of threat, but rather in terms of aversive incentive value. The greater the harm or loss associated with failure, the greater is the incentive to perform. Given this conception, it follows that harm and loss variables should impact CV responses in the same way that attractive contingencies should. Regarding the second intimation, that CV responses should be inversely related to control, the integrative analysis has no role for the control construct per se, but does have roles for three factors that determine it. One such factor is challenge difficulty. Whereas the traditional threat theme suggests that CV responsivity should increase with difficulty, the integrative analysis suggests that CV responsivity should increase with difficulty to a point and then decrease. A second such factor is outcome expectancy. Whereas the traditional threat theme suggests that CV responsivity should increase as outcome expectancy decreases, the integrative analysis suggests that outcome expectancy should determine potential motivation and thus affect CV responses in the same way that attractive and aversive behavioral contingencies should. The third such factor is personal efficacy. Whereas the traditional threat theme suggests that CV responsivity should increase as efficacy
266
REX A. WRIGHT AND LESLIE D. KIRBY
decreases, the integrative analysis suggests that the relation between CV responsiveness and efficacy should vary, depending on the fixedness and difficulty of a challenge. Differences between the integrative perspective and the BAR view are somewhat difficult to identify because of ambiguities associated with the BAR line of reasoning. Most notably, the BAR model does not tell us how resources (seemingly “apples”) can approach, meet, or exceed the underlying demand components of uncertainty and danger (seemingly “oranges”), which makes it hard to estimate when challenge and threat appraisals will occur. But focusing on the required effort demand component, one BAR implication seems to be that difficulty increases should influence CV responses in a single stepwise fashion. That is, up to a certain difficulty level people should evince a challenge pattern (strong cardiac activity and decreased TPR) and beyond it they should show a threat pattern (weaker cardiac activity and increased TPR—Fig. 3).4 This is different from the integrative suggestion that the relation between cardiac and cardiac-related (e.g., SBP) responses, on the one hand, and difficulty, on the other, will be sawtooth in form and the idea that TPR and DBP (which varies with TPR) may or may not track the sawtooth pattern. Another BAR implication is that ability perception should determine simply the point of conversion from challenge to threat, with low-ability people converting at a lower difficulty level than high-ability people (Fig. 4). This differs from the integrative suggestion that cardiac and cardiac-related responses will sometimes be greater in low- than high-ability people and the idea that TPR and DBP responses may or may not track this pattern.
IV. Evidence for the Primary Implications The full set of primary integrative implications identified earlier were (1) moderate challenges should tend to evoke more pronounced sympathetically mediated CV responses than mild and extreme challenges, (2) sympathetic CV responses to a possible challenge should be moderated by the perceived importance of success, (3) success importance should have no sympathetic CV impact where a challenge 4 In their recent chapter, Blascovich and Mendes (1999) propose that performance situations involving large resource/demand discrepancies may be viewed as “nonevaluative” and fall outside the predictive domain of the BAR model. This suggests that challenge appraisals should occur only after some (undefined) difficulty level has been achieved, but does not alter the implied stepwise character of the difficulty/CV-response function. Blascovich and Mendes also hypothesize that some CV findings may reflect gradations of threat, which intimates that threat appraisals may be amplified as demand increasingly outweighs ability. However, this intimation should be considered cautiously because Blascovich and Mendes do not elaborate on the reasoning underlying their gradation hypothesis and the hypothesis does not comport with the usual BAR model depiction.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
267
Fig. 3. BAR perspective I: Myocardial performance and vascular resistance as a function of task difficulty.
is impossible to meet, and (4) sympathetic CV effects in unfixed circumstances should be proportional to success importance up to the point that a performer’s performance peak has been reached. To date, three of these have been investigated and supported. One (the third) remains to be examined.5
A. SIMPLE DIFFICULTY EFFECT The implication that sympathetic CV responses will be greater under moderatechallenge conditions than under low- and extreme-challenge conditions is supported by a host of investigations that have examined CV responses as a function of task demand. Among the earliest studies to show the expected simple difficulty pattern were experiments by Elliott (1969, Experiment 2) and Obrist et al. (1978; see also Johnson, 1963). Elliott found a nonlinear relation between the difficulty of a tone discrimination task that participants performed for money and the participants’ HR response over trials, with the greatest HR response occurring where the task was moderately difficult. Obrist et al. found similar effects in participants told they could avoid electric shock by performing a (key release) reaction time (RT) task. 5
Investigators have not avoided this implication; they simply have not yet tested it.
268
REX A. WRIGHT AND LESLIE D. KIRBY
Fig. 4. BAR perspective II: Myocardial performance and vascular resistance as a function of task difficulty for people with low and high perceived ability.
For some (mild challenge), the requirement was to attain a relatively slow RT standard; for others (extreme challenge), the requirement was to attain a very fast RT standard; and for still others (moderate challenge), the requirement was to improve response time on successive trials. Results indicated greater CV responsivity—this time for SBP as well as HR (the Elliott study examined only HR)—where the challenge was moderate than where it was mild and where it was extreme. Subsequent difficulty studies have replicated the Elliott and Obrist et al. CV patterns repeatedly, using a range of tasks and incentives (Contrada, Wright, &
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
269
Glass, 1984; Light & Obrist, 1980, 1983; T. Smith, Baldwin, & Christenson, 1990; Steptoe, Moses, Mathews, & Edwards, 1990; Van Schijndel, DeMey, & N¨aring, 1984; Wright, Brehm, & Bushman, 1989; Wright, Brehm, Crutcher, Evans, & Jones, 1990; Wright, Contrada, & Patane, 1986). For example, T. Smith et al. (1990) found nonmonotonic effects on measures of blood pressure (SBP and DBP) and HR responsivity in participants presented the opportunity to earn a monetary incentive by making a speech that was mildly, moderately, or extremely convincing to an audience. Similarly, Wright et al. (1990) found nonmonotonic effects for SBP and HR responsiveness in a procedure that provided participants the chance to avoid a noxious noise by succeeding on an easy (2-trigram), moderately difficult (6-trigram), or extremely difficult (20-trigram) memory task. It is notable that the predicted difficulty pattern has been documented not only while participants were engaged, but also while they anticipated performance (e.g., T. Smith et al., 1990; Wright et al., 1990, 1986). Data indicating that difficulty effects are anticipatory comport with the view that physiological adjustments of this type serve to prepare people for action (Obrist, 1981). The data also limit concerns that observed adjustments may reflect group differences in somatic activity rather than group differences in engagement (Obrist, Webb, Sutterer, & Howard, 1970).
B. MODERATING EFFECT OF IMPORTANCE The implication that success importance should moderate the relation between difficulty and sympathetic CV responsivity is supported by a host of studies as well. Most relevant studies have crossed two-level difficulty manipulations with manipulations of variables that should affect success importance appraisals, including need (Wright, Shaw, & Jones, 1990, Experiment 2), incentive value (Wright et al., 1990, Experiment 1, 1995), and outcome expectancy (Murray, Wright, & Williams, 1993; Wright & Gregorich, 1989; Wright, Williams, & Dill, 1992; see also Gendolla, 1998, 1999; Manuck et al., 1978). However, a pair of studies have examined success importance effects across multiple difficulty levels. 1. Difficulty as a Two-Level Factor Results from studies including two-level difficulty factors have consistently borne out the expected interactional CV response effect. They have shown the interactional pattern most often for SBP, somewhat less so for HR, and on occasion for DBP. An illustration of the pattern is seen in an early study that included a social psychological (prosocial behavior) component (Wright et al., 1990, Experiment 2). Participants were assigned an easy (2-trigram) or moderately difficult (10-trigram) memorization task and told that, if they succeeded, they would earn a donation for a woman with various personal problems (Toi & Batson, 1982).
270
REX A. WRIGHT AND LESLIE D. KIRBY
Half had listened previously to an audiotape describing the woman’s plight with instructions to imagine how she must feel; the rest had listened with instructions to attend to technical aspects of the presentation. Based on research and theory implicating empathy in the activation of altruistic motives (e.g., Batson, 1987), it was assumed that the need to help would be greater in participants who took the victim’s perspective than in participants who took the technical perspective. Consequently, it was predicted that effort-related CV responses would increase with demand among victim perspective (high-need) participants, but not among technical perspective (low-need) participants. Systolic responses assessed just prior to performance were supportive. For those with a high need to help, responsivity was greater in the moderately difficult than easy condition; for those with a low need to help, responsivity was low regardless of the task assigned. A further illustration of the interactional pattern is seen in a later experiment that manipulated success importance by varying, in an avoidance context, outcome expectancy (Wright et al., 1992). It presented participants, by way of computer, trials of an easy or moderately difficult digit-recognition task (Sternberg, 1966) and led them to believe a good performance (90% success rate) would ensure a high (19/20) or low (1/19) chance of avoiding an aversive noise at the end of the work period. The recognition task involved the presentation of a series of digit strings, each of which was followed by a probe asking if a particular digit was in the preceding string. The challenge is made increasingly difficult by increasing the number of digits in each string. Thus, difficulty was manipulated by presenting some participants with three digits on each trial and the rest with seven digits on each trial. As anticipated, SBP and HR responses assessed during performance were greater in the moderately difficult condition than in the easy condition when the probabilistic link between success and noise avoidance was strong (i.e., when outcome expectancy was high), but were low in both difficulty conditions when that link was weak (i.e., when outcome expectancy was low). 2. Importance Influence at Multiple Difficulty Levels Results from studies that included multiple difficulty levels also have supported the interactional prediction. The first study of this type was another that included a social psychological component (Wright, Dill, Geen, & Anderson, 1998). It examined CV effects of incentive value across five difficulty levels, operationalizing incentive value in terms of the potential for social evaluation. Traditional conceptions in social psychology assume that evaluation enhances physiologic activation, either inherently (Zajonc, 1965, 1980) or through a secondary process, such as anxiety production (Cottrell, 1972; Cottrell, Wack, Sekerak, & Rittle, 1968) or distraction (Baron, 1986; Sanders, 1981). The integrative analysis, by contrast, presents the possibilities that CV effects of evaluation may be at least partially mediated by effort and, insofar as they are, manifested only under conditions where
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
271
evaluation causes effort to be increased (for a related view, see Seta & Seta, 1983). From this perspective, a publicity manipulation frequently constitutes a manipulation of incentive value because people who are evaluated often have greater reason to try than people who are not evaluated (Geen, 1991; Geen & Bushman, 1989). To the degree that publicity increases the importance of success, it should increase the peak of what people would be willing to do, but not increase actual effort and related CV responses. What should determine effort and its associated CV responses is the difficulty of the task with which performers are confronted. Where a possible task is easy, effort and CV responsiveness should be low irrespective of the potential for evaluation. On the other hand, where a possible task is difficult, effort requirements should sometimes be viewed as worthwhile only if an audience is present. In such circumstances, effort and CV responsiveness should be relatively great under public conditions, but low under private conditions. Finally, where a task is impossible, effort and CV responsiveness should be low regardless of the potential for evaluation. Participants were presented in random order five versions of the recognition task used by Wright et al. (1992), ranging in difficulty from very low to very high. Specifically, in different work periods, participants were presented with a series of character (as opposed to digit) strings, each followed by probe asking if a particular character was in the preceding string. Difficulty was manipulated by presenting in different work periods strings that were 2, 4, 6, 8, and 10 characters in length. Half of the participants (Audience) were led to believe their responses would be known to a senior graduate student and half (No Audience) were led to believe their responses would be private. Results revealed difficulty × audience interactions for SBP and HR responses measured during the work periods (Fig. 5). Where performance was public, both CV outcomes increased with difficulty to a point and then dropped to a low level. By contrast, where performance was private, the outcomes were relatively low at all difficulty levels. Participants’ DBP responses were similar to their SBP and HR responses; however, analysis of the DBP data revealed only a marginally reliable (p < .06) difficulty main effect. The other study that included multiple difficulty levels examined incentive value influence more directly, operationalizing the value construct in terms of the amount of money participants could win by doing well on character versions of the Wright et al. (1992) memory task (Eubanks, Wright, &Williams, 2001). Like participants in the previous study, participants in this study were presented five task versions. However, for them, character string lengths in different work periods were 1, 4, 7, 10, and 13. Furthermore, the participants were (1) presented with difficulty levels in a fixed order (low to high) and (2) told that they would earn chances to win either a $10 prize by doing well in the work periods or a $100 prize by doing so. Analysis of CV data collected during the work periods showed an incentive value × difficulty interaction for HR. As seen in Fig. 6, HR increased steadily with difficulty among those who could win the more valuable prize, but increased with difficulty only to a
272
REX A. WRIGHT AND LESLIE D. KIRBY
Fig. 5. Systolic pressure response (millimeters of mercury) and heart rate response (beats per minute) as a function of task difficulty for participants whose performance was described as private (No Audience) and public (Audience). Figures created based on data presented by Wright et al. (1998).
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
273
Fig. 6. Heart rate response (beats per minute) as a function of task difficulty and incentive value. From Eubanks, Wright, and Williams (2001).
point for those who could win the less valuable one. Comparisons at each difficulty level revealed that HR was equivalent for the incentive groups in the first four work periods, but greater for the high-incentive group than for the low-incentive group in the final one.
C. WHERE A CHALLENGE IS UNFIXED Support for the fourth primary implication, that CV responses in unfixed circumstances should be a function of success importance until a person’s performance peak has been reached, comes from at least four sources. One is older studies by Belanger and Feldman (1962); Hahn, Stern, and McDonald (1962); and others that showed a positive correspondence between fluid deprivation and HR in rats bar pressing for water. The animals were free to press at any rate; thus, the HR data could reflect different degrees of effort in the different need groups. Another source of support is more recent studies by Fowles and his colleagues that demonstrated a positive relation between the value of a monetary reward and HR in college students performing and RT task with the goal of surpassing as often as possible a preset RT standard (Fowles et al., 1982; Tranel, Fisher, & Fowles, 1982). Participants were free to perform as well as they pleased, at least
274
REX A. WRIGHT AND LESLIE D. KIRBY
within limits; thus, the CV data could, again, reflect different degrees of effort in different incentive conditions. Although findings from the preceding animal and human studies comports with the unfixed implication, their interpretation is uncertain because the studies did not include conditions necessary to identify task engagement, as opposed to need or incentive value per se, as the proximal CV determinant. A third source of support for the implication is recent studies that did include conditions necessary to identify effort as the CV response mediator. Two operationalized success importance in social evaluation terms. In the first (Wright et al., 1995, Experiment 2), participants performed a simple (three-character) version of the Wright et al. (1992) memory task with the understanding that their performance would be known to the experimenter and professor in charge of the project (importance high) or be private (importance low). For half (fixed challenge), instructions were to be 90% accurate. Because the task version was simple, the 90% performance target should have been easy to attain. Therefore, it was expected that these participants would manifest low effort-related CV responses regardless of whether their responses would or would not be observed. For the remaining participants (unfixed challenge), instructions were to be 90% accurate in the shortest time possible. The time aspect of the instructions these participants received allowed them to work at any level (i.e., to respond at any speed); hence, it was expected that those working publicly would manifest stronger effort-related CV responses than those working privately. Systolic responses assessed during work matched the anticipated effort pattern. Whereas publicity potentiated the responses when the challenge was unfixed, it had no impact on the responses when the challenge was fixed; furthermore, SBP responses in the fixed condition were uniformly low. Diastolic responses were in the same pattern for men, but not for women. In the second evaluation study (Wright, Killebrew, & Pimpalapure, 2001, Experiment 1), participants were directed either to press a computer “mouse” key once each time they heard a tone sound at 10-s intervals (challenge fixed) or to press the key repeatedly so long as each tone continued to sound, with the aim of pressing as many times as possible (challenge unfixed). Some (No Observer, NO) were told that their responses would be private; the rest were told their responses would be observed either by an undergraduate (Low Status Observer, LSO) or by a medical school professor (High Status Observer, HSO). Assumptions underlying the study were (1) that the social benefit accrued from success in an evaluator’s presence depends in part on the significance of the evaluator’s opinion (Kamarck, Peterman, & Raynor, 1998; Seta, Crisson, Seta, & Wang, 1989; Seta & Seta, 1992) and (2) that low-status evaluators’ opinions commonly are regarded as less significant than high-status evaluators’ opinions (e.g., Seta & Seta, 1992). Based on these, it was expected effort-related CV responses would be engendered by publicity insofar as the depicted evaluator had status where the challenge was unfixed and be low regardless of the presence or status of the evaluator where the challenge
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
275
Fig. 7. Systolic pressure response (millimeters of mercury) and heart rate response (beats per minute) as a function of the presence and status of an evaluator under fixed- and unfixed-challenge conditions. Error bars indicate standard error. Adapted from Wright et al. (2001).
was fixed. Analysis of CV measures taken during work indicated that SBP and HR responses in the unfixed/HSO condition were greater than those in the remaining conditions (Fig. 7). Diastolic responses tended to be greater in the unfixed/HSO condition than in the other conditions, although an ANOVA in that case yielded only a challenge effect.
276
REX A. WRIGHT AND LESLIE D. KIRBY
A limitation of the unfixed evaluation studies is that they may have confounded challenge fixedness with challenge difficulty. Whereas the studies presented fixed participants with simple tasks, they presented unfixed participants with tasks that called for quick responses. Some unfixed participants may have viewed their challenge as a difficult one that they could meet or not meet. If they did, their responses would be indicative of a fixed-challenge effect rather than the unfixed one of interest. This limitation was addressed in a third study of this type, which manipulated success importance in terms of money paid for success and avoided the difficulty confound by removing the instructional demand to perform at a high level (Wright et al., 2001, Experiment 2). Participants performed a letter-scanning task with the chance to earn 1 or 5 cents for every two “E”s they indentified. As in the second evaluation study, a tone sounded at intervals during the session. For some participants (fixed challenge), instructions were to circle an E each time it did. For others (unfixed challenge), they were to circle Es so long as the tone continued to sound (i.e., so long as the session continued), with no intimation that faster responses were preferred to slower ones. It was expected that CV effort effects would correspond to incentive value where the challenge was unfixed and be low irrespective of value where the challenge was fixed. Work-related responses for SBP and HR fit this expectation. Work-related responses for DBP, on the other hand, were in a crossover pattern compatible neither with the effort predictions nor with any DBP effect observed previously. A final source of support for the fourth primary (i.e., unfixed) implication is an experiment designed to extend the preceding investigation by examining unfixed blood pressure and HR effects across a wide range of incentive value levels (Anthony &Wright, 1998). Participants performed the unfixed version of the letter-circling task used previously. Some were offered no incentive for doing well, whereas other were offered 3, 6, 9, 12, or 15 cents for every two E letters identified. Analysis of CV responses measured during performance indicated that SBP responsivity increased steadily with the value of the available incentive (linear trend, p < .003; see Fig. 8). It was expected originally that effort-related responses would stabilize at one of the higher value levels, and this obviously did not occur. However, in retrospect, the lack of stabilization seems reasonable, as even the highest incentive may well have not been great enough to justify a maximum effort.
D. NOTABLE EXCEPTIONS Although most studies relevant to the primary integrative implications have shown heart or heart-related (SBP) effects compatible with effort expectations, it is important to note that not all have. One study that is exceptional in this regard crossed a difficulty manipulation with a manipulation of need, operationalizing the
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
277
Fig. 8. Systolic pressure response (millimeters of mercury) as a function of incentive value.
latter variable in terms of fluid deprivation (Storey, Wright, & Williams, 1996). Participants who either had (dehydrated) or had not (hydrated) refrained from drinking for an extended period were given the chance to earn a beverage by doing well on a version of the Wright et al. (1992) memory task. For half, trials were easy (three characters) and for half they were moderately difficult (eight characters). Analysis indicated that difficulty and need (deprivation) interacted to determine DBP responsivity, but not SBP or HR responsivity, during work. Whereas DBP responses were greater under moderately difficult than easy conditions for dehydrated participants, they were relatively low under both task conditions for hydrated participants. This study’s results are interpretable in integrative terms if one assumes that engagement produced a TPR increase. The absence of SBP and HR effects might be related to the fact that both HR responsivity (M = +1.55) and blood pressure responsivity (SBP M = +6.85, DBP M = +5.92) were relatively low even where effort was expected to be the greatest (the difficult/dehydrated condition), which suggests that the procedure may have elicited less engagement than most others have. If effort was limited, the effect on rate could have been neutralized by the countervailing effect of parasympathetic activity; furthermore, sympathetic
278
REX A. WRIGHT AND LESLIE D. KIRBY
influence on contractility could have been so slight as to be offset by a contrary nonautonomic influence, such as that of baroreceptors (see discussion above). On the other hand, the highest blood pressure and HR levels in the Wright et al. (1998) experiment were comparable to the highest blood pressure and HR levels in this study. So this interpretation would not explain why SBP and HR effects were observed only in former case. Another exceptional study assessed responses in husbands and wives attempting to persuade verbally their spouse to take a set of issue positions, half (high incentive) with the opportunity to earn chances toward a lottery prize by doing well and the rest (low incentive) without this opportunity (Brown, T. Smith, & Benjamin, 1998). Success was measured in terms of the number of desired positions spouses adopted. Difficulty was operationalized in terms of ratings the participants provided regarding their spouse’s dominance level, on the assumption that persuasion would be viewed as increasingly difficult the more dominant the opposing spouse was thought to be. Analysis indicated that DBP responses increased with perceived spouse dominance where incentive value was high, but were nonmonotonically related to perceived spouse dominance (greatest in the dominance midrange) where incentive value was low. Further analysis indicated a curvilinear relation between perceived dominance and SBP responsivity among wives and a linear relation between these variables among husbands. The investigators interpreted their results in the present theoretical terms, and we agree with them assuming that (1) effort induced a TPR increase and (2) the spouses viewed one another as no more than moderately dominant (which is reasonable, given contemporary trends away from unilateral decision making in marriages). But the interpretation plainly is weakened by the SBP effects, which did not match the DBP effects or agree between spouses. We know of no hemodynamic explanation for the discrepant blood pressure data and suggest that they be viewed guardedly, pending replication.
E. SUMMARY To summarize, considerable empirical support exists for the first, second, and fourth primary integrative implications. The first of the primary implications is the suggestion that moderate challenges should evoke greater CV reactivity than mild and extreme challenges; the second is the suggestion that the perceived importance of success should moderate the relation between CV responsivity and the difficulty of a possible challenge; the fourth is the suggestion that CV responses to an unfixed challenge will be proportional to perceived success importance up to the point that participants reach their performance peak. Evidence for these comes from CV studies that have examined simple difficulty effects, difficulty × success importance interactions, success importance effects under unfixed challenge
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
279
conditions, and fixed challenge × importance interactions. Most commonly, SBP and HR responses have been found to parallel implied effort patterns. However, DBP responses have sometimes been found to do so, and in two cases DBP responses corresponded to implied effort effects when other CV measures did not. Support is not available for the third primary implication, that success importance will have no influence on CV response where success is not possible. But that is because appropriate studies have not yet been conducted rather than because the implication has been examined and disconfirmed.
V. Evidence for Implications Deriving from the Ability Extension The set of ability-related implications identified earlier were (1) sympathetic CV responses should tend to be greater for low- than high-ability people where demand is modest, (2) sympathetic CV responses should sometimes be greater for high- than low-ability people where demand is high, (3) sympathetic CV responses should be low for both ability groups if demand is high enough, (4) success importance should moderate the relation between ability perception and CV responsivity to a possible challenge, (5) success importance should not affect the relation between perceived ability and CV responsivity where a challenge is impossible to meet, and (6) sympathetic CV responses for low- and high-ability groups should be equivalent and proportional to success importance until greater effort cannot be exerted where a challenge is unfixed. To date, three of these have been examined and supported. Three (the fourth, fifth, and sixth) remain to be investigated.6
A. ABILITY EFFECTS AT LOW-MODERATE- AND HIGH-DIFFICULTY LEVELS The first two ability implications are supported by studies that examined perceived ability effects at low, moderate, and high levels of task demand. In the earliest (Wright & Dill, 1993), participants first were encouraged to do their best on a brief letter-scanning task and then provided feedback indicating that they had performed well (87th percentile of previous performances) or poorly (12th percentile of previous performances) in comparison to other students. Shortly after receiving the feedback, participants were presented with a second scanning task— similar, but not identical, to the first—and told they could earn a prize by attaining a high (85th percentile of previous performances) or low (15th percentile of previous performances) performance standard. Analysis of SBP responses assessed 6
Again, investigators have not avoided these implications; rather, they simply have not yet tested them.
280
REX A. WRIGHT AND LESLIE D. KIRBY TABLE I BLOOD PRESSURE AND HEART RATE ELEVATIONS AS A FUNCTION OF PERCEIVED ABILITY AND TASK DEMAND Low ability
Systolic pressure Diastolic pressure Heart rate
High ability
Easy
Difficult
Easy
Difficult
11.2 12.3 11.0
7.7 5.0 4.5
10.6 9.9 5.8
12.7 10.0 12.7
Note. Values from Wright and Dismukes (1995). Pressure measured in millimeters of mercury; heart rate measured in beats per minute.
immediately prior to and during the second performance period revealed the expected feedback × standard interaction. Responses for high-ability (i.e., positive feedback) participants were greater under high- than low-standard conditions; by contrast, those for low-ability (negative feedback) participants were slightly (not reliably) reduced under high- compared to low-standard conditions. Furthermore, whereas responses were slightly greater for low- than high-ability participants when the standard was low, they were lower for low- than high-ability participants when the standard was high. Analysis of DBP responses yielded the same interaction and means in the same crossover pattern. Later studies conceptually replicated and extended this work. One by Wright and Dismukes (1995), for example, reproduced the crossover effect observed by Wright and Dill manipulating ability perceptions experimentally in an avoidance context. In this case, an interaction was obtained not only for SBP and DBP, but also for HR (Table I). Another study, by Wright et al. (1994), reproduced the interactional effect using a procedure that involved natural ability appraisals and three demand levels. Participants who previously had identified themselves as having high math ability (ratings of 7–10 on a 10-point scale, where 0 = very low and 10 = very high) or low math ability (ratings of 0–3 on the same scale) were given the chance to avoid noise by solving math problems described as easy, difficult, or very difficult. Analysis of CV data collected just prior to work showed that the relation between demand and SBP responsivity differed between the ability groups (Fig. 9). High-ability participants tended to evince greater responsiveness as demand increased (linear trend, p < .10), whereas low-ability participants evinced greater responsiveness in the difficult condition than in the low- and very difficult conditions (quadratic trend, p < .05). Comparisons within the difficulty conditions indicated that low-ability participants had stronger responses than high-ability participants where difficulty was high. Low-ability participants had slightly stronger responses than highability participants where difficulty was low, and slightly weaker responses than those participants where difficulty very high, but those effects were not reliable.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
281
Fig. 9. Systolic pressure response (millimeters of mercury) as a function of task difficulty for people with low and high perceived ability. (Adapted from Wright et al., 1994.)
Convergent CV findings from studies (1) examining ability effects in aversive contexts and (2) examining ability as a quasiexperimental factor are important, of course, because they give one faith that the original ability effects are not unique to a particular laboratory setting (see also Clements & Turpin, 1995; Wright & Hodges, 1999). The most recent study to examine ability effects as a function of demand was by Shannon Annis (1999; see also, Annis, Wright, & Williams, 2001). It assessed not only SBP, DBP, and HR, but also mean arterial pressure (MAP) and—via impedance cardiography (Sherwood, Allen, Fahrenberg, Kelsey, Lovallo, & van Doornen, 1990)— an array of additional CV parameters, including heart preejection period (PEP), stroke volume (SV), cardiac output (CO), and TPR. Preejection period is the most commonly used index of myocardial contractility and hence should in theory be the best index of sympathetic nervous system influence on the heart. Stroke volume refers to blood ejected in a heart cycle. It should be a function of (1) the amount of blood in the left ventricle (the heart chamber that ejects blood into systemic circulation) prior to contraction and (2) the force with which the blood is expelled on a heartbeat. Where a sympathetic discharge causes HR acceleration, the discharge might or might not increase SV because a faster HR is associated with reduced ventricular filling time and therefore blood volume
282
REX A. WRIGHT AND LESLIE D. KIRBY
(Obrist, 1981; Papillo & Shapiro, 1990). However, where a sympathetic discharge has no HR impact, it should increase SV because it should potentiate contractile force while leaving unaffected blood volume in the relevant ventricle. Cardiac output and MAP refer to blood pumped over time (SV × HR) and the average arterial pressure maintained over a heart cycle [theoretically, 1/3 (SBP − DBP) + DBP]. Although these could increase with sympathetic activity, they should not necessarily. By definition, CO can be influenced by sympathetic activity only insofar as SV and HR are so influenced, and a positive effect on one of these is not outweighed by a contrary effect on the other, Similarly, MAP should be influenced by sympathetic activity only insofar as its underlying parameters are so influenced, and the effect of an increase in one of the underlying parameters is not offset by the effect of a decrease in the other. The Annis study began as an attempt to create low and moderate ability perceptions and examine their impact on CV responses across six levels of objective task demand, ranging from very low to very high, with the expectation that the highest demand levels would be sufficiently challenging to cause both ability groups to disengage.7 But analysis of the participants’ perceived ability ratings indicated that although ability appraisals of the low- and moderate-ability groups differed reliably (low-ability M = 5.13, moderate-abilityM = 6.18; 0 = very low and 10 = very high), the appraisals overall were higher than anticipated, calling into question the original disengagement assumption as it applied to the higher ability group. Furthermore, although difficulty ratings distinguished statistically the demand levels assigned, differences between adjacent difficulty means were trivial in magnitude and thus likely to yield unstable effects. With the latter consideration in mind, the investigators analyzed the CV data collapsing across the lower two, middle two, and higher two demand levels, aiming to create three more powerfully defined categories of low demand, moderate demand, and high demand. With the former consideration in mind, they abandoned their expectation that the moderate-ability group would disengage where demand was high and reasoned instead that only the low-ability group would do so. Participants first performed a character version of the Wright et al. (1992) memory task and received feedback indicating that they had low ability (score of 22 of 100) or moderate ability (score of 52 of 100) with regard to it. Subsequently, they were provided in a series of additional work periods the opportunity to earn chances to win a prize by attaining scores of 24, 39, 54, 69, 84, and 99. Collapsing across demand levels as described above, the design was a 2 (perceived ability: low’s moderate) × 3 (performance standard: low’s moderate’s high) mixed factorial. In view of the perceived ability ratings, it was hypothesized (1) that effort-related CV 7
The study also included initially an attempt to manipulate success importance by offering some participants a valuable incentive and others a modest one. Unfortunately, ability ratings indicated that the ability manipulation was not effective for participants offered the more valuable incentive. Therefore, the data from the high-value-incentive conditions were excluded from primary analyses.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
283
responses should be linearly related to demand for the moderate ability participants, but nonmonotonically related to demand (i.e., higher where demand was moderate) for the low-ability participants and (2) that effort-related CV responses should be greater for the low- than moderate-ability group where demand was low and moderate, but greater for the moderate- than low-ability group where demand was high. Figure 10 shows that the PEP and SV values comported, or tended to comport, with predictions (contrast p < .05 for PEP; contrast p = .10 for SV). For those who received moderate-ability feedback, responsivity increased with the standard assigned; for those who received low-ability feedback, responsivity first increased and then decreased as the standard increased. Further, responses were, or tended to be, greater for the low-ability group where the standard was low and moderate, but greater for the moderate-ability group where the standard was high. Participants’ DBP and MAP responses were configured similarly to the PEP and SV responses (interaction ps = .05 and .075, respectively); however, surprisingly, analysis of the SBP data produced only an ability main effect, reflecting greater responses among low-ability participants than among moderate-ability participants. For reasons that remain, unclear, the SBP data indicated no reversal of the perceived-ability/CVresponse relation where demand was high. As would be expected, given the SV results, HR was low and did not differ across conditions.
B. WHERE DEMAND IS EXTREME The third ability implication, that sympathetic CV responsiveness should be low even for high-ability people if difficulty is high enough, is supported by a study that examined the possibility that the integrative analysis might shed light on reported sex differences in CV response to challenge (Polefrone & Manuck, 1987; Saab, 1989; Stoney, Davis, & Matthews, 1987). The most common sex effect is greater SBP elevations in men than women. However, some have found a reverse CV response pattern, others have obtained CV sex effects that were task-specific, and still others have observed no sex effects at all (Brown & T. Smith, 1992; Collins & Frankenhauser, 1978; Liberson & Liberson, 1975; Light et al., 1993; T. Smith, Limon, Gallo, & Ngu, 1996). Various investigators have speculated that men may try harder on “masculine” tasks, women may try harder on “feminine” tasks, and the sexes may try equally hard on tasks that are gender-neutral (Lash, Eisler, & Schulman, 1990; Lash, Gillespie, Eisler, & Southard, 1991; Matthews, Davis, Stoney, Owens, & Caggiula, 1991; Stoney, Matthews, McDonald, & Johnson, 1988; T. Smith et al., 1996). Conceptions of “masculine” and “feminine” have not been consistent. But the expressions commonly have been defined in terms of felt competence, with masculine tasks being those in regard to which men feel especially capable and feminine
284
REX A. WRIGHT AND LESLIE D. KIRBY
Fig. 10. Preejection period response (seconds) and stroke volume response (milliliters) as a function of task difficulty for participants who received low- and moderate-ability feedback. Note that lower PEP = greater contractility. Adapted from Annis (1999).
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
285
tasks being those in regard to which women feel especially capable. Although this “gender-relevance” hypothesis has been examined repeatedly, it has received little support (Davis & Matthews, 1996; Matthews et al., 1991), perhaps in part because it assumes that effort and associated CV responses are proportional to perceived ability in a performance realm. The integrative reasoning, of course, intimates something different. It suggests that when those of a certain sex feel especially capable with respect to a task (challenge), they may try harder and manifest greater CV responsivity than those of the other sex only at relatively high levels of task demand. At lower demand levels—where people in both sex groups view success as possible and worthwhile—the relation between sex, on the one hand, and effort and CV responsivity, on the other, may be reversed. Furthermore, if demand is high enough, there should be no sex difference in CV response because neither women nor men should exert much effort. The integrative version of the gender relevance hypothesis was examined initially in a study (Wright, Murray, Storey, & Williams, 1997, Experiment 1; see also Murray, 1994) that began with instructions indicating that the Wright et al. (1992) computer memory task was something on which men (masculine conditions) or women (feminine conditions) excel. Differential ability was conveyed not only by way of the instruction text, but also by way of scores described as typical for each sex. In the masculine conditions, men were depicted as typically scoring 84 of a possible 100 points, and women were depicted as typically scoring 43 of 100 points. In the feminine conditions, the scores for men and women were reversed. Once this information was clear, the study took CV measures while participants performed a version of the task with the chance to avoid noise by attaining a low (score of 30) or high (score of 90) performance standard. The prediction was that participant sex would combine with task type and standard to determine sympathetic CV responses, with relevant elevations forming a double crossover pattern across task type levels. Where task depiction was masculine, women were expected to show more responsiveness than men under low-standard conditions, but less responsiveness than men under high-standard conditions; where the depiction was feminine, women were expected to show less responsiveness than men under low-standard conditions, but more responsiveness than men under high-standard conditions. Analysis of SBP change scores assessed during work revealed effects broadly, though not completely, consistent with expectations (Fig. 11). Where depiction was masculine, women manifested greater responsiveness than men under lowstandard conditions, but slightly (nonreliably) less responsiveness than men under high-standard conditions. Where depiction was feminine, women evinced less responsivity than men under low-standard conditions, but equivalent responsivity to men under high-standard conditions. Within-sex comparisons indicated that responses for women were greater under low- than high-standard conditions where depiction was masculine, but in the reverse pattern where depiction was feminine.
286
REX A. WRIGHT AND LESLIE D. KIRBY
Fig. 11. Systolic pressure response (millimeters of mercury) for women and men under low- and high-difficulty conditions where the challenge was identified as masculine and feminine. Adapted from Wright et al. (1997), Experiment 1.
Responses for men did not differ as a function of standard where depiction was masculine, but were greater under low- than high-standard conditions where depiction was feminine.8 As a follow-up to this first study, a second study was carried out (Wright et al., 1997, Experiment 2; see also Storey, 1996), in part, to compare CV sex effects obtained under high-standard conditions to those obtained where a performance standard is extreme. As before, the expectation was that sex would combine with task type and standard to determine CV engagement effects. However, in this case, the prediction was for a pair of three-versus-one patterns rather than two crossover patterns. Where depiction was masculine, effort-related effects were expected to be greater for men than women under high-standard conditions, but low for both sexes under extreme-standard conditions. Where task depiction was feminine, effort-related CV effects were expected to be greater for women than men under high-standard conditions, but, again, low for both sexes under extreme-standard conditions. Predictions assumed that the extreme standard would be viewed as excessively difficult by both sexes. 8 All CV analyses reported by Wright et al. (1997) excluded change scores that exceeded cell means by 2 standard deviations. The original analysis of these SBP data (Murray, 1994) did not exclude the extreme scores, but yielded essentially the same results (see Wright, 1996, 1998).
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
287
Procedures were highly similar to those used originally. Major changes were threefold. First, participants worked to earn a prize rather than to avoid noise. Second, typical scores for the more and less capable sex were depicted as 64 and 23, respectively, instead of 84 and 43, respectively. Third, the required performance score in the high-standard condition was 75, instead of 90, and that in an extreme-standard condition was 95. It should be noted that even though this study’s high standard was lower than the high standard used originally, the discrepancy between it and the norm for the more capable sex was somewhat greater than the corresponding discrepancy in the original study. The discrepancy was expanded to make the high standard somewhat more challenging and possibly strengthen sex differences in the high-standard conditions. It also should be noted that, even though the extreme standard in this study was close to the high standard used previously, the discrepancy between it and the norm for the more capable sex was substantially greater than the discrepancy between the previous high standard and more capable sex norm. The design was a 2 (female vs male) × 2 (feminine vs masculine) × 2 (high standard vs extreme standard) factorial. Like before, analysis of the SBP data generally confirmed expectations (Fig. 12). An a priori contrast comparing responses of masculine/high-standard men and feminine/high-standard women with responses of other participants proved
Fig. 12. Systolic pressure response (millimeters of mercury) for women and men under high- and extreme-difficulty conditions where the challenge was identified as masculine and feminine. Adapted from Wright et al. (1997), Experiment 2.
288
REX A. WRIGHT AND LESLIE D. KIRBY
reliable. Follow-up contrasts indicated that responses were greater for masculine/high-standard men than for all other masculine task participants and greater for feminine/high-standard women than for all other feminine task participants. The individual female-to-male comparison under high-standard conditions was reliable where the task depiction was masculine, but not where the task depiction was feminine. Specifically consistent with the idea that CV responsiveness will be low irrespective of ability if difficulty is high enough, neither of the female-to-male comparisons was reliable under extreme standard conditions.
C. CARDIOVASCULAR EFFECTS OF MOOD Two recent mood experiments did not address directly implications of the ability extension, but provide data that are highly related to them if one assumes (as we do) that moods can influence appraisals of task demand, either by affecting assessments of personal resources (i.e., ability) or by affecting assessments of task characteristics. Demand appraisals should tend to be greater in negative moods than in positive moods, and where they are, mood/CV-response relations should be the same as the perceived-ability/CV-response relations we have discussed. One of the experiments (Gendolla, Abele, & Kr¨usken, 2001) made physiological assessments while participants went through one of two mood induction procedures and then again later while participants performed a moderately difficult letter-cancellation task. The induction techniques involved listening to (happy or sad) music and recalling (happy or sad) past life events. Just prior to performing, participants rated the difficulty of the task and the amount of effort it would require. As expected, mood had no impact on blood pressure or HR responsiveness during the induction period. On the other hand, mood did impact SBP responses during performance. Consistent with the investigators’ presumption that negative mood participants would view the task as especially difficult, but not excessively so, elevations were greater in the negative mood conditions (music M = 15.23, recollection M = 16.63) than in the positive mood conditions (music M = 5.08, recollection M = 7.84). The difficulty construal of the performance SBP data was supported by participants’ anticipatory difficulty and effort ratings, which proved to be higher for the negative mood group. The other experiment (Gendolla & Kr¨usken, 2001) made CV assessments during a single (music) mood induction period and a period in which participants performed a version of the letter-cancellation task intended to be easy or difficult. As before, results indicated no mood effects during the induction period. By contrast, during work there was a difficulty × mood interaction pattern for SBP and DBP responsivity. Where difficulty was low, blood pressure values were greater for the negative mood group than for the positive mood group; where difficulty was high, blood pressure elevations were in a reverse configuration (Table II). Findings in
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
289
TABLE II BLOOD PRESSURE ELEVATIONS DURING PERFORMANCE AS A FUNCTION OF MOOD AND TASK DEMAND Negative mood
Systolic pressure Diastolic pressure
Positive mood
Easy
Difficult
Easy
Difficult
13.2 8.9
4.9 3.7
5.5 5.5
10.2 8.0
Note. Values from Gendolla and Kr¨usken (2001). Pressure measured in millimeters of mercury.
the easy conditions were interpreted as conceptually replicating effects from the original mood study, whereas those in the difficult condition were taken to indicate that the difficult challenge exceeded what negative mood participants—but not positive mood participants—were willing or able to do. Like the original mood experiment, this study included a measure of perceived task demand. Unlike the original study, though, it produced no mood effect on the measure. D. SUMMARY In sum, the first—but not final—three secondary implications have been supported empirically. The first three secondary implications are (1) sympathetic CV responses should tend to be greater for low- than high-ability people where demand is modest, (2) such CV responses should sometimes be greater for high- than low-ability people where demand is high, and (3) such CV responses should be low for both ability groups if demand is high enough. Direct support for these comes from CV studies that have examined difficulty × perceived ability interactions, including the sex difference studies, which were conceptualized in perceived ability terms. Indirect support comes from CV studies that have examined demand × mood interactions. Most often, SBP has been found to track predicted effort patterns, although other CV parameters have been found to do so on occasion and in one study (Annis et al., 2001) interactional patterns were obtained for several CV outcomes, but not SBP. The final three secondary implications are (4) success importance should moderate the relation between ability perception and sympathetic CV responsivity to a possible challenge, (5) importance should not affect the relation between perceived ability and such responsivity where a challenge is impossible to meet, and (6) sympathetic CV responses for low- and high-ability groups should be equivalent and proportional to importance until greater effort cannot be exerted where a challenge is unfixed. Although these have not been supported, it is because appropriate studies have not yet been completed rather than because the implications have been examined and disconfirmed.
290
REX A. WRIGHT AND LESLIE D. KIRBY
VI. Observations When considered as a group, the studies reviewed in the preceding section paint a picture of CV response that is remarkably consistent. Where experimental or quasiexperiment groups have been, or would be, expected to exert different degrees of effort, they generally have displayed different degrees of CV responsiveness, usually in terms indicative of increased cardiac activity. Where these groups have been, or would be, expected to exert equivalent effort, they generally have displayed equivalent CV responses. To be sure, the base of evidence relevant to the integrative analysis is not complete, as four implications (one primary, three ability-related) remain to be examined. Furthermore, anticipated pairwise effects have not always proved to be statistically reliable and few studies have yielded anomalous blood pressure results. But to focus on these aspects of the data set would be to miss the forest for the trees. In broad view, the studies make an impressive case for the integrative line of thought.
A. TRADITIONAL THEMES AND THE BAR HYPOTHESIS Although the studies reviewed argue well for the integrative perspective on CV response determination, they call into question the traditional themes identified concerning the roles that incentive value and threat play in producing tonic CV responses. Regarding the incentive value theme, fixed-challenge studies have demonstrated that variables associated with incentive value do not influence CV responses directly, but rather moderate the relation between difficulty and those responses where success is possible. The moderating effect of incentive value is best illustrated in the Eubanks, Wright, and Williams (2001) experiment, which showed HR differences between participants offered the chance to win $100 and participants offered the chance to win $10 only where difficulty was very high. It also is illustrated in the Wright et al. (1998) experiment, which showed greater SBP and HR responses among evaluated participants (incentive value high) than among nonevaluated participants (incentive value low) only where difficulty was in a midrange. Unfixed challenge studies have complemented the fixed challenge ones by demonstrating that CV responses correspond to incentive value where performance is unconstrained, but do not correspond to incentive value where there is a call for low effort. The clear suggestions are that CV responsiveness is mediated by engagement and that engagement may or may not vary with value. Regarding the threat theme, fixed challenge studies have shown that threatrelated factors either simply do not predict CV responses or moderate the relation between those responses and task difficulty in a manner consistent with the
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
291
integrative formulation. One example of a threat factor being nonpredictive is seen in the Obrist et al. (1978) study that manipulated the difficulty of avoiding electric shock. The traditional threat theme suggests that CV responsiveness should have increased steadily with task demand because participants presented more difficult tasks should have experienced less control and, thus, more threat than participants presented less difficult tasks. But results indicated a nonmonotonic relation between difficulty, on the one hand and SBP and HR responsiveness, on the other, with the strongest CV responses occurring where demand was moderate. Further examples of threat nonpredictiveness are seen in the studies that crossed ability factors with manipulations of avoidant task demand (e.g., Wright et al., 1994; Wright & Dismukes, 1995). An example of a threat factor serving as a moderator is seen in the Wright et al. (1992) experiment that provided participants a low or high chance of avoiding noise by succeeding on an easy or difficult memory task. The traditional threat theme in this case intimates that the results should have shown two main effects because subjective threat should have been (1) greater where difficulty was high than where difficulty was low and (2) greater where the outcome expectancy was low than where the outcome expectancy was high. In fact, of course, results indicated an interactional pattern, with SBP and HR responsiveness increasing with demand where outcome expectancy was high and being low regardless of demand where outcome expectancy was low. The studies reviewed also raise questions about the BAR hypothesis (e.g., Blascovich & Tomaka, 1996). Hypothesis implications discussed earlier were (1) difficulty increases should influence CV responses in a single stepwise fashion, with lower difficulty levels being associated with strong myocardial responses and low TPR and high difficulty levels being associated with weaker myocardial responses and higher TPR; and (2) ability perception should determine the point of conversion from challenge to threat, with low-ability people converting at a lower demand level than high-ability people. The first of these does not accord with the nonmonotonic relation that has been observed repeatedly between difficulty, on the one hand, and SBP and HR, on the other. It also does not accord with the fact that difficulty effects on DBP have been inconsistent and comparable in some cases to those of difficulty on SBP and HR (e.g., T. Smith et al., 1990). The second BAR implication discussed does not comport with the data indicating that lower ability people sometimes manifest greater cardiac (e.g., HR, PEP) and cardiac-related (e.g., SBP) responsiveness than higher ability people or the data indicating that ability effects on DBP have been inconsistent and matched in some cases by those of ability on cardiac and cardiac-related measures. In light of the traditional and BAR views’ limited ability to account for findings from the studies reviewed, it is worth considering whether some investigations previously construed in terms of those views might not be better understood in the present integrative terms. Cases in point are studies interpreted as documenting CV
292
REX A. WRIGHT AND LESLIE D. KIRBY
effects of challenge and threat appraisals (e.g., Tomaka et al., 1997, 1999). These typically have measured CV responses in participants performing a moderately difficult task (e.g., backwards subtraction) and classified participants as “challenged” and “threatened” based on demand and ability ratings the participants made pertaining to the task. Ability usually has been assessed straightforwardly and demand has been assessed as an index made up of ratings relevant to that construct. The main result across studies has been stronger myocardial responses and diminished TPR responses in participants whose ability ratings exceeded their demand index ratings relative to participants for whom the reverse was true. Notably, the studies have not consistently shown the predicted TPR decrease relative to baseline in the former group and the predicted TPR increase relative to baseline in the latter. In one instance, challenged participants showed less of a TPR increase than did Threatened participants (Tomaka et al., 1999). An integrative interpretation of the broad pattern might be that Challenged participants have tended to perceive success as difficult, but possible and worthwhile, whereas Threatened participants have tended to perceive success as excessively difficult. Such perceptions would be expected to lead to lower effort in the latter group than in the former, which would account for the observed heart-related effects and not be incompatible with the TPR effects, since the relation between sympathetic activity and TPR is variable. Other cases in point are well-known studies by Bandura and his colleagues that have used a phobic fear model to investigate efficacy effects on “stress” responses (Bandura, Reese, & Adams, 1982; Bandura, Taylor, Williams, Mefford, & Barchas, 1985; Wiedenfeld et al., 1990). The usual procedure in these has been to first ask phobic participants to indicate how confident they are that they can perform various phobia-relevant tasks and then assess the participants’ physiological responses under conditions where the participants have the chance to perform tasks in regard to which they have indicated different degrees of confidence. Findings have not been entirely consistent. But overall patterns suggest an inverse relation between confidence (efficacy) and physiological responsiveness so long as the tasks presented are within the range of what participants are willing to attempt. Initial physiological assessments usually are followed by efficacy training, which tends to reduce responsiveness to the tasks first attempted. Data from the Bandura studies are widely taken to indicate that control buffers stress in aversive contexts (e.g., O’Leary & Brown, 1990). However, an integrative interpretation is at least as plausible. That is, prior to efficacy training, study participants should have found it more difficult to perform tasks in regard to which they had less confidence than tasks in regard to which they had more confidence. If they did, it follows that they should have tried harder on the tasks in regard to which they had less confidence so long as success seemed possible and worthwhile. After efficacy training, participants should have found it easier to perform the tasks that previously were hard and, thus, reduced their effort.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
293
B. FIXED VERSUS UNFIXED CHALLENGES The confirmation of predictions pertaining to both fixed and unfixed challenges is noteworthy for at least two reasons. For one, it suggests that the integrative CV-response framework can confidently be applied across a range of performance circumstances. In addition, it documents two ways in which incentive value and other importance factors can influence effort-related CV responses. One way is by determining whether challenges will be met (fixed cases); the other is by determining what challenges will be set (unfixed cases). Although unfixed challenges have received less empirical attention to date than have fixed challenges, this should not be taken to mean that they are rare. Unfixed challenges would appear to be at least as common as fixed ones, with social exemplars ranging from a persuasive speaking engagement to a boxing match and nonsocial ones ranging from a dash home for lunch to a history exam. The challenges generally should be easy to identify; however, there is one unfixed variation that is nonobvious and worth highlighting. It is where people do not have a clear grasp of what is called for by a challenge (commodity “price”). If people have no idea what they must do, then they are free to perform at any level. Accordingly, they should aim for increasingly higher performance levels, the greater the importance of success, up to the point that they believe they can perform no better. If people have some idea of what they must do, but are not sure, then they are free to perform at any level within some range of performance levels (the ones associated with uncertainty). And accordingly, they should aim for higher performance levels, the greater the importance of success, until they are sure they will succeed. Tentative support for this argument comes from a study in which participants were offered either a record album (importance high) or a pen (importance low) for performing a memorization task (Wright, Heaton, & Bushman, reported by Wright & Brehm, 1989). Half (Fixed) learned that the version of the task they had been assigned was exceptionally easy, 1 trigram. The rest (Unknown) learned that they could be assigned 1, 3, 5, 7, or 15 trigrams, but that they would not find out which task they would receive until the work period began. Analysis of CV measures taken just prior to work showed greater SBP elevations in the Unknown/HighImportance condition than in both Fixed conditions. Systolic elevations in the Unknown/Low-Importance condition were intermediate between these extremes and significantly different from neither. Appreciation of the conceptual link between performance situations involving uncertainty and the simple unfixed weight-lifting situation described earlier eases considerably the task of predicting CV responses where difficulty is not apparent. Appreciation also offers an explanation for why people sometimes (1) display reduced CV responses once they become familiar with a task (e.g., Elliott, 1965), (2) vary in their CV response to a fixed challenge (T. Smith et al., 1989; Turner, Sherwood, & Light, 1992), and (3) CV elevate to degrees that exceed metabolic
294
REX A. WRIGHT AND LESLIE D. KIRBY
demand (Herd, 1983; Obrist, 1981). It may be that familiarity sometimes reduces CV responses because people engage to the extent that they are willing to until they determine what is required. It task requirements are lower than people’s upper performance limit, then responses under familiar conditions should be lower than those under unfamiliar conditions. It may be that people sometimes vary in their response to a fixed challenge for the same reason, because they engage in proportion to success importance until the difficulty of the challenge is clear. If two individuals differ in how significant they perceive success to be, it follows that their engagement levels also should differ so long as task requirements remain uncertain. Finally, it may be that people sometimes respond excessively, from a metabolic standpoint, because effort in uncertain circumstances is correspondent to potential motivation, not what is actually called for.
C. UNANSWERED QUESTIONS An interesting question that the studies reviewed do not answer is whether people have to be cognizant of their effort levels for CV effects such as these to be in evidence. We suspect that they do not, that even relatively complex effort modulation can occur outside of conscious awareness, based on implicit appraisals of fixedness, difficulty, and benefit (Lewicka, 1986; Uleman & Bargh, 1989; for a similar idea, see Blascovich & Mendes, 1999). Engagement studies that have included subjective effort measures have in fact found only a limited and inconsistent relation between those measures and CV responses, which is compatible with a nonconscious processing view. But evidence of this sort is by no means definitive, and a worthy goal for future study would be to better understand the role that mindfulness plays in determining engagement-related CV outcomes. Another question that the studies do not answer pertains specifically to the studies that have manipulated incentive value in social evaluation terms. It is whether people always consider it especially important to do well in the presence of significant others. Although we and other investigators (e.g., Baumeister, 1982; Seta & Seta, 1983, 1992) have contended that significant others generally promote the willingness to try, it is unlikely that they invariably do. In some performance circumstances, meaningful evaluators may be seen as having no ability to reward or unqualified regard for the performer. In others, they may be seen as having a negative view of success or outcomes that accrue from it. In still others, they may be seen as potential comforters in the event of failure. Where evaluators are viewed as having no reward ability or unqualified performer regard, their presence could have no impact on the perceived importance of success. Where evaluators are viewed as disapproving of success or success-related outcomes, or as potential comforters, their presence could cause success to seem less important than it otherwise would (for possible examples, see Gerin, Pieper, Levy, & Pickering, 1992; Kamarck,
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
295
Manuck & Jennings, 1990; Kamarck et al., 1998; Gallo, T. Smith, & Kirchner, 2000). Social influence on effort-related CV responses plainly is not simple, and a second worthy goal for future research would be to determine boundary conditions of the present evaluation effects. Yet another unanswered question is whether CV effort effects in unfixed circumstances depend on the relation that exists between performance and benefit. In all of the unfixed challenge studies reviewed, Unfixed participants had the chance to gain more the better they did. In other words, the Unfixed participants were presented with a challenge that had the characteristic of a positive relation between performance and benefit. This characteristic fits well with Brehm and Self’s (1989) unfixed reasoning and the unfixed examples they discuss. However, it is not a logically necessary characteristic of unfixed challenges as Brehm and Self define them. For example, one can imagine situations in which performance is unconstrained and benefit decreases as performance rises or increases with performance to a point and then decreases. Even where benefit rises with performance, the slope and shape of the performance-to-benefit function could vary. It is reasonable to suppose that the performance/benefit relation in unfixed circumstances will influence effort in them, and, thus, a third worthy goal for future investigation would be to determine if and how this is true.
D. CARDIOVASCULAR INFLUENCE OF SOCIAL PSYCHOLOGICAL VARIABLES Although there is much work yet to be done related to the integrative CV response analysis, it is not too early to consider what the analysis might tell us about the CV influence of some specific social psychological variables. Social variables already discussed in terms of the analysis are empathy, social evaluation, and sex. We described in a previous section how people may be especially willing to promote a victim’s welfare when experiencing high empathy. If they are, the implication is that they have greater potential for experiencing CV arousal under high- than lowempathy conditions. However, the realization of this potential should depend on what can and must be done to help. We also made the case previously that people may generally place greater value on success when their performance is public as compared to private. Where evaluation enhances success importance, potential motivation should be higher than it otherwise would be and CV responses should be predictable based on the fixedness and difficulty of the performance challenge at hand. Where evaluation does not enhance success importance, it should have no effort-related CV impact. Regarding sex, we pointed out previously that the sexes may sometimes differ in their felt capacity to perform. Where sex is correlated with ability perception, it should be related to CV responsiveness in different ways under different task
296
REX A. WRIGHT AND LESLIE D. KIRBY
conditions. The sexes also could sometimes differ in (1) their need for an available incentive, (2) their appraisal of an incentive’s value, and (3) their judgment of how likely it is that they will attain an incentive if they succeed (i.e., their outcome expectancy). Needs, value appraisals, and outcome expectancies, of course, are thought to determine potential motivation. Thus, where the sexes differ on these dimensions, CV sex effects should have a different character than CV sex effects associated with ability perception. Where the sexes do not differ on these dimensions or in terms of ability perception, they should not differ in effort-related CV response. Other social variables that might be analyzed in these integrative terms are race (Black versus White; Anderson, Lane, Taguchi, Williams, & Houseworth, 1989; Falkner & Kushner, 1989; James, Hartnett, & Kalsbeck, 1983), anger (T. Smith & Christensen, 1992), and optimism (Armor & Taylor, 1998; Scheier & Carver, 1985). Just as men and women may have experiences that lead them to feel especially competent in different behavioral realms, so may Blacks and Whites. And just as the men and women may sometimes differ in their need to do well, their perception of goal value, and their expectation of how likely it is that success will lead to motive satisfaction, so may Blacks and Whites. Where race is related to the integrative parameters, particular CV race effects should be in evidence; where race is not related to the parameters, effort-related CV race effects should not be found.9 An integrative anger analysis might begin with the observation that anger derives from the perception of injustice and instigates action oriented toward restoring a just state (Averill, 1979; Brehm, 1999: Weiner, 1995). Theoretically, the greater the perceived injustice, the greater should be the potential motivation to carry out restorative behavior. However, the amount of effort and CV responsiveness engendered should depend on what can and must be done to this end. If the restorative challenge is fixed, effort and CV responsivity should be proportional to its difficulty so long as success is worthwhile, and low if success is impossible or unwarranted. If the challenge is unfixed, effort and CV responsivity should be proportional to the degree of injustice up to the point that further effort cannot be generated. This view is markedly different, of course, from the familiar notion that anger is proportional to injustice and invariably associated with high autonomic arousal (e.g., Izard, 1993). Like the variables sex and race, optimism may be related to more than one integrative analytic parameter. At minimum, people with optimistic outlooks should tend to (1) view challenges as easier, and (2) have more favorable expectancies regarding the likelihood that success will yield desired outcomes, than should 9 Complicating the sex and race analyses is recent research by Steele and his colleagues on disidentification and stereotype threat (Steel & Aronson, 1995). The idea behind the research is that people sometimes withhold effort instead of trying and failing, thereby confirming a negative stereotype. This suggests that where difficulty is high, those who are purported to be incapable may sometimes view success as less important than those who are purported to be capable; furthermore, where difficulty is modest, the reverse could sometimes be true. Thus, where negative sex and race stereotypes are salient, potential motivation may vary between the sexes and races at different levels of task demand.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
297
people with pessimistic outlooks. In addition, optimists may sometimes appraise incentives as more valuable than pessimists. Where optimism impacts demand appraisals, its relation to CV responsiveness should be the same as that between ability and CV responsiveness (Fig. 2). Where optimism impacts outcome expectancies and, possibly, value appraisals, its relation to CV responsiveness should be the same as that between potential motivation and CV responsiveness (Fig. 1).
E. DOCUMENTING SOCIAL EFFORT ASSUMPTIONS It also is not too early to consider what social psychological studies could benefit from the inclusion of CV measures in their protocols to document underlying effort assumptions. Certainly included in this group would be studies concerned with the social loafing and helplessness topics mentioned at the outset of this chapter. Historically, social loafing and helplessness studies have taken poor performance to be indicative of low effort (e.g., the hypothesized motivational deficit in the helplessness case). Poor performance could indicate low effort, but it also could indicate effort that is great enough to interfere with the performer’s ability to succeed (Ford & Neal, 1985.) Thus, by including CV measures in their research, loafing and helplessness investigators could greatly improve understanding of the performance outcomes they generate. Also included in the group of studies that could benefit from the inclusion of CV measures in their protocols would be studies concerned with (1) persuasive message processing (Chaiken, Giner-Sorolla, & Chen, 1996; Petty & D. T. Wegner, 1998), (2) person perception (Gilbert, 1998; Gilbert & Malone, 1995; Hilton & Darley, 1991; Neuberg, 1996), (3) thought suppression (Ansfield & D. M. Wegner, 1996; D. M. Wegner, 1992), and (4) the self-regulation, or inhibition, of behavior (Muraven & Baumeister, 2000; Pennebaker, 1989, 1995; Polivy, 1998; Traue & Pennebaker, 1993). Contemporary message processing and person perception studies tend to share the underlying notion that people sometimes think carefully and at other times think casually. When thinking carefully about persuasive communications, people are expected to evaluate the quality of arguments. By contrast, when thinking casually about such communications, people are expected to consider the communications’ secondary features, such as the number of arguments that they include. When thinking carefully about social targets, people are expected to take into account facts about them and possible situational influences on their behavior. By contrast, when thinking casually about such targets, people are expected to utilize social stereotypes and make simple characterological attributions. Irrespective of whether thought is focused on a message or a person, it is believed to be more effortful when it is careful than when it is casual. Therefore, researchers in these areas could use CV measures to gain information about how engaged participants are when confronted with influence attempts and social perception challenges.
298
REX A. WRIGHT AND LESLIE D. KIRBY
Thought suppression and self-regulation studies tend to share the notion that restraint is effortful, which intimates that CV measures could be usefully applied to determine the extent to which it is taking place. Relevant to this CV measurement possibility is research by Gross (discussed by Gross, 1998; Gross & Mu˜noz, 1995) that shows greater CV arousal in participants instructed not to respond to emotional stimuli than in participants allowed to respond naturally. Also relevant is the hypothesis expressed by Fowles some years back (e.g., 1980) that behavioral activation is associated with increased HR activity, whereas behavioral inhibition is associated with increased electrodermal activity. Although Fowles’ hypothesis is a venerable one, it is open to question in light of current understandings of CV response determination. There is broad agreement at this point that CV parameters should not be evaluated in isolation of one another and that HR, in particular, is hard to interpret (e.g., Berntson, Cacioppo, & Quigley, 1993). Evidence that HR was not responsive to inhibition in a given instance or set of instances does not demonstrate that it cannot be so responsive or preclude the possibility that CV outcomes more sensitive to sympathetic influence would have been responsive had they been measured.
F. EFFORT ASSUMPTION REFINEMENT As a final point of observation, it is worth noting that social psychological studies such as those above might benefit not only from the inclusion of CV measures in their research protocols, but also from a reconsideration of their effort assumptions in light of the effort ideas outlined here. Consider in this regard, for example, studies concerned with learned helplessness, which tend to have underlying them the assumption that extended failure leads to low effort. Effort propositions from the integrative analysis suggest that such failure may not always have the assumed effect. In theory, the influence of extended failure on effort should depend on (1) whether those who fail conclude that they cannot succeed, and (2) how important those failing feel it is to succeed in related performance realms. Failure experiences are likely to increase appraisals of task difficulty in relevant realms (Ford & Brehm, 1987; see also the earlier ability studies). Extended failure could lead people to believe they cannot succeed, but it also could lead them to believe that success is merely more difficult than they first thought. Where failure leads to the latter belief, it should produce either high or low effort, depending on the perceived importance of success (i.e., performers’ potential motivation).10 10
This analysis is, of course, similar in some respects to the Wortman and Brehm (1975) integration of the learned helplessness model and reactance theory (Brehm, 1966). However, the Wortman and Brehm integration did not consider the impact of failure on demand appraisals and had a role for success importance that is different from the one assumed here.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
299
To provide another example, studies of persuasive message processing generally assume that message recipients will process carefully when ability and (potential) motivation are high, and process lightly when these things are low. By contrast, effort propositions from the integrative analysis suggest that processing will sometimes be more intense under low-ability conditions than under high-ability conditions. Specifically, where processing success is viewed as possible and worthwhile irrespective of ability, task engagement should be inversely proportional to ability. Effort propositions from the integrative analysis also suggest that processing intensity will lower regardless of success importance if persuasive arguments are perceived as easy to comprehend. Depending on the research outcome studied, differences in perspective like these may or may not be of practical significance.
VII. Conclusion We have outlined, and reviewed evidence relevant to, an integrative analysis concerned with the determinants of CV responses in people confronted with behavioral challenges. Core elements of the analysis are the ideas (1) that sympathetically mediated CV adjustment varies with task engagement; and (2) that engagement is a function of what can, will, and must be done to satisfy motives. Although the base of evidence supporting the integrative analysis is not complete, it makes an impressive case for the general integrative line of thought. Moreover, it calls seriously into question several popular alternative CV-response hypotheses. The integrative analysis has potential for shedding light on the CV influence of social psychological variables, and suggests means of documenting and refining social effort assumptions. We will be gratified if investigators with social interests find it useful in their work.
Acknowledgments Preparation of this chapter was facilitated by Grant SBR97-27707 from the NSF to the first author and National Research Service Award 1 F32 MH12522-01 from the NIH to the second author.
References Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49–74. Anderson, N. B., Lane, J. D., Taguchi, F., Williams, R. B., & Houseworth, S. J. (1989). Race, parental
300
REX A. WRIGHT AND LESLIE D. KIRBY
history of hypertension, and patterns of cardiovascular reactivity in women. Psychophysiology, 26, 39–47. Annis, S. M. (1999). Effects of perceived ability, incentive value, and objective task demand on cardiovascular response. Unpublished masters thesis, University of Alabama at Birmingham, Birmingham, Alabama. Annis, S. M., Wright, R. A., & Williams, B. J. (2001). Interactional influence of ability perception and task demand on cardiovascular response: Appetitive effects at three levels of challenge. Journal of Applied Social Psychology, in press. Ansfield, M. E., & Wegner, D. M. (1996). The feeling of doing. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 482–506). New York: Guilford. Anthony, L., & Wright, R. A. (1998). Unpublished data. University of Alabama at Birmingham, Birmingham, Alabama. Armor, D. A., & Taylor, S. E. (1998). Situated optimism: Specific outcome expectancies and selfregulation. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 30, pp. 309– 379). New York: Academic Press. Averill, J. R. (1979). Anger. In H. E. Howe & R. A. Dienstbier (Eds.), Nebraska Symposium on Motivation: Human emotions (Vol. 26, pp. 1–80). Lincoln: Univ.of Nebraska Press. Ax, A. F. (1953). The physiological differentiation between fear and anger in humans. Psychosomatic Medicine, 15, 433–442. Bandura, A. (1986). The social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1991). Self-efficacy conception of anxiety. In R. Schwarzer & R. A. Wicklund (Eds.), Anxiety and self-focused attention (pp. 89–110). Chur, Switzerland: Harwood. Bandura, A., Reese, L., & Adams, N. E. (1982). Microanalysis of action and fear arousal as a function of differential levels of perceived efficacy. Journal of Personality and Social Psychology, 43, 5–21. Bandura, A., Taylor, C. B., Williams, S. L., Mefford, I. N., & Barchas, J. D. (1985). Catecholamine secretion as a function of perceived coping self-efficacy. Journal of Consulting and Clinical Psychology, 53, 406–414. Baron, R. S. (1986). Distraction-conflict theory: Progress and problems. Advances in Experimental Social Psychology, 19, 1–40. Batson, C. D. (1987). Prosocial motivation: Is it ever truly altruistic? In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 65–122). New York: Academic Press. Baumeister, R. F. (1982). A self-presentational view of social phenomena. Psychological Bulletin, 91, 656–662. Belanger, D., & Feldman, S. M. (1962). Effects of water deprivation upon heart rate and instrumental activity in the rat. Journal of Comparative and Physiological Psychology, 55, 220–225. Berntson, G. G., Cacioppo, J. T., & Quigley, K. S. (1993). Cardiac psychophysiology and autonomic space in humans: Empirical perspectives and conceptual implications. Psychological Bulletin, 114, 296–322. Blascovich, J., & Katkin, E. S. (Eds.) (1993). Cardiovascular reactivity to psychological stress and disease. Washington, DC: American Psychological Association. Blascovich, J., & Mendes, W. (1999). Challenge and threat appraisals: The role of affective cues. In J. Forgas (Ed.), Feeling and thinking: The role of affect in social cognition. Cambridge, UK: Cambridge Univ. Press. Blascovich, J., Mendez, W., Hunter, S. B., & Salomon, K. (1999). Social “facilitation” as Challenge and Threat. Journal of Personality and Social Psychology, 77, 68–77. Blascovich, J., & Tomaka, J. (1996). The biopsychosocial model of arousal regulation. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 29, pp. 1–51). New York: Academic Press.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
301
Brehm, J. W. (1966). A theory of psychological reactance. New York: Academic Press. Brehm, J. W. (1999). The intensity of emotion. Personality and social Psychology Review, 3, 2–22. Brehm, J. W., & Self, E. (1989). The intensity of motivation. In M. R. Rozenweig & L. W. Porter (Eds.), Annual review of psychology (pp. 109–131). Palo Alto: Annual Reviews, Inc. Brown, P. C., & Smith, T. W. (1992). Social influence, marriage, and the heart: Cardiovascular consequences of interpersonal control in husbands and wives. Health Psychology, 11, 88–96. Brown, P. C., Smith, T. W., & Benjamin, L. S. (1998). Perceptions of spouse dominance predict blood pressure reactivity during marital interactions. Annals of Behavioral Medicine, 20, 286–293. Chaiken, S., Giner-Sorolla, & Chen, S. (1996). Beyond accuracy: Defense and impression motives in heuristic and systematic information processing. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 553–578). New York: Guilford. Clements, K., & Turpin, G. (1995). Effects of feedback and task difficulty on electrodermal activity and heart rate: An examination of Fowles’ three arousal model. Journal of Psychophysiology, 9, 231–242. Collins, A., & Frankenhauser, M. (1978). Stress responses in male and female engineering students. Journal of Human Stress, 4, 43–48. Contrada, R. J., Cather, C., & O’Leary, A. (1999). Personality and health: Dispositions and processes in disease susceptibility and adaptation to illness. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed.). New York: Guilford. Contrada, R. J., Wright, R. A., & Glass, D. C. (1984). Task difficulty, Type A behavior pattern, and cardiovascular response. Psychophysiology, 21, 638–646. Cottrell, N. B. (1972). Social facilitation. In C. G. McClintock (Ed.), Experimental social psychology. New York: Holt. Cottrell, N. B., Wack, D. L., Sekerak, G. J., & Rittle, R. H. (1968). Social facilitation of dominant responses by the presence of an audience and the mere presence of others. Journal of Personality and Social Psychology, 9, 245–250. Cox, T. (1978). Stress. Baltimore, MD: University Park Press. Davis, M. C., & Matthews, K. A. (1996). Do gender-relevant characteristics determine cardiovascular reactivity? Match versus mismatch of traits and situations. Journal of Personality and social Psychology, 71, 527–535. Dienstbier, R. A. (1989). Arousal of physiological toughness: Implications for physical and mental health. Psychological Review, 96, 84–100. Elliott, R. (1965). Reaction times and heart rates as functions of magnitude of incentive and probability of success. Journal of Personality and Social Psychology, 2, 604–609. Elliott, R. (1969). Tonic heart rate: Experiments on the effects of collative variables lead to a hypothesis about its motivational significance. Journal of Personality and social Psychology, 12, 211–228. Eubanks, L., Wright, R. A., & Williams, B. J. (2001). Reward influence on the heart: Cardiovascular response as a function of incentive value at five levels of task demand. Unpublished manuscript. Ewart, C. K. (1995). Self-efficacy and recovery from heart attack: Implications for a social-cognitive analysis of exercise and emotion. In J. E. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: Theory, research, and application (pp. 203–224). New York: Plenum. Falkner, B., & Kushner, H. (1989). Race differences in stress-induced reactivity in young adults. Health Psychology, 8, 613–627. Ford, C. E., & Brehm, J. W. (1987). Effort expenditure following failure. In C. R. Snyder & C. E. Ford (Eds.), Coping with negative life events: Clinical and social psychological perspectives (pp. 51–79). New York: Plenum. Ford, C. E., & Neale, J. M. (1985). Learned helplessness and judgments of control. Journal of Personality and Social Psychology, 49, 1330–1336.
302
REX A. WRIGHT AND LESLIE D. KIRBY
Fowles, D. C. (1980). The three-arousal model: Implications of Gray’s two-factor theory for heart rate and electrodermal activity and psychopathology. Psychophysiology, 17, 87–104. Fowles, D. C. (1982). Heart rate as an index of anxiety: Failure of a hypothesis. In J. T. Cacioppo & R. E. Petty (Eds.), Perspectives in cardiovascular psychophysiology (pp. 93–126). New York: Guilford. Fowles, D. C., Fisher, A. E., & Tranel, D. T. (1982). The heart beats to reward: The effect of monetary incentive on heart rate. Psychophysiology, 19, 506–513. Gallo, L. C., Smith, T. W., & Kircher, J. C. (2000). Cardiovascular and electrodermal responses to support and provocation: Interpersonal methods in the study of psychophysiological reactivity. Psychophysiology, 37, 289–301. Geen, R. G. (1991). Social motivation. In M. R. Rosenzweig & M. R. Porter (Eds.), Annual review of psychology (pp. 377–399). Palo Alto: Annual Reviews, Inc. Geen, R. G. (1995). Human motivation: A social psychological approach. Pacific Grove, CA: Brooks/Cole. Geen, R. G., & Bushman, B. J. (1989). The arousing effects of social presence. In H. Wagner & A. Manstead (Eds.), Handbook of social psychophysiology (pp. 261–281). London: Wiley. Gendolla, G. H. E. (1998). Effort as assessed by motivational arousal in identity relevant tasks. Basic and Applied Social Psychology, 20, 111–121. Gendolla, G. H. E. (1999). Self-relevance of performance, task difficulty, and task engagement assessed as cardiovascular response. Motivation and Emotion, 23, 45–66. Gendolla, G. H. E. (2000). On the impact of mood on behavior: An integrative theory and a review. Review of General Psychology, in press. Gendolla, G. H. E., Abele, A. E., & Kr¨usken, J. (2001). On the role of mood in the energization process. Emotion, in press. Gendolla, G. H. E., & Kr¨usken, J. (2001). The joint impact of mood state and task difficulty on cardiovascular and electrodermal reactivity in active coping. Psychophysiology, in press. Gerin, W., Pieper, C., Levy, R., & Pickering, T. G. (1992). Social support in social interaction: A moderator of cardiovascular reactivity. Psychosomatic Medicine, 54, 324–336. Gilbert, D. T. (1998). Ordinary personology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 89–150). New York: McGraw–Hill. Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 117, 21– 38. Gollwitzer, P. M., Wicklund, R. A., & Hilton, J. L. (1982). Admission of failure and symbolic selfcompletion: Extending Lewinian theory. Journal of Personality and Social Psychology, 43, 358– 371. Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2, 271–299. Gross, J. J., & Mu˜noz, R. F. (1995). Emotion regulation and mental health. Clinical Psychology: Science and Practice, 2, 151–164. Hackett, G., & Betz, N. (1995). Self-efficacy and career choice and development. In J. E. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: Theory, research, and application (pp. 249– 276). New York: Plenum. Hahn, W. W., Stern, J. A., & McDonald, D. G. (1962). Effects of water deprivation and bar pressing activity on heart rate in the male albino rat. Journal of Comparative and Physiological Psychology, 55, 786–790. Hassett, J., & Danforth, D. (1982). An introduction to the cardiovascular system. In J. T. Cacioppo & R. E. Petty (Eds.), Perspectives in cardiovascular psychophysiology (pp. 4–18). New York: Guilford. Heckhausen, H. (1991). Motivation and action (P. K. Leppman, Trans.). Berlin: Springer-Verlag. Herd, J. A. (1983). Physiological basis for behavioral influences in atherosclerosis. In T. M. Dembroski,
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
303
T. H. Schmidt, & G. Blumchen (Eds.), Biobehavioral bases of heart disease (pp. 248–256). Basel, Switzerland: Karger. Hilton, J. L., & Darley, J. M. (1991). The effects of interaction goals on person perception. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 24). New York: Academic Press. Izard, C. E. (1993). Organizational and motivational functions of discrete emotions. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (631–641). New York: Guilford. James, S. A., Hartnett, S. A., & Kalsbeck, W. D. (1983). John Henryism and blood pressure differences among black men. Journal of Behavioral Medicine, 6, 259–277. Johnson, H. J. (1963). Decision making, coflict, and physiological arousal. Journal of Abnormal and Social Psychology, 2, 114–124. Kamarck, T. W., Manuck, S. B., & Jennings, J. R. (1990). Social support reduces cardiovascular reactivity to psychological challenge: A laboratory model. Psychosomatic Medicine, 52, 42–58. Kamarck, T. W., Peterman, A. H., & Raynor, D. A. (1998). The effects of the social environment on stress-related cardiovascular activation: Current findings, prospects, and implications. Annals of Behavioral Medicine, 20, 247–256. Kahneman, D., Slovic, P., & Tversky, A. (1982). Judgment under uncertainty: Heuristics and biases. Cambridge, UK: Cambridge, Univ. Press. Kukla, A. (1972). Foundations of an attributional theory of performance. Psychological Review, 79, 454–470. Lash, S. J., Eisler, R. M., & Southard, D. R. (1995). Sex differences in cardiovascular reactivity as a function of the appraised gender relevance of the stressor. Behavioral Medicine, 21, 86–94. Lash, S. J., Gillespie, B. L., Eisler, R. M., & Southard, D. R. (1991). Sex differences in cardiovascular reactivity: Effects of gender relevance of the stressor. Health Psychology, 10, 392–398. Latane, B., Williams, K., & Harkins, S. (1979). Many hands make light the work: The causes and consequences of social loafing. Journal of Personality and Social Psychology, 37, 822–832. Lazarus, R. S. (1968). Emotion and adaptation: Conceptual and empirical relations. In W. J. Arnold (Ed.), Nebraska symposium on motivation (pp. 173–270). Lincoln, NE: Univ. of Nebraska Press. Lazarus, R. S., & Folkman, S. (1994). Stress, appraisal, and coping. New York: Springer-Verlag. Lewicka, P. (1986). Nonconscious social information processing. Orlando, FL: Academic Press. Liberson, C. W., & Liberson, W. T. (1975). Sex differences in autonomic responses to electric shock. Psychophysiology, 17, 243–252. Light, K. C., & Obrist, P. A. (1980). Cardiovascular response to stress: Effects of the opportunity to avoid shock, shock experience, and performance feedback. Psychophysiology, 17, 243–252. Light, K. C., & Obrist, P. A. (1983). Task difficulty, heart rate reactivity, and cardiovascular responses to an appetitive reaction time task. Psychophysiology, 20, 301–312. Light, K. C., Turner, J. R., Hinderliter, & Sherwood, A. (1993). Race and gender comparisons. I. Hemodynamic responses to a series of stressors. Health Psychology, 12, 354–365. Locke, E. A., & Latham, G. P. (1984). Goal setting: A motivational technique that works! Englewood Cliffs, NJ: Prentice-Hall. Locke, E. A., & Latham, G. P. (1994). Goal setting theory. In H. F. O’Neil, Jr. & M. Drillings (Eds.), Motivation: Theory and research. Hillsdale, NJ: Erlbaum. Maddux, J. E. (1995). Self-efficacy theory: An introduction. In J. E. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: Theory, research, and application (pp. 3–27). New York: Plenum. Manuck, S. B., Harvey, S. H., Lechleiter, S. L., & Neal, K. S. (1978). Effects of coping on blood pressure responses to threat of aversive stimulation. Psychophysiology, 15, 544–549. Matthews, K. M., Davis, M. C., Stoney, C. M., Owens, J. F., & Caggiula, A. R. (1991). Does the gender relevance of the stressor influence sex differences in psychophysiological responses? Health Psychology, 10, 112–120. Meyer, W. U. (1987). Perceived ability and achievement-related behavior. In F. Halisch & J. Kuhl (Eds.), Motivation, intention, and volition. Berlin: Springer-Verlag.
304
REX A. WRIGHT AND LESLIE D. KIRBY
Muraven, M. B., & Baumeister, R. F. (2000). Self-regulation and the depletion of limited resources: Does self-control resemble a muscle? Psychological Bulletin, 126, 247–259. Muraven, M. B., Tice, D. M., & Baumeister, R. F. (1998). Self-control as a limited resource: Regulatory depletion patterns. Journal of Personality and Social Psychology, 74, 774–789. Murray, J. B. (1994). Gender relevant tasks and the elicitation of sex differences in cardiovascular response. Unpublished doctoral dissertation, University of Alabama at Birmingham. Murray, J. B., Wright, R. A., & Williams, B. J. (1993). Difficulty as a determinant of cardiovascular response: Moderating effect of instrumentality in an alleviation paradigm. International Journal of Psychophysiology, 15, 135–145. Neuberg, S. L. (1996). Expectancy influences in social interaction: The moderating role of social goals. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 529–552). New York: Guilford. Obrist, P. A. (1976). The cardiovascular-behavioral interaction as it appears today. Psychophysiology, 13, 95–107. Obrist, P. A. (1981). Cardiovascular psychophysiology: A perspective. New York: Plenum. Obrist, P. A., Gaebelein, C. J., Teller, E. S., Langer, A. W., Grignolo, A., Light, K. C., & McCubbin, J. A. (1978). The relationship among heart rate, carotid dp/dt, and blood pressure in humans as a function of type of stress. Psychophysiology, 15, 102–115. Obrist, P. A., Webb, R. A., Sutterer, J. R., & Howard, J. L. (1970). The cardiac–somatic relationship: Some reformulations. Psychophysiology, 6, 569–587. O’Leary, A., & Brown, S. (1995). Self-efficacy and the physiological stress response. In J. E. Maddux (Ed.), Self-efficacy, adaptation, and adjustment: Theory, research and application (pp. 227–246). New York: Plenum. Papillo, J. F., & Shapiro, D. (1990). The cardiovascular system. In J. T. Cacioppo, & L. G. Tassinary (Eds.), Principles of psychophysiology: Physical, social, and inferential elements (pp. 456–512). New York: Cambridge Univ. Press. Pennebaker, J. W. (1989). Confession, inhibition, and disease. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 22, pp. 211–244). New York: Academic Press. Pennebaker, J. W. (Ed.) (1995). Emotion, disclosure, and health. Washington, DC: American Psychological Association. Petty, R. E., & Wegner, D. T. (1998). Attitude change: Multiple roles for persuasion variables. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 323–390). New York: McGraw-Hill. Polefrone, J. M., & Manuck, S. B. (1987). Gender differences in cardiovascular and neuroendocrine response to stressors. In R. C. Barnett, L. Biener, & G. K. Baruch (Eds.), Gender and stress (pp. 13–38). New York: Free Press. Polivy, J. (1998). The effects of behavioral inhibition: Integrating internal cues, cognition, behavior, and affect. Psychological Inquiry, 9, 181–204. Pyszczynski, T., & Greenberg, J. (1983). Determinants of reductions in intended effort as a strategy for coping with anticipated failure. Journal of Research in Personality, 17, 776–783. Phodewaldt, F., & Fairfield, M. (1991). Claimed self-handicaps and the self-handicapper: The relation of reduction in intended effort to performance. Journal of Research in Personality, 25, 402–417. Saab, P. G. (1989). Cardiovascular and neuroendocrine response to challenge in males and females. In N. Schneiderman, S. M. Weiss, & P. G. Kaufmann (Eds.), Handbook of research methods in cardiovascular behavioral medicine (pp. 453–481). New York: Plenum. Sanders, G. S. (1981). Driven by distraction: An integrative review of social facilitation theory and research. Journal of Experimental Social Psychology, 17, 227–251. Schachter, S. (1964). The interaction of cognitive and physiological determinants of emotional state. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 1, pp. 49–80). New York: Academic Press.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
305
Scheier, M. F., & Carver, C. S. (1985). Optimism, coping, and health: Assessment and implications of generalized outcome expectancies. Health Psychology, 4, 219–247. Sch¨onpflug, W. (1983). Coping efficiency and situational demands. In G. R. J. Hockey (Ed.), Stress and fatigue in human performance (pp. 299–330). New York: Wiley. Seta, C. E., & Seta, J. J. (1992). Increments and decrements in mean arterial pressure as a function of audience composition: An averaging and summation analysis. Personality and Social Psychology Bulletin, 18, 173–181. Seta, J. J., Crisson, J. E., Seta, C. E., & Wang, M. A. (1989). Task performance and perceptions of anxiety: Averaging and summation in an evaluative setting. Journal of Personality and Social Psychology, 56, 387–396. Seta, J. J., & Seta, C. E. (1983). The impact of personal equity processes on performance in a group setting. In P. B. Paulus (Ed.), Basic group processes (pp. 121–147). New York: SpringerVerlag. Shepperd, J. A. (1993). Productivity loss in performance groups: A motivational analysis. Psychological Bulletin, 113, 67–81. Sherwood, A., Allen, M. T., Fahrenberg, J., Kelsey, R. M., Lovallo, W. R., & van Doornen, L. J. P. (1990). Methodological guidelines for impedance cardiography. Psychophysiology, 27, 1–23. Smith, C. A., & Pope, L. K. (1992). Appraisal and emotion: The interactional contributions of situational and dispositional factors. In M. S. Clark (Ed.), Review of personality and social psychology: Emotion and social behavior (Vol. 14, pp. 32–62). Newbury Park, CA: Sage. Smith, T. W., Allred, K. D., Morrison, C. A., & Carlson, S. D. (1989). Cardiovascular reactivity and interpersonal influence: Active coping in a social context. Journal of Personality and Social Psychology, 56, 209–218. Smith, T. W., Baldwin, M., & Christenson, A. J. (1990). Interpersonal influence as active coping: Effects of task difficulty on cardiovascular reactivity. Psychophysiology, 27, 429–437. Smith, T. W., & Blascovich, J. (1992). Social psychophysiology. Journal of Social and Clinical Psychology, 11, 197–231. Smith, T. W., & Christensen, A. J. (1992). Life-style and hostility. In H. S. Friedman (Ed.), Hostility, coping, and health (pp. 33–48). Washington, DC: American Psychological Association. Smith, T. W., & Gerin, W. (1998). The social psychophysiology of cardiovascular response: An introduction to the special issue. Annals of Behavioral Medicine, 20, 243–246. Smith, T. W., Limon, J. P., Gallo, L. C., & Ngu, L. Q. (1996). Interpersonal control and cardiovascular activity: Goals, behavioral expression, and the moderating effects of sex. Journal of Personality and Social Psychology, 70, 1012–1024. Smith, T. W., Nealey, J. B., Kircher, J. C., & Limon, J. P. (1997). Social determinants of cardiovascular reactivity: Effects of incentive to influence and evaluative threat. Psychophysiology, 34, 65–73. Smith, T. W., Ruiz, J. M., & Uchino, B. N. (2000). Vigilance, active coping, and cardiovascular reactivity during social interaction in young men. Health Psychology, 19, 382–392. Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of AfricanAmericans. Journal of Personality and Social Psychology, 69, 797–811. Steptoe, A., Moses, J., Mathews, A., & Edwards, S. (1990). Aerobic fitness, Physical activity, and psychophysiological reactions to mental tasks. Psychophysiology, 27, 264–274. Sternberg, S. (1966). High speed scanning in human memory. Science, 153, 652–654. Stoney, C. M., Davis, M. C., & Matthews, K. A. (1987). Sex difference in physiological responses to stress and in coronary disease: A causal link? Psychophysiology, 24, 127–131. Stoney, C. M., Matthews, K. A., McDonald, R. H., & Johnson, C. A. (1988). Sex differences in lipid, lipoprotein, cardiovascular and neuroendocrine responses to acute stress. Psychophysiology, 27, 125–135.
306
REX A. WRIGHT AND LESLIE D. KIRBY
Storey, P. L. (1996). Cardiovascular gender effects as a function of task difficulty and the gender relevance of a behavioral challenge. Unpublished doctoral dissertation, University of Alabama at Birmingham. Storey, P. L., Wright, R. A., & Williams, B. J. (1996). Need as a moderator of the difficulty– cardiovascular response relation: The case of fluid deprivation. Journal of Psychophysiology, 10, 228–238. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193–210. Toi, M., & Batson, C. D. (1982). More evidence that empathy is a source of altruistic motivation. Journal of Personality and Social Psychology, 43, 281–292. Tomaka, J., Blascovich, J., Kibler, J., & Ernst, J. M. (1997). Cognitive and physiological antecedents of threat and challenge appraisal. Journal of Personality and Social Psychology, 73, 63–72. Tomaka, J., Palacios, R., Schneider, K. T., Colotla, M., Concha, J. B., & Herrald, M. M. (1999). Assertiveness predicts threat and challenge reactions to potential stress among women. Journal of Personality and Social Psychology, 6, 1008–1021. Tranel, D. T., Fisher, A. E., & Fowles, D. C. (1982). Magnitude of incentive effects on heart rate. Psychophysiology, 19, 514–519. Traue, H. C., & Pennebaker, J. W. (1993). Emotion inhibition and health. G¨ottingen: Hogrefe & Huber. Turner, J. R., Sherwood, A., & Light, K. C. (Eds.) (1992). Individual differences in cardiovascular response to stress. New York: Plenum. Van Schijndel, M., DeMey, H., & N¨aring, G. (1984). Effects of behavioral control and Type A behavior on cardiovascular responses. Psychophysiology, 21, 501–509. Uleman, J. S., & Bargh, J. A. (Eds.) (1989). Unintended thought. New York: Guilford Press. Wegner, D. M. (1992). You can’t always think what you want: Problems in the suppression of unwanted thoughts. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 193–225). San Diego, CA: Academic Press. Weiner, B. (1992). Human motivation: Metaphors, theories, and research. Newbury Park, CA: Sage. Wicklund, R. A., & Braun, O. L. (1987). Incompetence and the concern with human categories. Journal of Personality and Social Psychology, 53, 373–382. Wicklund, R. A., & Gollwitzer, P. M. (1982). Symbolic self-completion. Hillsdale, NJ: Erlbaum. Wiedenfeld, S. A., O’Leary, A., Bandura, A., Brown, S., Levine, S., & Raska, K. (1990). Impact of perceived coping efficacy on components of the immune system. Journal of Personality and social Psychology, 59, 1082–1094. Williams, S. L. (1995). Self-efficacy and anxiety and phobic disorders. In J. E. Maddux (Ed.), Self– efficacy, adaptation, and adjustment: Theory, research, and application (pp. 69–102). New York: Plenum. Wortman, C. B., & Brehm, J. W. (1975). Responses to uncontrollable outcomes: An integration of reactance theory and the learned helplessness model. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 8, pp. 277–336). New York: Academic Press. Wright, R. A. (1996). Brehm’s theory of motivation as a model of effort and cardiovascular response. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 424–453). New York: Guilford. Wright, R. A. (1998). Ability perception and cardiovascular response to behavioral challenge. In M. Kofka, G. Weary, & G. Sedek (Eds.), Personal control in action: Cognitive and motivational mechanisms (pp. 197–232). New York: Guilford. Wright, R. A., & Brehm, J. W. (1989). Energization and goal attractiveness. In L. A. Pervin (Ed.), Goal concepts in personality and social psychology (pp. 169–210). Hillsdale, NJ: Erlbaum. Wright, R. A., Brehm, J. W., & Bushman, B. (1989). Cardiovascular responses to threat: Effects of the difficulty and availability of a cognitive avoidance task. Basic and Applied Social Psychology, 10, 161–171.
EFFORT DETERMINATION OF CARDIOVASCULAR RESPONSE
307
Wright, R. A., Brehm, J. W., Crutcher, W., Evans, M. T., & Jones, A. (1990). Avoidant control difficulty and aversive incentive appraisals: Additional evidence of an energization effect. Motivation and Emotion, 14, 5–73. Wright, R. A., Contrada, R. J., & Patane, M. J. (1986). Task difficulty, cardiovascular response, and the magnitude of goal valence. Journal of Personality and Social Psychology, 51, 837–843. Wright, R. A., Dill, J. C., Geen, R. G., & Anderson, C. A. (1998). Social evaluation influence on cardiovascular response to a fixed behavioral challenge: Effects across a range of difficulty levels. Annals of Behavioral Medicine, 20, 277–285. Wright, R. A., & Dill, J. C. (1993). Blood pressure responses and incentive appraisals as a function of perceived ability and objective task demand. Psychophysiology, 30, 152–160. Wright, R. A., & Dismukes, A. (1995). Cardiovascular effects of experimentally-induced efficacy (ability) appraisals at low and high levels of avoidant task demand. Psychophysiology, 32, 172– 176. Wright, R. A., & Gregorich, S. (1989). Difficulty and instrumentality of imminent behavior as determinants of cardiovascular response and self-reported energy. Psychophysiology, 26, 586–592. Wright, R. A., & Hodges, D. (1999). The domain specificity of cardiovascular ability effects. Basic and Applied Social Psychology, 24, 291–300. Wright, R. A., Killebrew, K., & Pimpalapure, D. (2001). Cardiovascular incentive effects where a challenge is unfixed: Demonstrations involving social evaluation, evaluator status, and monetary reward. Unpublished manuscript. Wright, R. A., Murray, J. B., Storey, P., & Williams, B. J. (1997). Ability analysis of gender relevance and sex differences in cardiovascular response to behavioral challenge. Journal of Personality and Social Psychology, 73, 405–417. Wright, R. A., Tunstall, A. M., Williams, B. J., Goodwin, J. S., & Harmon-Jones, E. (1995). Social evaluation and cardiovascular response: An active coping approach. Journal of Personality and Social Psychology, 69, 530–543. Wright, R. A., Shaw, L. L., & Jones, C. R. (1990). Task demand and cardiovascular response magnitude: Further evidence of the mediating role of success importance. Journal of Personality and Social psychology, 59, 1250–1260. Wright, R. A., Wadley, V. G., Pharr, R. P., & Butler, M. (1994). Interactive influence of self-reported ability and avoidant task demand on anticipatory cardiovascular reactivity. Journal of Research in Personality, 28, 68–86. Wright, R. A., Williams, B. J., & Dill, J. C. (1992). Interactive effects of difficulty and instrumentality of avoidant behavior on cardiovascular reactivity. Psychophysiology, 29, 677–689. Zajonc, R. B. (1965). Social facilitation. Science, 149, 269–274. Zajonc, R. B. (1980). Compresence. In P. Paulus (Ed.), Psychology of group influence. Hillsdale, NJ: Erlbaum. Zillmann, D. (1983). Transfer of excitation in emotional behavior. In J. T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A sourcebook (pp. 215–240). New York: Guilford.
This Page Intentionally Left Blank
INDEX
A Accountability, see Videotape confession bias Affordances, perception effect on behavior, 2, 4 Ambivalence toward Men Inventory (AMI) content, 155–156 feminism and hostility towards men, 165 hostility versus benevolence towards men, 116–117, 153, 163–165, 181 validation across cultures correlations with Ambivalent Sexism Inventory, 161–162 fitting of data, 161 hostility and benevolence towards men mean scores, 162–163 study design, 161 validation in United States convergent and discriminant validity of hostility and benevolence towards men scales, 159–160 factor loadings for items, 157–158 factor structure, 157–159 predictive validity, 160–161 sex differences in hostility and benevolence towards men scores, 159 study design, 155, 157 Ambivalent Sexism Inventory (ASI) ambivalence concept, 149–150 comparison with other measures Attitudes toward Women Scale, 125–126 Modern Sexism scale, 126 Neo-sexism scale, 126 consequences of hostile and benevolent sexism attitudes toward and stereotypes about women explicit attitudes, 143–146 implicit attitudes, 146–147
309
disarming resistance to sexism by benevolent sexism, 147–149, 180–181 overview, 140–141 structural gender inequality, 141–142 content, 123–125 hostile versus benevolent sexism measurement, 116–117, 125 reconciliation of hostile and benevolent sexism, 149–152 validation across cultures average hostile and benevolent sexism scores across nations, 136–138 correlations between hostile and benevolent sexism across nations, 135–136 within nations, 134–135 factor structure, 134 overview, 133–134 summary, 138–140 validation in United States convergent and discriminant validity of hostile and benevolent sexism scales, 132–133 factor structure correlation of hostile and benevolent sexism, 129 overview, 127 preferred and alternative models, 127–129 subfactors and item loadings, 129–130 summary, 129–131 sex differences in hostile and benevolent sexism scores, 131 study design, 127 AMI, see Ambivalence toward Men Inventory ASI, see Ambivalent Sexism Inventory
310
INDEX
Attitudes toward Women Scale (AWS), Ambivalent Sexism Inventory comparison, 125–126 Authoritarianism, see Prejudice AWS, see Attitudes toward Women Scale
Feminism, see Ambivalence toward Men Inventory; Sexism F scale, personality correlations in prejudice, 42, 45 G
B Behavior segmentation hypothesis, mediation of videotape confession bias direct effect of segmentation on causality judgements, 238–241 initial test of perceptual segmentation hypothesis, 232–233 segmentation rate confounding with amount of information recalled, 236–238 segmentation task alteration of causality judgements, 234–236 technique overview, 231–232 Benevolence towards men, see Ambivalence toward Men Inventory Benevolent sexism, see Ambivalent Sexism Inventory
GDI, see Gender-Related Development Index GEM, see Gender Empowerment Measure Gender Empowerment Measure (GEM), sexism effects, 141–142 Gender-Related Development Index (GDI), sexism effects, 141–142 H Hostile sexism, see Ambivalent Sexism Inventory Hostility towards men, see Ambivalence toward Men Inventory Humanitarian–Egalitarianism scale, personality correlations in prejudice, 88–89 I
C Camera perspective bias, see Videotape confession bias Capital punishment, error rate, 189–190 Common-coding hypothesis, imitation, 8 Competitive jungle worldview, 51–52, 68, 70 Confession, see Videotape confession bias D Dangerous World scale, 61 Death row, error rate of capital punishment, 189–190 E Envious prejudice, see Prejudice Exploitive Manipulative Amoral Dishonesty scale, personality correlations in prejudice, 51
Ideomotor action, imitation, 7, 23 Imitation, see also Perception–behavior relationships automatic imitation benefits, 32–33 facial expressions, 10–12 human tendencies, 1, 32–33 movements and gestures, 12–13 perception–behavior relationships common-coding hypothesis, 8 ideomotor action, 7, 23 neurophysiological evidence, 6–7 strategic nature studies, 14–15 speech, 13–14 trait inference generation role, 16–17 Intellectual performance, stereotype activation effects knowledge tasks, 19–20 math testing, 22 memory, 20–21 relation to stereotype threat, 20–22
F Facial expression, perception effect on behavior, 10–12
J Juror, see Videotape confession bias
INDEX L Liking, enhancement of relationships, 31–32 M Mental performance, stereotype activation effects knowledge tasks, 19–20 math testing, 22 memory, 20–21 relation to stereotype threat, 20–22 Miranda ruling, status, 246 Modern Sexism scale, Ambivalent Sexism Inventory comparison, 126 N Need for Cognition Scale, measurement of individual susceptibility to bias, 202–204 Neo-sexism scale, Ambivalent Sexism Inventory comparison, 126 P Paternalistic prejudice, see Prejudice Perception–behavior relationships animals versus humans, 3–4, 32 behavioral assimilation, 24–26 behavioral contrast, 25–28 imitation outcome common-coding hypothesis, 8 ideomotor action, 7 neurophysiological evidence, 6–7, 23 inhibition of action, 5–6 mediators, 22–24 moderation of link conflict with goals and purposes, 29 disincentives, 28–29 liking enhancement of relationship, 31–32 self-focused attention, 29–31 natural selection perspective, 3 observables and imitation facial expressions, 10–12 movements and gestures, 12–13 overview, 9–10 speech, 13–14 overview, 1–2 social stereotype perception
311
intellectual performance effects of activation knowledge tasks, 19–20 math testing, 22 memory, 20–21 relation to stereotype threat, 20–22 interpersonal behavior effects of activation, 19–20 motor behavior effects of activation, 18–19 overview, 9 trait activation mediation of imitation, 24 trait inference generation experimental evidence, 16–17 imitation role, 17 overview, 9, 16 Personal Power, Meanness, and Dominance scale (PP-MAD), prejudice studies, 51, 54 Positive coercion bias, see Videotape confession bias PP-MAD, see Personal Power, Meanness, and Dominance scale Prejudice, see also Sexism individual versus group phenomenon cognitive-motivational basis of ideological attitudes, 90 ethnocentric cultures and societies, 92–94 ethnocentrism in White Afrikaners versus European New Zealanders, 94–95 overview, 41 personality of individuals, 90–91 social worldview and role of situation, 91–92 integration of individual and intergroup dynamics of prejudice additive and interactive effects, 101–102 authoritanarianism intergroup threat interactions, 102–103 social dominance orientation interactions in minimal groups, 104 competitive–dominance motivation of prejudice, 98–100 duality of prejudice and intergroup dynamics, 100–101 minority versus majority prejudice, 97–100 motive activation and intergroup dynamics, 96–97 social dominance orientation and challenged dominance, 104–105 threat–control motivation of prejudice, 98–100
312
INDEX
Prejudice, see also Sexism (continued ) model of personality, ideology, and prejudice competitive jungle worldview, 51–52 intergroup attitude dimensions, 87, 89 prejudice types, 85–86, 88–90 racism types, 87–88 social dominance orientation relationship with right-wing authoritarianism, 58–59, 74, 78–80 socialization, personality, and worldview, 52–54 social threat relationship with authoritarianism, 50–51 study of personality and ideological attitudes correlations between personality and ideology, 57–58 samples and measures, 54–55 social conformity measurement as personality trait dimension, 55 tough-mindedness measurement as personality trait dimension, 56–57 summary, 105–106 two dimensions as motivational goal-schemas, 50–52, 85 personality correlations authoritarian personality F scale, 42, 45 Right-wing Authoritarianism scale, 42–43 Exploitive Manipulative Amoral Dishonesty scale, 51 Humanitarian–Egalitarianism scale, 88–89 Personal Power, Meanness, and Dominance scale, 51, 54 Protestant Work Ethic scale, 88–89 social dominance orientation age effects, 44–45 items in scale, 45–46 Right-wing Authoritarianism scale comparison, 43–50, 54, 57–59, 63–66, 74 sociocultural values in two dimensions, 48–49 sociopolitical attitudes in two dimensions, 46–48 stereotype content theory ambivalence as common reaction to out-groups, 171–172 comparison with Duckitt’s theory, 175–178
consequences of positive stereotypes, 178–179, 182–183 dimensions of stereotype content, 169–171 envious prejudice, 174, 182 paternalistic prejudice, 174, 182 social structure prediction of stereotypes, 172–175 status and interdependence types between groups, 172–175 structure equation modeling Auckland University 1998 study of personality, worldview, and ideology correlations among variables, 63–67 Dangerous World scale, 61 European/Pakeha sample, 61, 64–66 fitting of data, 62–63 maximum likelihood estimation analysis, 62 personality and ideology measures, 60–61 promajority and antiminority ethnic attitudes, 61–62 sample, 60 scope and objectives, 60 social worldview conceptualization, 61 Auckland University 1999 study of European/Pakeha sample childhood socialization dimension measurement, 70 competitive-jungle worldview measurement, 68, 70 correlations among variables, 72–74, 77 ethnic attitude scale items, 71 fitting of data, 70–72 measures, 68–70 participants, 67–68 paths to antiminority attitudes, 74–76 punitive socialization scale items, 72, 76 scope and objectives, 67 social conformity relationship with punitive socialization, 76 South Africa study of 1999 correlations among variables, 78–80 measures, 77–78 objectives, 77 participants, 77 summary of studies authoritarianism and social dominance as ideological dimensions, 82–83
INDEX core model of personality, worldview, and ideology, 80–81 ideology and intergroup attitudes, 81–82 methodological issues, 83–84 socialization and personality dimensions, 82 Protestant Work Ethic scale, personality correlations in prejudice, 88–89 R Racism, see Prejudice Right-wing Authoritarianism scale, see Prejudice S SDO, see Social dominance orientation Segmentation hypothesis, see Behavior segmentation hypothesis SEM, see Structure equation modeling Sexism ambivalent sexism domains and structural foundations men against women gender differentiation, competitive versus complementary, 121–122 heterosexual intimacy and hostility, 122–123 patriarchy and dominative versus protective paternalism, 120–121 women against men gender differentiation, compensatory and complementary, 154 heterosexual intimacy and hostility, 154–155 overview, 152–153 patriarchy and resentment of paternalism and maternalism, 153 cross-dimensional ambivalence, 181–182 development of gender prejudice adolescent ambivalence, 167–168 adult amibivalence, 168 childhood prejudice, 166–167 overview, 165–166, 168–169 interdependence between sexes, 117–119 measures, see Ambivalence toward Men Inventory; Ambivalent Sexism Inventory power differences between sexes, 117–119
313
prejudice versus close relationship approaches to gender relations, 115–117 stereotype content theory ambivalence as common reaction to out-groups, 171–172 comparison with Duckitt’s theory, 175–178 consequences of positive stereotypes, 178–179, 182–183 dimensions of stereotype content, 169–171 envious prejudice, 174, 182 paternalistic prejudice, 174, 182 social structure prediction of stereotypes, 172–175 status and interdependence types between groups, 172–175 structure of gender relations, 118–119, 179–180 Social dominance orientation (SDO) age effects, 44–45 integration of individual and intergroup dynamics of prejudice challenged dominance, 104–105 interactions in minimal groups, 104 items in scale, 45–46 Right-wing Authoritarianism scale comparison, 43–50, 54, 57–59, 63–66, 74, 78–80 S-R bond, perception effect on behavior, 2 Stereotype, see also Prejudice; Sexism ideomotor action, 7, 23 perception intellectual performance effects of activation knowledge tasks, 19–20 math testing, 22 memory, 20–21 relation to stereotype threat, 20–22 interpersonal behavior effects of activation, 19–20 motor behavior effects of activation, 18–19 overview, 9 trait activation mediation of imitation, 24 prejudice, see Prejudice Structure equation modeling (SEM), prejudice Auckland University 1998 study of personality, worldview, and ideology correlations among variables, 63–67 Dangerous World scale, 61 European/Pakeha sample, 61, 64–66 fitting of data, 62–63
314
INDEX
Structure equation modeling (SEM), prejudice (continued ) maximum likelihood estimation analysis, 62 personality and ideology measures, 60–61 promajority and antiminority ethnic attitudes, 61–62 sample, 60 scope and objectives, 60 social worldview conceptualization, 61 Auckland University 1999 study of European/Pakeha sample childhood socialization dimension measurement, 70 competitive-jungle worldview measurement, 68, 70 correlations among variables, 72–74, 77 ethnic attitude scale items, 71 fitting of data, 70–72 measures, 68–70 participants, 67–68 paths to antiminority attitudes, 74–76 punitive socialization scale items, 72, 76 scope and objectives, 67 social conformity relationship with punitive socialization, 76 South Africa study of 1999 correlations among variables, 78–80 measures, 77–78 objectives, 77 participants, 77 summary of studies authoritarianism and social dominance as ideological dimensions, 82–83 core model of personality, worldview, and ideology, 80–81 ideology and intergroup attitudes, 81–82 methodological issues, 83–84 socialization and personality dimensions, 82 Syntactic persistence, imitation, 13–14 T Trait activation, mediation of imitation, 24 V Videotape confession bias
behavior segmentation hypothesis and mediation of bias direct effect of segmentation on causality judgements, 238–241 initial test of perceptual segmentation hypothesis, 232–233 segmentation rate confounding with amount of information recalled, 236–238 segmentation task alteration of causality judgements, 234–236 technique overview, 231–232 coercion techniques, 193–194 criteria for admission as evidence, 191–192 elaborate trial simulation studies of camera perspective bias camera view of interrogator effects equal-focus versus detective-focus outcomes, 230 rationale for study, 227–228 study design, 228–230 judicial instruction effects outcomes, 226 rationale for study, 223–224 study design, 225 timing of instruction, 224 multiple viewing of confession effects confession case, 220–221 juror profile, 222 likelihood of guilt assessment, 222–223 possible outcomes, 219–220 study design, 221–222 voluntariness index, 222–223 rationale, 218–219 existence and robustness of camera perspective bias accountability of jurors, study of effects, 213–216 attribution index, 201 coercion ratings by camera view, 200–201, 227 content attention directions, study of effects, 209–210 credibility of suspect impact, 216–217 crime type effects, 202 deliberation effect study, 205–207 forewarning of bias, study of effects, 207–209, 224 limitations of intial studies, 218–219 longer/case-based confession bias, 211–213
INDEX Need for Cognition Scale measurement of individual susceptibility to bias, 202–204 positioning of camera, 195, 200, 205, 241 study design, 199–200 voluntariness index scoring, 203–204 exposure control as remedy, 245–246 mental contamination concept, 242–245 Miranda ruling implications, 246 point-of-view/salience effects in causal attribution fundamental attribution error, 196–197
315 implication for videotaped confessions, 197–198 mediation, 231 summary of studies, 195–197, 231 popularity of use, 194–195, 227, 241 positive coercion bias, 192 rationale for recording of confession, 190–191 research program overview and goals, 198–199 state mandates, 195 voluntary versus involuntary confessions, 191–192
This Page Intentionally Left Blank
CONTENTS OF OTHER VOLUMES
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 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 3
Volume 2
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
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
Volume 4
317
The Theory of Cognitive Dissonance: A Current Perspective Elliot Aronson
318
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 toTransmit 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
319
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
320
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 IdentityAnalytic 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 Psychology from the Standpoint of a Structural Symbolic Interactionism: Toward an Interdisciplinary Social Psychology 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
321
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
322
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 SelfBeliefs 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 The Social Psychology of Stanley Milgram Thomas Blass
323
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, Shortcomings, 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 RelationalRegulation 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 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
324
CONTENTS OF OTHER VOLUMES
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 Na¨ıve
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 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
CONTENTS OF OTHER VOLUMES 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
325
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
This Page Intentionally Left Blank