The Mass Media and the Dynamics of American Racial Attitudes

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

The Mass Media and the Dynamics of American Racial Attitudes Paul M. Kellstedt explains the variation in Americans’ racial attitudes over the last half-century, particularly the relationship between media coverage of and American public opinion on race. The analyses reveal that racial policy preferences have evolved in an interesting and unpredicted (if not unpredictable) fashion over the past fifty years. There have been sustained periods of liberalism, where the public prefers an active government to bring about racial equality, but these periods have invariably been followed by eras of conservatism, where the public wants the government to stay out of racial politics altogether. These opinions respond to cues presented in the national media. Kellstedt then goes on to examine the relationship between attitudes on the two major issues of the twentieth century: race and the welfare state. Although formerly distinct in the public mind, during the mid-1960s, race was “fused onto” the cluster of welfare-state issues. Since then, the over-time fluctuations of these two series have been indistinguishable. Moreover, the shift in the relationship is a function of media coverage of the issues of race and poverty. Paul M. Kellstedt is Assistant Professor of Political Science at Texas A&M University. He has been awarded grants from the National Science Foundation and was a Fellow at Harvard University’s Shorenstein Center for the Study of the Press, Politics, and Public Policy in 1999. His research has appeared in journals such as the American Journal of Political Science and Political Analysis.

i

CY252/Kellsted-FM

0 521 821711

June 13, 2003

ii

8:37

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

The Mass Media and the Dynamics of American Racial Attitudes

PAUL M. KELLSTEDT Texas A&M University

iii

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

published by the press syndicate of the university of cambridge The Pitt Building, Trumpington Street, Cambridge, United Kingdom

cambridge university press The Edinburgh Building, Cambridge cb2 2ru, uk 40 West 20th Street, New York, ny 10011-4211, usa 477 Williamstown Road, Port Melbourne, vic 3207, Australia ´ 13, 28014 Madrid, Spain Ruiz de Alarcon Dock House, The Waterfront, Cape Town 8001, South Africa http://www.cambridge.org c Paul M. Kellstedt 2003
This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2003 Printed in the United States of America Typeface Sabon 10/13.5 pt.

System LATEX 2ε

[tb]

A catalog record for this book is available from the British Library. Library of Congress Cataloging in Publication Data Kellstedt, Paul M., 1968– The mass media and the dynamics of American racial attitudes / Paul M. Kellstedt. p. cm. Includes bibliographical references and index. isbn 0-521-82171-1 – isbn 0-521-52915-8 (pb.) 1. Mass media and race relations – United States. i. Title. p94.5.m552u6467 2003 305.8 – dc21 2002041442 isbn 0 521 82171 1 hardback isbn 0 521 52915 8 paperback

iv

CY252/Kellsted-FM

0 521 821711

June 13, 2003

To the Kellstedt women: Deb, Abigail, and Elizabeth

v

8:37

CY252/Kellsted-FM

0 521 821711

June 13, 2003

vi

8:37

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

Contents

List of Figures List of Tables Preface 1 2 3 4 5 6

page ix xi xiii

Toward a Dynamic Perspective on Racial Attitudes Eras of Media Coverage of Race Eras of Racial Liberalism and Conservatism Media Framing and the Dynamics of Racial Policy Preferences The Fusion of Race and the Welfare State in the Public Mind A New American Dilemma for a New Millennium?

1 13 55

106 130

Appendix

142

83

References

147

Index

153

vii

CY252/Kellsted-FM

0 521 821711

June 13, 2003

viii

8:37

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

List of Figures

2.1. Number of Stories on Race in Newsweek per year, 1950–1994 2.2. Mentions of Segregation in Newsweek, 1950–1994 2.3. Egalitarian Cues in Newsweek, 1950–1994 2.4. Individualism Cues in Newsweek, 1950–1994 2.5. Mentions of the Word “Ghetto” in Newsweek, 1950–1994 2.6. Mentions of Poverty in Newsweek, 1950–1994 3.1. Four Indicators of Racial Policy Preferences, 1970–1996 3.2. Racial Policy Preferences, 1950–1996 3.3. Black and White Preferences for Busing, 1972–1996 (Uniform Axis) 3.4. Black and White Preferences for Busing, 1972–1996 (Separate Axes) 3.5. Racial and Welfare-State Policy Preferences, 1966–1996 3.6. Racial and Welfare-State Policy Preferences, 1950–1965 3.7. Racial and Welfare-State Policy Preferences, 1950–1996 4.1. Egalitarian Media Cues and Racial Policy Preferences, 1950–1994 4.2. Individualist Media Framing and Racial Policy Preferences, 1950–1994 5.1. Racial and Welfare-State Policy Preferences, 1950–1996

page 31 36 38 42 45 50 66 68 73 73 78 79 80 88 89 108 ix

CY252/Kellsted-FM

x

0 521 821711

June 13, 2003

8:37

List of Figures

5.2. Merged Public Opinion and Media Cues on “Black Poverty”: 1958–1994 5.3. Economic Equality and Public Opinion: An Alternative Causal Model 5.4. Economic Racial Equality, 1958–1994 5.5. Policy Liberalism and Public Opinion: An Alternative Causal Model

116 119 122 124

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

List of Tables

2.1. Correlations between Types of Media Framing page 52 3.1. Correlations of Indicators of Racial Policy Preferences with Overall Index 69 3.2. Correlations between White and Black Opinion Trends 74 3.3. Correlations between Year-to-Year Shifts in Racial and Welfare-State Policy Preferences for Various Time Periods 80 4.1. The Media and Public Opinion: Granger Test Results 92 4.2. The Media and Racial Policy Preferences: Preliminary Regression Analyses 93 4.3. A Multivariate Model of Racial Policy Preferences 101 5.1. The Media and the Relationship between Welfare-State and Racial Policy Preferences 117 5.2. Correlations of Indicators of Racial Equality with Overall Index 121 5.3. The Media and the Relationship between Welfare-State and Racial Policy Preferences: Controlling for Economic Equality 122 5.4. The Media and the Relationship between Welfare-State and Racial Policy Preferences: Controlling for Policy Liberalism 125

xi

CY252/Kellsted-FM

0 521 821711

June 13, 2003

xii

8:37

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

Preface

The kernel of insight that eventually led to my doctoral dissertation – and, finally, to this book – occurred in my apartment in Minneapolis on a cold winter day in 1994. As a young graduate student, I was fortunate to be taking seminars at the time when two pathbreaking works in the study of public opinion were first being published: John Zaller’s The Nature and Origins of Mass Opinion and James Stimson’s Public Opinion in America. Although these books were both immediately hailed as important new works, and both spoke in a fresh voice about the nature of public opinion, they were actually quite disconnected from one another, and remain so today. This, in my view, is unfortunate. The reason for this disconnection is that they approached the subject from very different perspectives – Zaller from the more traditional individual level, and Stimson from the quite different macro, timeseries level. But why, I wondered, weren’t these two insights connected? Didn’t Zaller’s theory of the survey response have aggregate, over-time implications? Weren’t these implications testable? Couldn’t the tools and methodologies that Stimson pioneered be used in the analysis? In the strictest intellectual sense of the word, I am indebted to Zaller and Stimson for showing the rest of us new ways to study public opinion. Without both of their works, mine would not be possible. My debt to Stimson goes beyond this, however. As a student of his at the University of Iowa and the University of Minnesota, I had grown accustomed to approaching all political questions from his longitudinal point of view. On that day in my apartment, I did what I usually did xiii

CY252/Kellsted-FM

xiv

0 521 821711

June 13, 2003

8:37

Preface

when I had an idea – I tried to talk to Jim immediately. I reached him on the phone at his home across town and relayed the core of the idea, unsure of whether it would excite him at all. I remember Jim’s reaction exactly: “Start writing.” He meant it quite literally. It was vintage Stimson in every sense of the word (including the brevity). Most important, it was intellectually energizing. No one who has ever worked with Jim will find any of this at all surprising. I relay the story as a bit of my own intellectual history, and to acknowledge the enormous debt I owe him. He is welcome to claim as much of the credit for this work as he wants. Knowing Jim, he wouldn’t dream of taking any of the credit, though – and not because it might not be up to his high standards. That’s just the kind of advisor and friend he has been, and I’m greatly in his debt. Others who helped shape my work in its embryonic stages include other members of my dissertation committee at Minnesota: John Freeman, John Sullivan, David Fan, and Antonio Merlo. Fan, in particular, patiently taught me the ins and outs of content analysis, especially the rigors of creating a dictionary of ideas. Freeman offered tireless encouragement as well as the perspective of someone not specializing in American politics but engrossed by any idea that was scientifically testable. Also at Minnesota, Wendy Rahn offered a number of challenging and constructive criticisms. The dean of the faculty at Brown University, Kathryn Spoehr, generously provided me with a sabbatical leave during the 1999–2000 academic year, without which this book would not have been possible. I spent part of that year as a Fellow at the Joan Shorenstein Center for the Press, Politics, and Public Policy at the John F. Kennedy School of Government at Harvard University. The staff, faculty, and my fellow Fellows provided an environment rich in ideas and caffeine that led to a significant reshaping of this project. Edie Holway of the center did everything possible to make my life as comfortable and productive as possible. I thank Rich Morin, B. J. Bullert, Christina Holtz-Bacha, Jonathan Mirsky, Richard Parker, Anna Greenberg, Tami Buhr, and particularly Lance Bennett, Tim Cook, Taeku Lee, and Tom Patterson for their encouragement and stimulating conversations. This book is a better product for their input.

CY252/Kellsted-FM

Preface

0 521 821711

June 13, 2003

8:37

xv

Various colleagues and friends have read a portion or, in the case of a generous few, all of this manuscript in its various stages. This list includes John Transue, Eric Lawrence, Bud Kellstedt, Christina Wolbrecht, Suzie DeBoef, Don Green, and Claudio Cioffi-Revilla. I had the opportunity to present various portions of this work at several seminars, which led to feedback that helped correct several mistakes I had made and clarify portions that were vague. Thank you to Nicole Krassas and her colleagues at Eastern Connecticut State University, to the Shorenstein Center at the Kennedy School of Government at Harvard, and to the Taubman Center for Public Policy and American Institutions at Brown University for lively discussions and for allowing me to present unfinished work. Also at Brown, I feel indebted to several years’ worth of students in my undergraduate classes on public opinion and American democracy. In their active listening to and engagement with my lectures on the nature of racial attitudes, they helped to clarify my thinking considerably. In particular, the geological metaphor that I use in the concluding chapter is the product of discussions that took place in class. I am certain that my teaching has benefited my research; whether my research has benefited my teaching will have to be left up to my students to decide. None of this work would have been possible without the considerable help of numerous research assistants who helped me amass the data used in this book. At the University of Minnesota, Greg Belshe, Rowzat Shipchandler, and Ethan Cherin all helped with the scanning of Newsweek archives. At Brown University, Jackie Delamatre, Vera Brunner-Sung, Stefanie McGowan, and especially David Primo helped to sift through the media data to help me make sense of it. Those media data were collected with the aid of a grant from the National Science Foundation (SBR-9423125), for which I am grateful. Thanks to the American Journal of Political Science, which granted permission to revise my article “Media Framing and the Dynamics of Racial Policy Preferences” and include it here as Chapter 4. On a more personal level, I feel grateful for all of the support I received in the writing of this book from my closest friends and family members. My friends at the Evangelical Covenant Church in Riverside, Rhode Island, provided unceasing encouragement. They allowed me to be real with my feelings about how the book was going – celebrating the

CY252/Kellsted-FM

xvi

0 521 821711

June 13, 2003

8:37

Preface

joys and helping me deal with the rough spots – without ever making me feel that their love for me depended on the book’s success. The number of individuals that comes to mind here is simply too big to list them all, which goes a long way in showing what a wonderful, healing community I am lucky to be a part of. They have been instruments of God’s grace to me. My familial debts are substantial. My father, Lyman (Bud) Kellstedt, a just-retired member of my own profession, read every page of this book and provided by far the least unbiased feedback I have ever received. Try as he did to be an objective critic, he failed miserably at this task and instead turned into a relentless cheerleader. The good news is that I needed him to be the cheerleader, and for succeeding splendidly in that role I thank him profoundly. But the largest familial debts I owe are to my wife, Deb, and our two daughters, Abigail and Elizabeth. On days when things didn’t go so well for me, I could always, always count on walking through the front door to the joyous cries of “Daddy’s home!” and tight hugs around my neck. I would never have had the energy to finish this book without them. I have been forever changed by the unconditional nature of their love. I dedicate this book to my Kellstedt women with all of the love I can give back.

CY252/Kellsted-FM

0 521 821711

June 13, 2003

8:37

The Mass Media and the Dynamics of American Racial Attitudes

xvii

CY252/Kellsted-FM

0 521 821711

June 13, 2003

xviii

8:37

0521821711c01

0 521 821711

June 2, 2003

7:56

1 Toward a Dynamic Perspective on Racial Attitudes

On the question of the social integration of the races our people draw the line. No decent and self-respecting Negro would ask for a law to force people to accept him where he is not wanted. They themselves do not want social intermingling. They are entitled to equality of opportunity, and they will get it through our efforts. But all the laws of Washington and all the bayonets of the army cannot force the Negro into our homes, our schools, our churches, and our places of recreation. – Governor J. Strom Thurmond, D-SC, 1948 (quoted in Cohodas, 1993) When I was governor, the laws said the races should be separated. But now the law is different, customs are different, public opinion has changed, and it’s an entirely different situation. – Senator J. Strom Thurmond, R-SC, 1978 (quoted in Cohodas, 1993)

The year is 1950. The place, the post–World War II South. Southern society is completely divided along racial lines. In fact, the notion that a single “southern society” even exists strains credibility. Instead, there are two southern societies – one white and one black. Black children and white children do not attend the same schools. Blacks and whites do not drink from the same water fountains or eat at the same restaurants. They do not buy groceries at the same stores. They do not participate in social or recreational activities together. Only a few short 1

0521821711c01

2

0 521 821711

June 2, 2003

7:56

Mass Media and American Racial Attitudes

years before, they did not even fight in the same units in the armed forces. They do not participate equally in the selection of the government officials who preside over both groups. Perhaps most ironically, though they worship the same God, cherish the same scriptures, and cling to the same cross, they do not attend the same churches. These societal divisions were perfect reflections of American public opinion on race and civil rights, which had a distinctly segregationist, nonegalitarian flavor to it. Indeed, a 1942 survey revealed that only 41 percent of all Americans believed that black and white soldiers should serve together in the armed forces. Only 55 percent agreed that a black man could be just as good a soldier as a white man. Only 42 percent agreed that “Negroes are as intelligent as white people.” The perceived differences between blacks and whites are hard to exaggerate, as only 36 percent believed that “Negro blood is the same as white blood.” In addition to these strikingly prejudicial sentiments, the bulk of white Americans preferred to maintain distance between the races. Fifty-one percent claimed that streetcars and buses should be segregated. Fifty-five percent favored active job discrimination. Sixty-eight percent believed that white and black children should go to segregated schools. When it came to restaurants, 69 percent favored separate facilities, and fully 84 percent wanted segregated neighborhoods. This complex of attitudes existed in an atmosphere of denial that race was even a social problem worth addressing, as three Americans in five said that “Negroes are getting all the opportunities they deserve in this country.”1 Indeed, southern society in 1950 was not so different from southern society in 1865. Eighty-five years after the Thirteenth Amendment to the Constitution ended slavery, things had not changed all that much for blacks in America. Blacks were still concentrated disproportionately in the South. There, blacks still tended to live in rural rather than urban areas. In absolute terms as well as in comparison to whites, blacks remained poor. In spite of the language of the Emancipation Proclamation, blacks were considered subhuman by most whites, and the society’s laws and customs reflected this derision. Discrimination against blacks was more than tolerated; in many cases the law required 1

These examples all come from surveys conducted between 1942 and 1946, and are reprinted in Cantril 1951.

0521821711c01

0 521 821711

June 2, 2003

7:56

Toward a Dynamic Perspective on Racial Attitudes

3

it. For example, it was against the law in most southern states for blacks and whites to attend the same public schools and universities. In most places in the South, blacks were not allowed to participate in electoral politics. In nearly every case where the interests of blacks and whites came into conflict, the scales of justice were unfairly tilted in favor of whites. The testimony of blacks in court was typically discounted simply because whites controlled the justice system. Eighty years of freedom had not produced equality of any sort – not equality before the law, not economic equality, and, as the survey results just reported confirm, certainly not equality in the eyes of the American public. And eighty years of freedom had not even produced much movement toward those goals. Although slavery was gone, blacks were in almost every respect second-class citizens. The year is 2000. Southern society is not as distinctive as it was a halfcentury ago. Many blacks have migrated to the northern cities in search of better-paying jobs. Many whites from the North have migrated to the South, and the formerly agrarian-dominated region is now one of the centers of America’s economic growth. Beliefs about the innate inferiority of blacks are now dismissed as racist – witness the recent reaction to the controversial book The Bell Curve, which purported to show racial differences in intelligence but was quickly labeled pseudoscholarship thinly masking a racist political agenda. Public policy debates that focus on the role of blacks in society have evolved considerably; whereas in the past the debates revolved around whether discrimination against blacks should be allowed, the debate today centers on the question of whether discrimination in favor of blacks (in the form of affirmative-action policies) should be allowed. And it is not just the policy debates that have changed; government policies themselves have changed. The Civil Rights Act of 1964, the Voting Rights Act of 1965, and the Fair Housing statutes enacted in 1968 have given blacks the legal equality they formerly lacked. Legal segregation is a thing of the past, presumably never to return. The United States Supreme Court has made this clear in decision after decision. Laws and policies that discriminate on the basis of race are deemed “suspect” by the court and demand “strict scrutiny.” Laws that either explicitly or in effect forbid black political participation have given way to full black participation in the polity, and today blacks actually participate at higher rates than

0521821711c01

4

0 521 821711

June 2, 2003

7:56

Mass Media and American Racial Attitudes

whites, once socioeconomic status is controlled for.2 Although blacks are still disproportionately poor compared to whites, the gap has narrowed. And the expansion of the welfare state has provided a safety net for blacks (as well as whites) who cannot yet achieve the American Dream on their own. These shifts, unsurprisingly, have been accompanied by – or, in some cases, caused by – a sea change in American public opinion. The percentage of Americans favoring integrated schools rose from 32 percent in 1942 to 96 percent in 1995. The last time a pollster asked about segregated streetcars and buses, in 1970, fully 88 percent of Americans favored integration. The percentage of people favoring job discrimination has revealed a similarly dramatic movement: a 1972 survey showed that 97 percent preferred giving blacks an equal chance to get good jobs. Over the past half-century, clearly, a wave of egalitarian, antidiscriminatory sentiment – at least in principle – has swept over the American public.3 What happened in the last half-century that did not occur in the eighty or so years before that? To be sure, it is hard to exaggerate the profound change in American racial politics over the past fifty years, and this change is all the more remarkable given the equally profound lack of change over the previous eighty years. It is obvious that these societal shifts have occurred on several levels. The first level is that of public opinion. White society did not accept blacks as full partners in the American experiment in 1950, but at the turn of the millennium (with few exceptions) it does. In addition, the public’s preferences have evolved with respect to the proper role of government in ensuring that blacks have access to the American Dream. Whereas whites formerly viewed government as a mechanism to insulate white society from blacks, today government is seen, at least in the minds of many, as a tool to protect minority interests from discrimination. But perhaps the most significant change in American racial politics has occurred at the level of the national debate itself. In 1950, 2

3

See Verba, Schlozman, and Brady 1996. It is true that blacks vote less frequently than whites and donate less money to campaigns than whites, among other forms of participation. This masks the fact that these differences are more than completely accounted for by the different socioeconomic standings of blacks and whites today. For example, middle-income blacks are more likely to vote than middle-income whites. These examples are from Schuman et al. 1997.

0521821711c01

0 521 821711

June 2, 2003

7:56

Toward a Dynamic Perspective on Racial Attitudes

5

American society was struggling with the concept of the innate worthiness of blacks. By the mid-1950s, with the Brown v. Board of Education decision serving as the catalyst, the debate shifted to the issue of government actively discriminating against blacks by enforcing segregation. (Is “separate but equal” a good thing? Is it consistent with the Constitution? Can it be implemented?) In the 1960s, the debate shifted again, to the issue of governmental protection of blacks from discrimination in the private sector (such as public accommodations, employment, and housing). In the 1970s, the public discussion focused on the legitimacy of activist government policies like busing to end segregation in schools. And in the most recent period, debate has centered on the issue of affirmative-action policies that give preferences to blacks and other minorities in an effort to compensate for past discrimination or to ensure diversity in the workplace.

blind spots: what don’t we know? The scholarly literature on race in American politics is enormous, encompassing analyses that bridge such disciplines as economics, sociology, political science, public policy, and cultural studies. But the literature on race, vast though it is, has not produced much in the forms of theory or evidence on the role that the mass media have played in the evolution of race politics (or, more specifically, in public opinion) in the twentieth century. Our understanding of how the mainstream press has covered race – and especially how that coverage has evolved in anticipation of, in the middle of, and in the wake of the civil rights movement – is practically nonexistent. This despite the fact that nearly every scholar who has studied the politics of race would concede that the media have helped shape the course of race politics. On the face of it, few would deny that press coverage was critical to the unfolding civil rights drama of the 1950s and 1960s, for example. Most of the key events of the movement, after all – desegregation battles in Little Rock and elsewhere, bus boycotts, freedom rides – took place in communities that few non-southerners would have known about were it not for the national press. And yet the role played by the media in the unfolding drama of race in America has largely been ignored by scholars.

0521821711c01

6

0 521 821711

June 2, 2003

7:56

Mass Media and American Racial Attitudes

The literature on public opinion on race in America is similarly vast. Quite literally, decades of scholarly research have focused on the important question of why some Americans are racial liberals and others racial conservatives. Two leading theories have emerged, one focusing on the role played by overarching ideology, the other on the role of prejudice.4 Scholars who have pursued the connection between ideology and racial attitudes have seen policy preferences on race, in the main, as a function of larger considerations about the proper role of government in American society. Citizens predisposed to believe that government cannot help solve society’s problems will straightforwardly oppose governmental action in the case of race, whereas citizens who believe government can be an active part of the solution will tend to favor governmental action to level the playing field between white and black Americans.5 Another set of researchers has identified some form of racial prejudice (or the absence of it) as the leading factor that causes Americans to oppose or support liberal racial policies. These studies paint a very different picture of American public opinion, one that continues to be tainted with racial prejudice. In this view, when it comes to explaining why some Americans support or oppose liberal racial policies, ideological considerations are dwarfed by prejudice – and these ideological considerations are sometimes, in fact, merely convenient and “politically correct” covers for prejudice. Those individuals who harbor racial

4

5

There are actually at least two other plausible explanations, though they are considerably less common in the academic literature. The first revolves around the concepts of group conflict and self-interest. The core idea here is that attitudes toward groups are a function of competition in the individual’s environment and arise out of self-interest and a realistic conceptualization of intergroup conflict. This dates back at least to the political scientist V. O. Key (1949). For more recent examples, see Bobo 1983 and Giles and Evans 1986. Another theory, focusing around positions of social dominance, suggests that individuals in a dominant societal position seek to maintain that dominance; hence, whites will adopt policy views consistent with their historically dominant position over blacks in a social hierarchy. This is not a function of prejudice, but of a desire to maintain a position of power. See Sidanius et al. 2000 and the citations therein. The list of scholars endorsing this view is extensive. Most prominent among them, though, is the political scientist Paul Sniderman. See, for example, Sniderman and Carmines 1997; Kuklinski et al. 1997; Sniderman and Piazza 1993; Sniderman et al. 1991; and Sniderman and Tetlock, 1986a, 1986b. Often, in this literature, the notion that ideology determines racial policy preferences comes across mainly by implication, as competing theories are dismissed as not causing racial policy attitudes. Ideology, then, tends to be the only explanation left standing.

0521821711c01

0 521 821711

June 2, 2003

7:56

Toward a Dynamic Perspective on Racial Attitudes

7

prejudice overwhelmingly oppose liberal government policies on race, and those without prejudice tend to support those same policies.6 This debate is serious, substantive, and likely to rage on for years to come. The variation that these scholars seek to explain – why some individuals are racial liberals and other individuals are racial conservatives at a given point in time – represents something that is important to society. But the intense nature of the debate obscures the fact that variation among different individuals is not the only source of variation worth explaining. In our zeal to explain the differences between liberals and conservatives, another kind of variation has been almost completely ignored. For in addition to zeroing in on variation among individuals at one point in time, as this literature has exclusively done, scholars could also ask questions focusing on variation over time. As it stands, our understanding of public opinion on race is excessively static, wrongly assuming fixed attitudes that cannot and do not change. As a scholarly community, we have very little sense of how, as a whole, American racial policy preferences have varied over time. Indeed, we have almost no sense of whether they have or have not varied at all. Has American public opinion become more liberal on matters of race? More conservative? Or has it moved in both conservative and liberal directions at different points in time? Do racial policy preferences in the aggregate, over time, move in roughly parallel fashion, with the public becoming more liberal (or conservative) on all facets of race policy simultaneously, or does each subissue in the domain of race politics have its own unique dynamics?7 The scholarly community has no answers to these questions, largely because we have never asked them due to our preoccupation with individual-level variation. In turn, we have little sense of the causal dynamics that underlie any over-time change. If American racial policy preferences have varied over time, why have they done so? What forces have made American opinion more liberal (or more conservative) on race? By focusing 6

7

Prejudice is often referred to, in these studies, as “new racism,” or “symbolic racism,” or “racial resentment.” As with those focusing on ideology, the list of scholars emphasizing the role of prejudice in explaining racial attitudes is long. It is most closely identified, however, with Donald Kinder. See, for example, Kinder and Sanders 1996; Kinder and Mendelberg 1995; Sears 1988; Kinder 1986; Sears and Citrin 1985; Kluegel and Smith 1983; Kinder and Sears 1981; and Sears and Kinder 1971. Here I have in mind the famous articles published in Scientific American by Hyman and Sheatsley (1956; 1964). For more recent work in this vein, see Schuman et al. 1997.

0521821711c01

8

0 521 821711

June 2, 2003

7:56

Mass Media and American Racial Attitudes

exclusively on variation among individuals at one point in time, the scholarly community has neglected to identify and explain variation in the nation’s racial policy preferences over time. And yet such variation is critical to understanding the ebbs and flows of race politics in America; indeed, given the centrality of race in American politics, we would not be exaggerating too much to say that identifying and explaining the over-time dynamics of racial attitudes would go a long way toward understanding all of American politics. Here the two deficiencies of the literature on race in America – that it has neglected the role of the media and that it has been preoccupied with statics, not over-time dynamics – become related. The media’s important role in shaping American racial politics can only be understood in a longitudinal framework. When we begin to appreciate this reality, we will find considerable power in explaining the over-time dynamics of American racial attitudes. This book will make a connection that has been the source of a good deal of speculation, a great deal of polemic, but virtually no systematic analysis whatever – namely, the relationship between media coverage of race and American public opinion on race. In the process, the dynamic feature of racial politics in America will be emphasized. That is, the most interesting questions about race that have not been adequately addressed by the scholarly community revolve around how and why things today are (or are not) different then they were last year or ten years ago. Not surprisingly, then, the analyses in this book will have a distinctly time-series flavor. Given the focus on how and why racial politics has evolved over the last several decades, this is the most natural methodological approach. The end result will be a greater understanding of the sea change in racial politics that America has experienced over the last half-century. The focus on over-time change will yield one other benefit as well. Although it is clear that both the national debates and public opinion have changed considerably over the past fifty years, we will also notice that most of these shifts have not been of the that-was-then-this-isnow variety. Indeed, despite the fact that the historical sketches that introduce this chapter show real change in race politics, our focus on time will prove that each of the years in between is an indispensible element of the story. Instead of viewing the changes as being akin to the differences between night and day, we will see that they are more

0521821711c01

0 521 821711

June 2, 2003

7:56

Toward a Dynamic Perspective on Racial Attitudes

9

like shifts in the tide, which rolls in, then out, and then back in again, leaving the beach a bit different each time. In a sense, this book represents a revival of the first studies that examined racial attitudes over time, a tradition begun in the 1950s by Herbert Hyman and Paul Sheatsley in their famous articles in Scientific American, and continued more recently by Howard Schuman and colleagues.8 This volume, though, benefits from recent advances in the study of public opinion – to be described in detail in Chapter 3 – that enable us to extract information from numerous opinion surveys at once, thus making a unified analysis possible. Instead of taking each survey item about every racial issue as something unique, I will strive to find communality among the various subdomains of racial politics. This new approach will empower us to tie the trends in public opinion to systematic causes like the national media in previously undiscovered ways.

a macro perspective How and why do Americans’ attitudes about race change? The illustrations that introduce this chapter show unquestionable and massive change in the span of roughly fifty years or less. These attitudinal shifts are surely, to a degree, a function of generational replacement. As older generations that were socialized in an America in which blacks were considered to be inherently inferior begin to die off and are replaced by generations that witnessed the struggles of Rosa Parks, Martin Luther King, Jr., and many others, surely the attitudes about the humanity of black people will change forever. But how, if at all, do policy preferences about race change? It is worth noting that it has traditionally been assumed that such preferences do not change at all. In his famous 1964 treatise on public opinion, Philip Converse made the case that attitudes and policy preferences about race (along with attitudes about political parties) were the only stable and meaningful elements of a typical person’s political “belief system.” Perhaps because Converse’s theory was so widely accepted, scholarly investigations into the shifts (and, implicitly, the causes of those shifts) in racial policy preferences have been rare. It has until 8

See Hyman and Sheatsley 1956, 1964 and Schuman et al. 1997.

0521821711c01

10

0 521 821711

June 2, 2003

7:56

Mass Media and American Racial Attitudes

very recently been assumed that individuals’ policy preferences on race were stable, even over the long run. Their attitudes on race, like their partisanship, were apparently a function of childhood socialization and education: they were the product of emotion, not cognition. Without trivializing the issue, one could say that the process by which a person acquires attitudes about race was, in many ways, seen as quite similar to the ways in which a person espouses a favorite baseball team. The process can essentially be boiled down to something not much more complicated than the fact of which team a person’s parents rooted for. But what if people do change their minds on race? What if the evolution of racial attitudes occur over the span of years and decades instead of over the course of generations? Perhaps still slow-moving, to be sure, and never to be confused with a fickle, moody public, what if there is evidence of shifts in racial policy preferences taking place at a rate that is too rapid to be accounted for by birth and death? What forces lead to such attitude change? And what does such change say about the politics of race in America, and indeed about American politics as a whole? The answers, I believe, lie in the fact that Americans’ attitudes about race are the product of their underlying political values – values that sometimes conflict with one another. Because different currents of the American ethos pull citizens in varying directions on the issue of race, with some core parts of their value system embracing government action while other parts simultaneously resist it, most Americans vacillate on the subject. That is, most of us see at least some truth in both sides of contemporary policy arguments about race policy. And, crucially, the American ethos is not a static, etched-in-stone body of ideas; it, too, is dynamic, with certain values becoming more prominent at some points in time, then, years or decades later, receding into the background, never eliminated from our consciousness but surely less prominent. In a sense, American attitudes on race resemble an internal tug-ofwar between cherished values that conflict with one another – a struggle where one side gains ground over a period of time but the other side never truly loses, regaining strength and pulling back the other way. But there is only one way to answer such questions definitively: go to survey marginals in search of stability and change in policy preferences on race. Once such shifts are discovered, we will need to generate entirely new theories of racial policy preferences, because the

0521821711c01

0 521 821711

June 2, 2003

7:56

Toward a Dynamic Perspective on Racial Attitudes

11

cross-sectional literature is largely predicated on the assumption of individual stability. We will draw from recent advances in the study of public opinion that focus on internal ambivalence, and will require the development of theories of political learning and information acquisition that will reveal much about the nature of public opinion itself. And, once we see the evolution in racial attitudes over time, we will hunger for an explanation for the trends we have discovered. This is where the twin gaps in the literature on race in America – not sufficiently dynamic, no explicit role for the media – will require us to develop new dynamic theories of racial attitudes that incorporate the role of the mass media. The longitudinal approach taken in this book will produce different types of questions – and therefore different types of answers – than does the more traditional approach. I am advocating a perspective that focuses instead on over-time change in aggregate sentiment on racial policy preferences. This different perspective will show how the portrait of racial attitudes in America today differs from the similar portrait of a year ago, a decade ago, or a half-century ago – and, if successful, it will explain why as well. As such, the analyses to follow will resemble a motion picture more than a snapshot.

the plan of this book In the next chapter, I describe the evolution of media coverage of race in America. There, I introduce a new data base, comprised of every story written on race in Newsweek magazine between 1950 and 1994, along with a supplementary sample of New York Times stories. This data base will allow for a comprehensive analysis of previously unanswered questions such as: How, in terms of quantity and tenor, is media coverage of race today different – and how is it similar to – such coverage in the past? Have certain themes been more prominent at some times than others? Is there a discernable pattern in why coverage shifts? Is it driven by savvy political actors who inject symbols into public debate for their own advantage, or by dramatic events in the real world, or by something else? In Chapter 3, I describe recent developments in the study of public opinion and examine their implications for racial attitudes over time in the aggregate population. I ask the questions: Have racial policy

0521821711c01

12

0 521 821711

June 2, 2003

7:56

Mass Media and American Racial Attitudes

preferences shifted in an interpretable, understandable manner, or have they moved in a chaotic and inexplicable manner over the decades, or have they perhaps been stable and inmoving? Do racial policy preferences move through time in a coherent fashion? That is, are there liberal and conservative eras in racial policy preferences? Do periods of liberalism in American opinion on one racial issue coincide with liberal periods on other racial issues? Or does each racial issue represent a truly distinct dimension, with its own path and trajectory? Is there any coherence to public opinion on race? How does public opinion on race compare to public opinion on America’s other enduring issue cluster of the twentieth century, the welfare state? The focus there will be entirely on public opinion in the domain of time – representing roughly the last half-century – on ebbs in a liberal direction and flows toward conservatism. This discussion will lead to a focus on shifts in the national dialogue on race as portrayed in the national media, and on political learning. In Chapter 4, I combine theory and data, examine the time series of racial policy preferences in America, and build empirical models that describe the ebbs and flows of liberalism and conservatism in racial opinion. What forces drive aggregate sentiment on racial policy over time? Why is the American public sometimes liberal and sometimes conservative on race? The explanation will revolve around the complementary forces of media framing, economic cycles, feedback from the policy process, and generational change, all of which contribute to the changes in racial policy preferences. Chapter 5 then brings into focus a discovery from Chapter 3 about the relationship between opinion on race and opinion on the welfare state, and uses data on the media to describe how the two great issue domains in twentieth-century America – race and the welfare state – came to be associated with one another. How has that relationship evolved over time? What causal factors have influenced this relationship? What does this tell us about American politics? In Chapter 6, I conclude with implications for democracy and the future of race politics in twenty-first-century America. Are the media manipulating public opinion? Can the national media be trusted to portray race in America accurately? To look elsewhere for culprits, if public opinion is shaped in the way I describe, can public opinion be trusted?

0521821711c02

0 521 821711

June 2, 2003

8:6

2 Eras of Media Coverage of Race

Reality is problematic not only because news stories inevitably select only some aspects of reality and leave out others. More important, over time, the specific realities depicted in single stories may accumulate to form a summary message that distorts social reality. – Robert M. Entman (1994; emphasis in original) All too often, we demonstrate an amazing ignorance of past events. – Claude Sitton, news director, New York Times (quoted in Fisher and Lowenstein, 1967)

Has there ever been a news story where context – that is, the events and causes leading up to present-day realities – is more important than it is in the case of race? Has there ever been a story where providing that context fairly and accurately is so complex and contested, so difficult? Has there ever been a story where, if the context is not provided, the dangers of misrepresenting reality are greater? And yet, the most frequent criticism leveled against the news media, beginning at least with the famous Kerner Commission Report, is that coverage of race has failed to provide the consumer of mass media with the proper context.1 This must be an exceedingly difficult subject for journalists, for in the news business the antonym of “news” might be “context.” Clearly, though, journalists are in the news business, not the 1

See Report of the National Advisory Commission on Civil Disorders 1968.

13

0521821711c02

0 521 821711

June 2, 2003

14

8:6

Mass Media and American Racial Attitudes

context business. In the case of race, “context,” taken to its extreme, would imply attention to social and historical forces such as the slave trade, a civil war, regional isolation and then migration, and a social movement designed to win political equality. This is a tall order, more fit for an academic tome than for a 90-second evening-news segment or a 750-word newspaper article. And in most cases, of course, it is exceedingly unrealistic. The overarching assumption that will be made in this chapter (and tested in subsequent ones) is that both the quantity and the tenor of media coverage matter. They matter, primarily, because the media have the potential to affect public opinion in a variety of ways. The media are widely believed to possess agenda-setting power, and the power to make certain considerations more salient in the public’s political evaluations, an effect known as “priming.”2 Moreover, the news media are a powerful tool for social learning. This is true generally, and particularly so in the case of race. Americans learn about racial groups – in particular, what members of other racial groups “are like” – from a variety of sources. Certainly, personal experience and interactions are one such source, which is one reason why neighborhood and school integration has historically been such an important goal of the civil rights movement. However, there is evidence to suggest that the power of personal experience has some severe limitations with respect to learning about racial groups. For example, the personal-interaction hypothesis would lead one naturally to expect that individuals who live in racially diverse environments are less likely to endorse exaggerated stereotypes of African Americans. Living in a racially diverse area, the reasoning goes, enables one to have more encounters with people of different racial backgrounds, which leads to a more realistic and nuanced understanding of other people, and, as a result, a reduced likelihood of endorsing uninformed and stereotypical judgments about members of other groups. It could provide, to paraphrase Martin Luther King, Jr., the opportunity to judge people by the content of their character rather than by the color of their skin. Straightforward as this perspective seems, there is not much evidence to support it; indeed, some evidence contradicts it. For example, Martin Gilens cites a 1991 national survey that revealed there is 2

See Iyengar and Kinder 1987.

0521821711c02

0 521 821711

June 2, 2003

8:6

Eras of Media Coverage of Race

15

virtually no relationship between the racial diversity of respondents’ state of residence and their estimate of the percentage of the nation’s poor who are black. According to 1998 Census Bureau figures, that percentage has fluctuated in a relatively narrow range over the past several decades, hovering within a point or two of 30 percent. But, according to the 1991 survey, the median response to the question “What percent of all the poor people in this country would you say are black?” was 50 percent. Significantly, this varied relatively little from state to state, with relatively diverse states such as Pennsylvania and Michigan giving similar aggregate responses (50 percent, matching the national average) to relatively white states such as Montana and the Dakotas (47 percent).3 If diverse personal experiences do not produce diverse endorsements of stereotypical judgments, then from what sources do people learn about members of other racial or socioeconomic groups? The obvious suggestion, if only by the process of elimination, is the national news media. This suggests the central question that will be addressed in this chapter: When Americans consume information from the news media, what messages are they receiving about what blacks are like? And, just as important, has the nature of those messages changed noticeably over the last half-century? In this chapter, I describe the quantity and nature of media coverage of racial issues during that period. What themes emerge from this coverage? Are these themes given a relatively constant amount of coverage over the years, or does the type and tone of coverage fluctuate?

previous studies: insights and limitations If the number of studies of American public opinion on race were a raging river, then the comparable number of academic studies of American media coverage of race would be a paltry trickle. Indeed, in the domain of public-opinion scholarship, there are competing camps and theories of cause and effect, whereas in that of media coverage there are few systematic and ambitious scholarly efforts at all. That said, at least one prominent issue has emerged from the piecemeal studies on race and the media: the prominence of stereotypical portrayals of blacks. 3

See Gilens 1996, p. 530.

0521821711c02

16

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

The earliest of these studies were conducted in the 1950s. Perhaps more surprising than the conclusions, at the time, that blacks were portrayed in stereotypical and negative fashion is the fact that scholars noticed that black Americans were so portrayed. One of the first examples of studies of media portrayals of blacks was conducted by Audrey Shuey and her colleagues and published in 1953. Interestingly, it examined photographs, not stories, of blacks in six national magazines, scouring the photographs in both advertisements and news stories. The authors discovered that, overwhelmingly, black Americans were stereotypically depicted as “servants” in the advertisements and as “primitives, entertainers, and sportsmen” in the news articles.4 Of course, this was 1953. The idea that black Americans were portrayed as “servants” in advertisements then cannot be surprising to anyone who has cracked open a magazine from the period. Such a reader will invariably be confronted with pictures of, say, a black butler in a tuxedo serving a cocktail or a box of Chesterfield cigarettes to a white man in an easy chair. But at the end of the 1960s, the study by Shuey and her colleagues was updated by Keith Cox, who examined the advertisements in the same six magazines in 1967 and 1968. He found that the portrayal of blacks in below-skilled-labor positions had declined dramatically since the original study. Furthermore, blacks were increasingly shown in magazine advertisements in high-status positions, far exceeding the rates from the early 1950s. The stereotype of blacks as servants was apparently on the decline.5 But, as was to become characteristic of this literature, just as one scholar interprets changes positively, another comes along with the opposite spin. The challengers in this case were J. David Colfax and Susan Frankel Sternberg, who argued that Cox was far too charitable to the advertising industry in his interpretations. In particular, they criticized his definition of a “high-status” position, claiming that many of the positions Cox considered to be high status were really just new (and sometimes quite negative) stereotypes of black Americans. They wrote: Indeed, if the advertising image is to be believed, the black is a record star, an entertainer, a celebrity; if he is not one of these, he is a child, a woman, or a 4 5

See Shuey, King, and Griffith 1953. The six magazines examined were Time, Life, The Saturday Evening Post, The New Yorker, Ladies’ Home Journal, and Colliers. See Cox 1969; but for a different view, see Kassarjian 1969.

0521821711c02

0 521 821711

June 2, 2003

8:6

Eras of Media Coverage of Race

17

foreigner. As a male, he is in need of public or private charity, and he seldom if ever enjoys the occupational status of the whites with whom he is depicted. Missing from these ads are black families and black males, at work and at leisure – in short, the black American, rather than the black stereotype.6

As it turns out, whether the glass is half full or half empty depends on who’s looking at it. The subsequent literature documents this half-full-but-half-empty progress (and absence of progress) repeatedly. Whereas a more recent look at the images of blacks in national magazines by Paul Lester and Ron Smith presents the optimistic finding that stereotypical coverage of blacks is on the wane, a thorough review of the literature by Paula Poindexter and Carolyn Stroman shows quite the opposite. Focusing on television, they conclude that stereotypes of blacks continue to abound on television in the news and advertisements as well as in entertainment programming.7 The search for stereotypes has led scholars into some rather esoteric corners of the media universe. Raymond Rainville and Edward McCormick, for example, compared the treatment that black and white professional football players have received from announcers and color commentators. Football, until very recently, was a game where stereotypes flourished, especially about the intelligence of blacks and whites – which is why quarterbacks were almost always white. Rainville and McCormick’s analysis of the play-by-play of sixteen games revealed that, indeed, black players got harsher verbal treatment after making a bad play, and less praise for a good one.8 And, in recent work, White and Fuentes looked for, and found, stereotypical presentations of blacks in the cartoon section of papers, although they concluded that stereotypical depictions were declining.9 Even in these more remote segments of the media world, scholars have found evidence of negative stereotyping of black Americans. The most comprehensive study of press coverage of blacks that focuses on stereotyping was Carolyn Martindale’s 1986 book The White Press and Black America. She examined a small sample of 6 7 8 9

See Colfax and Sternberg 1972, p. 17. See Lester and Smith 1990 and Poindexter and Stroman 1981. See Rainville and McCormick 1977. See White and Fuentes 1997.

0521821711c02

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

18

selected articles from four newspapers – the Atlanta Constitution, the Boston Globe, the Chicago Tribune, and the New York Times – over the span of several decades. Stereotypical coverage of blacks, especially coverage that highlighted criminal behavior, was found in abundance, exceeding stories reporting on the everyday lives of blacks. Although, over the period of her study, Martindale claimed that progress had been made in addressing stereotypical reporting, the problem had far from vanished. Again, the glass is half full and half empty.10 As the above demonstrates, the vast majority of the existing literature has shown that media coverage of blacks in America focuses disproportionately on stories that confirm preexisting (and almost always negative) white stereotypes of blacks. Coverage, then, generally serves to reinforce, rather than challenge, stereotypes. If a typical media consumer holds unflattering beliefs about what black people are like, then he or she is not likely to encounter much information that disconfirms those beliefs; on the contrary, he or she will likely find much that conforms to preexisting beliefs. Many of these stories paint a picture of blacks as a “problem” people – with an accompanying implication that they are at the very core of postwar America’s defining social problems, such as the erosion of the traditional family, rising violent crime, urban decay, and the increase in illegal drug use. Some consumers of mass media may infer that blacks are to blame for these problems. At the least, it seems likely that repeated coverage that emphasizes, say, the rapid increase in teenage motherhood among young black women can do little to disabuse readers and viewers of such notions of black responsibility for social problems. Certainly, this presents something of a dilemma for the producers of news. In fact, it is true that blacks are indeed disproportionately involved in many of America’s pressing social problems. Black teenagers are more likely to become pregnant than white teenagers; blacks are more likely to be convicted of a crime than whites; blacks are disproportionately poor; and so on. Combine these facts with the mainstream media’s standard script for a story, which typically discusses a problem by using a narrative, story-telling method in examining the life of the person concerned – often, in the cases of social problems, a black person who has been afflicted with a problem – and 10

See Martindale 1986.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

19

the promulgation of stereotypes becomes more likely.11 Contrast this with the less commonly employed analytical approach to reporting such a story, which might describe the problem in abstract terms and be more likely to discuss its possible causes and solutions. In the case of teenage pregnancy, for example, these two treatments would produce quite divergent stories. The narrative approach might introduce the audience to a poor, inner-city black teenaged girl, startlingly young, who became pregnant. It would highlight her personal tragedy – how the pregnancy affects her status in school, her ability to hold down a job, her future, and the like. This is the type of story that most analysts readily identify as typical of network television news. The analytical story, in contrast, might involve quotations from experts – scholars or social workers – who are familiar with the problem of teenage pregnancy and can discuss its causes in abstract terms, with some idea of what policy options might be available to remedy the situation. This type of story does not seem typical of the mainstream media but would be more likely to show up in a fifteen-minute segment on PBS’s The News Hour with Jim Lehrer. The narrative frame of reporting on social problems presents an opportunity, depending on the characteristics of the person who is the subject of the story (such as, whether the person is black or white), either to confirm or refute preexisting stereotypes. To continue with the teenage pregnancy example, when the media tell a narrative story by choosing a black, inner-city girl as the subject they are confirming the stereotype that teen pregnancy is a black problem as well as an urban problem. The media could tell the story differently by choosing a wealthy white suburbanite as the subject of their story, which in this case would refute a stereotype. Of course, occasionally such stereotype-disconfirming stories do appear in the media. But the literature that exists on the issue is nearly unanimous in its finding that such stories are the exception, not the rule. The media are far more likely to tell stories in ways that confirm, rather than refute, the preconceptions of their viewers and readers. This much we know, and it is easy to see how the press might foster stereotypical beliefs among the public. There is much more we do not know about media coverage of black Americans. Beyond stereotyping, we know precious little about 11

See Gans 1979.

0521821711c02

20

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

either the quantity or the content of press coverage of race. To be sure, it is valuable to recognize that stereotypes have existed for a long time, and that they are still around. Stereotypes may have particular political consequences, though it is critical to note that no study to date has even attempted a serious investigation of what those political consequences might be. But other facets of media coverage might also have consequences for the dynamics of race politics in America. This must be true. The only question is: what kind of coverage? The literature provides precious little guidance on this front. At their worst, the existing academic studies of media coverage of black Americans collectively suffer from five serious deficiencies. First, the literature is insufficiently based on real media data and too often on mere speculation about what coverage actually has shown. Most of the studies that do look at actual data select extremely small samples of articles or issues to examine, and quite often those articles are not selected randomly. This means that the examined data run the risk of being unrepresentative. The second shortcoming of the literature is perhaps a consequence of the lack of credible data: the literature is excessively normative. There is a rightful place for normative studies, of course. Normative prescriptions, though, ought to flow from solid, empirically based descriptions. Because these descriptions are lacking, the literature has a hand-wringing, self-flagellating tinge to it – and this is particularly true of the bulk of the literature that dwells on stereotyping. Too often, the conclusions seem to amount to the general conclusion that stereotyping is bad, and that there ought to be less of it. Third, one of the criticisms of media coverage of race that appears in the literature – an exclusive focus on “crisis” events – is typified by the following comment in the published proceedings of a conference convened on the issue of race and the media: “What is not a crisis is usually not reported. . . . The news media . . . disregard the problems that seethe beneath the surface until they erupt in the hot steam that is a ‘live’ news story.”12 Such a criticism is undeniably true. Yet, ironically, the same accusation can be leveled at scholarly attention to media coverage of race: academic researchers have been disproportionately obsessed with coverage of “crisis” issues and, in the process, have

12

Fisher and Lowenstein 1967, p. 4.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

21

ignored day-to-day coverage of race. This is implicitly to assert that media consumers really only listen, watch, and learn when there is a crisis event. The trouble with this position is that it is assumed, not proven. And I suspect it is wrong. Every day when we read the morning paper or watch the evening news, we learn. Fourth, far too many studies of press coverage are static, focusing exclusively on data from a single point in time. Most studies that look at actual coverage do not make any over-time comparisons at all. At best, they examine two time periods, perhaps separated by a decade or more, for a “before and after” glance. These minimal (and perhaps unrepresentative) samples of data are not truly dynamic analyses, where scholars systematically and deliberately incorporate the element of time into their comparisons. Thus, the scholarly community is left with little credible knowledge of whether the stereotypical themes have been covered at roughly the same rates over the past several decades. In the absence of such analyses, it has been assumed that coverage has been relatively constant. Is this assumption warranted? For example, we are told that the media stereotype blacks as poor. Has it always been the case? Or are such portrayals more recent phenomena? If these stereotypes have not always been prominent features of media coverage of race, when did they appear? Did they appear suddenly or gradually? To date, there are no answers to questions such as these. Fifth, but perhaps most critically, the scholarly literature has provided no clear evidence that media coverage of race actually affects public opinion in any systematic way. Of course, this is an underlying assumption of research in this area – after all, if what the press says doesn’t affect those who read, watch, or listen to it, then why study it? Despite the universality of this assumption, there is no evidence to substantiate these connections. For example, there are no studies documenting that those who consume what the media say about race believe anything different than those who are not media consumers. Nor are there any studies connecting shifts in media coverage of race – in its tenor, its emphases, or the like – to any shifts in public opinion over time.13 In a scientific sense, then, media coverage exists as a 13

And this stands in contrast to other areas of public opinion, such as presidential popularity, where explicit connections between media priming and framing have been shown to alter public opinion. See Iyengar and Kinder 1987 and Krosnick and Kinder 1990.

0521821711c02

22

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

presumed cause without any demonstrated consequences and as a presumed consequence without any known causes. In the rest of this chapter, I will outline a strategy to correct for these five shortcomings. It will be based on a systematic and large universe of stories in the national press. Moreover, it will focus on an entire half-century of coverage, observing both the crisis coverage as well as the messages conveyed in day-to-day stories about black Americans. In later chapters, this approach will – for the first time – allow for systematic hypotheses about both the origins of press coverage and the effects it has on public opinion.

a dynamic approach to media frames Imagine for a moment that you are a scholar interested in the analysis of media coverage of race in America. Imagine further that you are unencumbered by the understandable problems that plague the literature – principally, the absence of a comprehensive data base of media stories on race over an extended period of time. In this hypothetical world, you have at your disposal all of the media data conceivable – presumably of the textual, audio, and video varieties – and that, furthermore, they are in readily analyzable fashion. The only limits are your creative ability to ask (and answer) interesting questions. Where to begin? Like the kid in the candy store who thinks that everything looks delicious, you somehow have to decide what to try first. To what questions would you first want answers? For what concepts would you first want to develop measures? What kind of theory about cause and effect would you first want to answer? I will pursue three main avenues of inquiry with the media data in this book. First, and at the most basic level, I believe it is important to document the volume of coverage that has been dedicated to racial issues in America. How much attention has been paid to racial issues in the mainstream press, and how has this level of attention varied over time? Presumably, media coverage peaked in the mid-1960s, at the height of the legislative push for civil rights. Did coverage increase dramatically then, or was it already at a high level before that? And how quickly, if at all, did it dwindle afterward? In the decades since the 1960s, have there been periods of resurgent interest in race politics on the part of the mainstream press? Metaphorically, when people

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

23

listen to news about race, how loud is it in comparison to other salient issues? The second major area of media coverage of race that I will focus on pertains to echoes of core American values. If race is “the” American issue, then it should not be surprising to find references – most often, subtle ones – to core American beliefs in news stories of race. In particular, for reasons relating to values that are prominent strands of American public opinion, I will search for mentions of two core American beliefs: individualism and egalitarianism. Because this book revolves around the over-time evolution of public opinion, facets of coverage that are directly related to key features of racial policy preferences – like values – seem especially relevant. That is, individualist and egalitarian frames in press coverage of race may be central to public opinion on race, particularly if racial policy preferences have their basis in these core values.14 In Chapter 3, I argue that this is precisely the nature of racial policy preferences. Therefore, to the degree that opinion has its basis in values, tracking coverage that resonates with these values becomes important. The final area of press coverage of race that strikes me as immediately relevant also has its origin in public opinion – namely, coverage that highlights the disproportionate poverty evident in the black community. Recent years have witnessed a proliferation of scholarship focusing on the relationship between attitudes about race and those about the welfare state.15 But how much press coverage of race has even mentioned – to say nothing of emphasized – the real-world connection between race and poverty? What, when, and how much has the American public been learning about this connection? This is my scholarly wish list. And, as noted, my idiosyncratic curiosities are surely a function of my interest in public opinion. That is, my interest is drawn to those areas of media coverage which are most likely to have a connection to American public opinion. The list of worthy topics a scholar might analyze – but which, in this book, I will not – is far longer. With a wealth of media data at hand, in-depth 14 15

The word frames, in reference to the nature of media portrayals of an issue, comes from Gamson and Modigliani 1966. Much of the aforementioned literature on racial attitudes touches on this relationship obliquely, but for direct references see Abramowitz 1994, Gilens 1999, 1996, 1995, Carmines and Stimson 1989, and Stimson 1999.

0521821711c02

24

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

analysis of particular movement leaders and personalities becomes possible. Tracing, for example, the political history of Martin Luther King, Jr., as depicted in the national media – from his emergence in the Southern Christian Leadership Conference through the peak of his influence in the march on Washington to his assassination in Memphis in 1968 and even to his posthumous legacy – becomes possible. So, too, does an analysis of King’s contemporary and rival for the hearts of followers, Malcolm X, or an analysis of how Lyndon Johnson, once a senator from the Deep South, became the champion of the most sweeping pieces of civil rights legislation of the century. The portrayals of the decades-long careers of George Wallace and Jesse Jackson also have obvious interest. So, too, would an analysis of particular political issues, and how they were covered. Government policies like desegregation, school busing, and affirmative action all have long and contentious histories, and newer racial issues such as racial profiling continue to emerge. How were such issues first portrayed? Did the frames originate from the pronouncements of political elites engaged in the struggle or from somewhere else? How did coverage of these issues evolve? Did the coverage shift because of changes in government policy, or because of the dynamics of social movements and countermovements? For an issue like desegregation, which is no longer prominent on the political agenda, how did coverage eventually wane and die out? With an inexhaustible supply of media data, all-encompassing analysis of the civil rights movement as an example of a social movement, and how it was covered, becomes possible.16 Analysts could document press coverage of the movement’s origins as it grew in pockets of the rural South and coalesced into a broader social movement, eventually crossing regional borders and becoming less and less a “southern problem” and more and more America’s defining problem. How, for example, did the press document the history of repression and backlash that had the whole world watching in the 1950s and 1960s? If social movements are defined by conflict between societal groups with grievances and governments, then the national press has a distinct role in portraying the origins, processes, and outcomes of that 16

Chong 1991, See McAdam 1982.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

25

conflict. This possibility is highlighted by eminent social movement scholar Sidney Tarrow, who writes: “Only in the modern world – when public opinion and national states began to mediate between claim makers and their targets – has contention become a true performance for the benefit of third parties.”17 Although Tarrow does not make this connection, such movements, including the civil rights movement, only have their third-party audience because of the modern media.18 Sociologist Bert Klandermans echoes this sentiment, writing that “The transformation of social issues into collective action frames does not occur by itself. It is a process in which social actors, media and members of a society jointly interpret, define and redefine states of affairs.”19 Several potentially fertile analyses spring to mind, including the possibility of tracking the use of biblical analogies by movement leaders – such as “the Promised Land” – in coverage of the civil rights movement. I can envision studies that would seek to document the coverage of black Americans and their involvement in social problems like crime and the erosion of the two-parent family. Coverage like this is rich in stereotype-confirming content; and a good amount of scholarship suggests that the image of black people as criminals can have powerful political meaning.20 But, as I noted above, scholars have not yet provided a history of such coverage, much less any theories about how it came into being. Finally, a wealth of media data on race would make possible studies of the use of visceral, emotion-laden symbols, and how, over time, they have been used by movement leaders. For instance, I have in mind a detailed breakdown of how terms such as “civil rights” or “discrimination” came to be used in media stories. In the abstract, civil rights is a good thing and discrimination a bad one. The latter concept, in 17 18

19 20

See Tarrow 1998, p. 94. Surely one of the least investigated elements of the origins and maintenance of social movements is the role played by the mass media. The one major exception to this is Gitlin (1980), who makes media coverage the centerpiece of his analysis, and the one minor exception is McAdam (1982), who incorporates some elements of media coverage into his analysis, although it is not central theoretically. See Klandermans 1997, p. 44. See, for example, Skogan 1995; Gordon, Michels, and Nelson 1996; and Rahn and Transue 1996.

0521821711c02

26

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

particular, is a fecund area for symbolic coverage: over time, its use may have shifted – in particular, who is portrayed as the object of discrimination. Later, I will begin to undertake many of these analyses, particularly those focusing on the sheer quantity of news, on the coverage of core American values, and on stereotypes like poverty. Others – especially those involving particular leaders, political issues, social problems, movement dynamics, and symbolic words – must wait for another day and other scholars with different but complementary interests. I have emphasized the disparity between what we assume we know versus what we actually know, based on evidence, about the dynamics of media coverage. And I believe it is critically important that we test these assumptions. The results of such analyses will reveal much about American politics in the descriptive realm that is inherently interesting. When Americans consume news from the mass media, what are they learning about blacks – what they are like, what kind of people they are? From this perspective, one justification for this research is that many Americans, whites especially, learn a great deal about what they think blacks are like indirectly rather than directly, through the mass media rather than through (or at the very least in addition to) personal experience. What, exactly, are they learning? It is equally important to underscore that these assumptions have never been examined. And with good reason. To subject them to empirical scrutiny would place extreme demands on the researcher. For example, it would require large amounts of textual data, which, because of the nature of the subject matter, must span several decades. As interesting as it might be to look into the dynamics of media coverage of blacks during, say, the 1980s and 1990s (when textual data became more readily available), there would be an obvious desire to see how such results would compare to coverage in the 1950s and 1960s. Moreover, the source of such material must be consistent over time. Finally, even if all of the data were available, determining a method for analyzing the text in a meaningful way presents a daunting task. In the next section I present a solution to these problems. My solution does not match my expressed ideal, where any and every research whim could be satisfied. At the very least, it represents a good first step in that direction.

0521821711c02

0 521 821711

June 2, 2003

8:6

Eras of Media Coverage of Race

27

the newsweek data base and new york times supplement No analysis that purports to speak of “the media” does justice to the breadth of this broad and constantly evolving institution. The lone exceptions to this generalization are those (relatively rare) works that deliberately set out to compare content across media. But these analyses, because of the sheer scope of the task, always compare media at a single point in time. For the longitudinal analyst, whose task is to collect data across large spans of time, this is even more problematic. The recent availability of data bases like Nexis, which contains a wide variety of print-news sources, and the Vanderbilt archives, which contains data from television news, is a positive development. However, for a project like this, which focuses on the evolution of racial politics over the last half-century, problems of data availability are obvious hurdles from the start, for no existing source stretches far enough back in time to permit the most informative types of analysis. Therefore, the analyst interested in studying race in the media is left in something of a predicament. A wide variety of sources is available in electronic form from around 1980 onward. This permits more general conclusions about what “the media” are saying, but only for a severely restricted time period – a time period that does not cover the most interesting time in racial politics in recent history. Alternatively, the analyst can pick a single media source and attempt to supplement recent electronic data back in time. This strategy has the intended benefit of yielding data over a longer time span, but at the cost of examining a smaller slice of the possible media sources. In this book, I have adopted a modified version of the latter strategy, focusing on a single news source for the last half-century but making comparisons to another source of data for the more recent time period in an effort to enhance measurement validity. The magazine Newsweek is the primary media source that will be used in these analyses.21 For pragmatic reasons as well as issues of comparability, it was the best available source. For several decades, Newsweek has consistently had 21

For those who are interested in analyzing these data further, all of the Newsweek stories are freely available in full-text format from my website, http://www-polisci. tamu.edu/kellstedt/.

0521821711c02

28

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

a broad national circulation of over two million. Its audience is less highbrow than that of elite newspapers like the New York Times or the Washington Post. So an analysis of Newsweek is less susceptible to the criticism that only an elite few read it. And it is probably safe to say that Newsweek’s influence (like its circulation) has been relatively consistent over the years of interest. Contrast this with television (and hence television news), which was not widely available until the mid1960s. Finally, the layout of Newsweek has not changed significantly over the period of analysis. Not surprisingly, the more recent versions of the magazine are considerably more slick and colorful in their graphical presentation. But the key sections of the magazine in the 1990s – the Periscope, National Affairs, Business and Finance, Perspective, and nearly all others – are identical to their predecessors from the 1950s. The articles tend to be of similar lengths, and the magazine has roughly the same number of articles per issue today as it had in the past. In short, on matters of format, I have been able to find no significant differences between current issues of Newsweek and issues from the 1950s. For more recent years, media data are available from the Nexis data base. A total of 3,243 Newsweek articles from 1975 onward were obtained from Nexis.22 This figure comes from a Boolean search in Nexis for at least two mentions of the words “Negro,” “black,” or “African-American” from 1975 through 1994.23 This search criterion surely resulted in the retrieval of many irrelevant stories; any story that contains two or more mentions of, for example, black miniskirts would fit the search criterion. To avoid any possible contamination, these were eliminated by manually examining each of the 3,243 stories to see whether its content was manifestly about race in 22

23

No other source was available in electronic form from Nexis as far back as 1975, with most sources coming on-line some time in the 1980s. For practical reasons, this makes Newsweek an attractive choice for a media source, because it minimized the amount of work needed to supplement the material that was readily available in electronic form. This search criterion was compared with several alternatives, including a Nexis subject search (subject = blacks), and was found to be superior. In particular, Newsweek defined which stories were about blacks very narrowly – too narrowly for my purposes. For example, stories about poverty that mentioned higher black rates of poverty were not classified by Newsweek as stories about blacks. Such stories obviously belong in my analysis, and therefore creating my own search criterion became necessary.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

29

America or about something else. A total of 2,199 stories remained for the analysis for the period after 1975; all stories were used in their entirety. Obviously, any time-series analysis of race that strives to be complete must stretch back in time before 1975. Indeed, the most dramatic events of the century that are relevant for race all took place before data become readily available in electronic form in 1975. Therefore, optical scanning of stories from library archives became necessary. The decision about which stories to scan and which to ignore from each issue of the magazine was made primarily by reference to the index of each volume of Newsweek, and these decisions were corroborated by checking the table of contents of each issue. For the period 1950 to 1974, there were 1,953 stories about race in America in the magazine. Certainly, it would be desirable to have data from a wide variety of media sources for this project. It is possible that newspapers, magazines, television, and radio news have covered race differently, and that these differences might not even be constant with respect to time. Transcripts of television and radio broadcasts, alas, are impossible to obtain for such a broad time span, eliminating them from consideration. And, for the period of time before newspaper sources came on-line in the Nexis data base, optically scanning newspaper articles for analysis is impractical – and this is especially true for the large number of articles that would be required for this analysis in the case of race. Add to this the poor quality of archived newspapers, which were not designed to last intact for decades, and the similarly poor quality of most microfiche text, and this option becomes very quickly unappealing. Scanning magazine text, though occasionally difficult, proves to be far easier by comparison. In particular, even the older issues of Newsweek have retained the integrity of the paper with minimal smudging of the text, making the optical scanning far easier. This situation is far from ideal, of course, and creates a somewhat risky situation for inference. Namely, by relying so heavily on the Newsweek data base, there is a risk that any results of these analyses might not generalize to other media sources, which would restrict the authoritativeness of these results. In fact, that is a possibility. But it is one that can be minimized, if not entirely eliminated. In an effort to alleviate these legitimate concerns, I have retrieved 2,500 stories on race from the New York Times from the beginning

0521821711c02

30

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

of January 1981 through the end of December 1994 for purposes of comparison to the Newsweek series. The 2,500 New York Times stories represent a random sample of around 10,000 stories on race that appeared in the paper over that time period.24 The New York Times data base is a useful supplement to the Newsweek data. Although the time period covered is too short to permit its use in any statistical analyses, it will be instructive to compare the time trends of a variety of types of coverage between these two sources over the more recent periods. The goal of this exercise is to increase confidence in the generality of the Newsweek data base. In those cases where the year-to-year fluctuations in the Times coverage closely mirror the Newsweek coverage for the later period, it is sensible to conclude that the Newsweek articles are representative of broader trends in the media. In those cases where there are substantial differences in the trends of coverage, then the longer Newsweek series should not be considered to be representative. This, of course, is an empirical issue, and can only be resolved by case-by-case comparisons.

blacks in the news media: a half-century of coverage When Americans consume information about race from the mainstream press, what are they ingesting? More important, how does their diet of news about race vary from year to year, from decade to decade? Does the quantity of information vary systematically over time, and does the content vary accordingly? Most of us could generate somewhat raw speculation about these questions. Most would hazard a guess that the quantity of coverage, for example, ballooned during the peak years of the civil rights era – roughly from 1963 to 1965 – and then dropped off the mainstream media’s radar screen. And many would add that those same years witnessed the pinnacle of coverage that was cloaked in the language of American egalitarianism, which has subsequently been replaced by coverage cloaked in its countervailing value, individualism. How do these intuitions square with the facts? 24

The search criterion employed was a subject search (subject = blacks in U.S.A.), which was a much more inclusive subject than the subjects of Newsweek articles.

0521821711c02

0 521 821711

June 2, 2003

8:6

Eras of Media Coverage of Race

31

180 160 140

Story Count

120 100 80 60 40 20

94

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

52

19

19

19

50

0

Year

figure 2.1. Number of Stories on Race in Newsweek per Year, 1950–1994

Volume of Coverage Our first step in examining media coverage of race is, quite simply, to assess the volume of coverage over time. Combining our two sources of Newsweek stories, we have 4,152 stories at our disposal from January 1950 through December 1994. How prominently did race figure on the national media’s agenda over those years? Figure 2.1 shows the number of separate stories in Newsweek for each year during the period under study.25 There was only a modest amount of coverage of race in the early 1950s. Interestingly, even in 1954, the year of the Supreme Court’s landmark decision in Brown v. Board of Education, the magazine published only twenty-two stories on race. The first volley in the battle of desegregation, at least by this measure, was hardly a shot heard round the world. By the time desegregation was beginning to be implemented in 1957, in places like Arkansas, coverage increased more than threefold, to sixty-eight stories – still a modest amount. Perhaps in retrospect this disparity should not be surprising, as the nature of conflict in the hallowed halls of the United States Supreme Court is 25

Story counts are only one way to measure the quantity of coverage, but they are surely the simplest. Other strategies, such as word counts or even the number of electronic bytes in the text files, are also possible. It turns out that all three measures produce essentially identical results. The year-to-year correlation coefficient for the number of stories and number of words is r = 0.88; the same figure for stories and file bytes is 0.84; for words and bytes it is 0.99. Therefore I opt for the more intuitive and simple story count, with assurances that alternate measures would yield the same findings.

0521821711c02

32

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

less eye-catching than conflict in the streets of Little Rock. But it is unsurprising only in hindsight, for surely a greater volume of coverage would have been expected for such major events as the Brown decision. The day-to-day struggles of the civil rights movement – in this case, desegregation – warranted more coverage than one of the epic legal battles in the Court’s history. The quantity of coverage of race hovered in this same vicinity – about one story per weekly issue – through the late 1950s and into the early 1960s, until 1963, when there was an explosive growth in the number of stories (152 in that year). The apex of the civil rights movement, consistent with our intuitions, did receive disproportionate shares of coverage. The peak in the number of stories lasted for several years through the colossal legislative battles of the next two years and through the riots and long, hot summers of the last years of the 1960s. Throughout the period, Newsweek published in the range of three stories per issue on race. No other subject received such sustained amounts of coverage over these years, though, of course, major events like the assassination of President Kennedy, the election of 1964, and the exploration of outer space loomed large as well. Even the escalating conflict in Vietnam, during the mid-to-late 1960s, was overshadowed by coverage of race. This began to change at the dawn of the 1970s, which witnessed a slow, steady decline in the amount of coverage of race in the magazine. Space on the public agenda once dedicated to race was increasingly taken up by the Vietnam War and the furor over American involvement (particularly after the 1968 presidential election). The Watergate scandal also began to consume the attentions of the mass media in the years after Richard Nixon’s reelection in 1972. But, interestingly, race did not disappear from the front pages for long. As American involvement in the Vietnam War wound down and Richard Nixon resigned from office, race again assumed a more prominent place in Newsweek. The number of stories rose again as controversies over school busing to achieve desegregation spread across the nation, and continued to proliferate into the late 1970s and early 1980s as affirmative action (and the backlash it spawned) captured headlines. Coverage of race in the late 1980s and early 1990s still revolved, to a large degree, around the candidacies (and potential candidacies) of Jesse Jackson for the Democratic presidential nomination,

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

33

and by affirmative action. In addition, it was buoyed by two great scandals involving prominent black men: the harassment scandal surrounding Clarence Thomas after his nomination to the Supreme Court, and O. J. Simpson and the mystery surrounding the murder of his ex-wife. The hard edge of hip-hop and rap music culture also emerged as a minor theme in these later years.26 Comparing article counts to cover stories and photographs also provides an illuminating view of Newsweek’s coverage of race. Its covers focused almost exclusively on desegregation in the 1950s, and on the continued struggle for civil rights and the southern resistance in the early 1960s. The magazine published many covers on Martin Luther King, Jr., and the marches in Selma and Washington. In the later years of the 1960s, the images shifted to the riots and other urban problems such as poverty in the nation’s cities. In the mid-1970s, race did disappear from the cover of Newsweek – there was a gap of over five years between 1973 and 1978 when there were no references to race or pictures on the cover. In the 1980s, electoral politics dominated the covers during election years, including several cover stories of black politicians like Jesse Jackson and Chicago Mayor Harold Washington. Four cover stories on the Clarence Thomas nomination appeared in the early 1990s, as well as five on the O. J. Simpson case, and three more on rap music and violence. In short, the covers, like the stories in the magazine, followed a similar pattern. Black Americans were virtually invisible in the earliest years under study, the 1950s.27 There was an increase, though not perhaps to the degree expected, toward the end of the decade, as coverage highlighted the struggles over desegregation. As the civil rights movement reached its pinnacle in the early and mid-1960s, coverage blossomed, and remained high as it struggled to deal with the urban riots in the latter part of that decade. After a lull in the early and mid1970s, a more diverse blend of stories began to appear, with surprising frequency, revolving around black political candidates, affirmative action, and national scandals tinged with sensationalism. 26

27

None of these over-time fluctuations, in my view, can be attributed to a fluctuation in the overall number of stories in Newsweek, which have remained roughly constant over the period of study. The term “black invisibility” in reference to race and the media was first used by Johnson, Sears, and McConahay (1971).

0521821711c02

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

34

This initial, broadbrush look at media coverage tells us about the volume of coverage, about how loudly the messages were being communicated. It does not, however, tell us much about the nuances of coverage, about the themes that Newsweek raised and echoed, and how those themes were likely to affect public opinion. That is our next task. Core Values: Egalitarianism Unlike the straightforward measuring of the raw quantity of information in Newsweek, by comparison the task of quantifying the magazine’s mood, its more qualitative connotations, presents more difficult challenges. The three concepts that we will track over time – egalitarianism, individualism, and poverty – are, to varying degrees, elusive. Taking the raw stories and creating a dictionary that will detect individualist or egalitarian emphasis or references to the impoverished status of blacks – content analysis – is time-consuming, for understandable reasons. These stories never overtly trumpet the fact that “liberal racial policies can violate America’s belief in individualism” or that “America’s egalitarian values demand a level playing field for all races.” (On the other hand, they do, with some regularity, make clear and unmistakable references to black poverty.) Allusions to values are subtle, and explicit references are the exception. That is, Newsweek does not often (if ever) use the word “individualism” when making a reference to individualism. Not surprisingly, then, the resulting scheme for measuring these mentions of core values is complex.28 But what do we mean by “egalitarian cues”? Simply put, Newsweek is framing an issue in egalitarian language whenever it reminds its readers about that strain of American beliefs which emphasizes the equal worth of all people. (In Chapter 3, I will examine how egalitarianism is particularly relevant for American public opinion on race.) When their coverage of day-to-day events stresses this in a decipherable way, the magazine is using an egalitarian context to “frame” 28

In the content analyses that follow, I use InfoTrend 1.0 (InfoTrend, Inc., 2115 Dudley Ave., St. Paul, Minn. 55108). This technology allows for the translation of words (and specified combinations of words) into user-defined ideas (such as individualism and egalitarianism). The analyst provides a dictionary of key words and how they combine to form ideas.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

35

its coverage. To be very specific, the detection of egalitarian cues revolves around the recognition of several types of ideas, each of which may have unique words or combinations of words associated with it: statements about blacks having (or deserving) equal rights to those of whites or about equal rights in general; phrases focusing on fairness or equality (or the lack thereof) in various parts of the political process (e.g., voting rights); sentences referring to “equal protection of the laws” or to its counterpart, Jim Crow laws; reports of blacks seeking equal access to facilities such as buses, schools, lunch counters, and the like; stories referring to access to open housing; sentences describing bigotry directed toward blacks or hate crimes committed against them; statements focusing on racial discrimination in employment and laws that attempt to prevent it; and finally, stories that describe segregation. This is an extensive list. What unites these subthemes is that they all deal with the presence or absence of equality of treatment in some public context. When making references such as these, the national media were framing their coverage of race in egalitarian language, portraying American society as one in which blacks are (or have been) treated as less than full citizens, as subequals. Such cues have the effect of making a connection in the minds of the American public between a highly abstract value that most citizens believe in – equality – and a set of highly concrete circumstances in American politics – the political and social position of black Americans. It is as if the magazine were saying, subtly, “Apply the equality principle to the case of blacks.” To return to the theme introduced at the beginning of this chapter, it is providing an interpretive context for consumers of the news. The sum total of these subthemes will yield an over-time measure of media framing that highlights egalitarianism. Before presenting those results, though, it will be helpful to describe one set of the findings that helps to compose the overall index. Stories about segregation, not surprisingly, were a more prominent part of the egalitarianism landscape in the 1950s and 1960s; and, as the issue faded from the forefront of the civil rights movement, so did its presence in Newsweek. Figure 2.2 tracks, on an annual basis, the number of references to racial segregation in Newsweek between 1950 and 1994. By 1954 – the year the Supreme Court issued its decision in the landmark case Brown v. Board

0521821711c02

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

36 250

Number of Mentions

200

150

100

50

94

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

52

19

19

19

50

0

Year

figure 2.2. Mentions of Segregation in Newsweek, 1950–1994

of Education, which had segregation and the “separate but equal” doctrine at its very core – the magazine referred to segregation almost one hundred times. But, as was the case with the quantity of coverage, the events of the subsequent few years – where the continuing day-to-day politics of desegregation, not isolated Supreme Court decisions, took center stage – garnered more coverage in the magazine. There were twice as many references to segregation in both 1957 and 1958 – almost three mentions per weekly issue – as there had been in 1954. After a lull in references to segregation in 1960 and 1961, the largest spike in the series occurs in 1963, when the magazine referred to segregation 221 times. But after two more years of consistent and frequent coverage of segregation, it slowly but steadily began to vanish from the pages of Newsweek. (Remembering the series of story counts in Figure 2.1, this is obviously not because overall coverage declined.) Interestingly, even during the open-housing debate surrounding the Civil Rights Act of 1968, little sustained coverage of segregation as an issue appeared in Newsweek, which mentioned segregation 54 times that year. And it has scarcely remained on the proverbial radar screen since then, never quite hitting 50 mentions per year since 1970, and being mentioned only 13 times in 1989. The press’s use of egalitarian language in describing segregation is easy to see, even when the word “segregation” is not overtly used in

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

37

the reporting. Consider the following portion of a March 18, 1957, story involving segregation in the Florida university system: Squarely in the face of a U.S. Supreme Court ruling, the Florida Supreme Court denied immediate entrance of a Negro educator to the all-white University of Florida law school. Just a year ago, the highest tribunal decreed that Virgil D. Hawkins, a 49-year-old instructor at Bethune-Cookman College in Daytona Beach, should be promptly admitted to the law school. But the Florida court, in a 5–2 majority opinion, held that it was the “compelling duty” of the state to prevent the violence which it foresaw if Hawkins should enter the school at this time.

The reference to an “all-white” institution makes the unequal treatment of blacks perfectly clear, and hence draws the reader’s attention to their egalitarian beliefs. Political leaders frequently made references to segregation, as this example from President Kennedy shows. It is from the July 1, 1963, issue: Exclusion of Negroes from public places “is a daily insult which has no place in a country proud of its heritage – the heritage of the melting pot, of equal rights, of one nation and one people,” Mr. Kennedy said. As for protests that such legislation violates private property rights: “ . . . there is an age-old saying that ‘property has its duties as well as its rights.’ . . . ”

Again, the word “segregation” was not even used here, but the phrase opening the paragraph, referring to the “exclusion” of blacks “from public places” makes an unmistakable reference to unequal treatment based on race. The more comprehensive measure of egalitarian cues – including all of the subthemes like segregation and racial discrimination and the like – is presented in Figure 2.3, again from 1950 to 1994. The battles over desegregation in the South led to an abundance of egalitarian cues in the mid-to-late 1950s (including 219 in 1957 alone). Many of these articles drew attention to the denials of blacks’ political rights, but a larger share revolved around segregation and the political forces resisting integration. The heyday of egalitarian coverage, not surprisingly, was the 1960s. In 1963, an astonishing 482 egalitarian cues were found in the magazine – an average of roughly nine per week. Most of these revolved around the civil rights movement and its persistence in placing civil rights at the center of the nation’s political agenda. The

0521821711c02

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

38 600

Egalitarian Cues (Number)

500

400

300

200

100

94

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

52

19

19

19

50

0

Year

figure 2.3. Egalitarian Cues in Newsweek, 1950–1994

Kennedy administration, though, also accounted for many of these references, as the quotation about segregation highlights; the administration was willing to engage in a very public discussion with Martin Luther King, Jr., over civil rights. Remarks from both sides had the effect of reinforcing the same principle – egalitarianism. The 1970s witnessed a decline in egalitarian cues. By 1975, the magazine made only 105 references to equality in their stories about black Americans, the fewest it had made in twenty years. As was the case with overall coverage, surely a part of this decline was a function of the prominence of two other earth-shaking political issues – the Vietnam War and Watergate – occurring at the time. Part of it, too, can be attributed to having leaders in the administration less sympathetic to the plight of blacks. It is perhaps surprising to discover that egalitarian cues made somewhat of a comeback during the 1980s. When examining the articles, it is easy to see how the conflictual nature of politics in a media age produced such a result. The conservatism of the Reagan administration inspired many liberal activists to decry the administration’s conservatism on race. Though never returning to their 1960s-level prominence, portrayals of the inequities that face black Americans reappeared during the 1980s and early 1990s. The majority of references to the broader theme of egalitarianism are quite straightforward, easily digested by most consumers of

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

39

mainstream news. Consider the following paragraph from the March 1, 1965, issue of Newsweek: The Selma demonstrations, for all their burlesque aspects, have served the purpose of dramatizing a conspicuous failure in the whole civil rights effort. Nothing so far done has assured qualified Negroes that they may vote now, or that they have any very good prospect of voting in the near future, in many Southern communities. On the fringes of the Old South and in some Southern cities they have made gains. But there are still counties in Alabama and Mississippi where not a single Negro is registered. Literacy tests applied to Negroes are often so difficult that college professors can’t pass them. The tests applied to whites are so simple that any fool can qualify.

Here, the combination of a negative phrase (“not a single”) with reference to blacks and the political process (“registered”) indicates how blacks are excluded from participating in the electoral process, and hence in egalitarian values. In a similar vein, consider the following paragraph from the January 24, 1966, issue that focused on the relationship between President Johnson and the Congress: The president urged a law banning racial discrimination in the sale or rental of any private housing. The proposal was at once a pleasant surprise and an annoyance to civil rights leaders, who prefer a presidential decree expanding JFK’s partway 1962 order against housing discrimination. LBJ’s way – bucking the issue to Congress – would be touchy any time, doubly so in an election year.

The two uses of “discrimination” in close proximity to “racial” and then “housing” produce two egalitarian cues in the paragraph and clearly portray the disadvantages that blacks face. The references to egalitarianism in the 1980s and 1990s often show the other faces of discrimination, or lingering discrimination. For example, the well-documented practice of “red-lining” in the lending industry is documented in the following story from the May 16, 1988 issue: The Sunday, May 1, edition of the Atlanta Journal and Constitution led off with a shocker. It was the beginning of a series of articles accusing Atlanta’s banks of discriminating against blacks. According to the painstaking analysis of computer data provided by the banks themselves, lenders were five times more likely to give mortgage and home-improvement loans to whites than

0521821711c02

40

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

to blacks. The paper quoted an Atlanta mayoral candidate, Fulton County Commission Chairman Michael Lomax, who had to approach three banks before he was able to get a $115,000 home-improvement loan. Lomax asked, “If I, a powerful black elected official, can’t get a loan, what black person can?”

Again, the clear phrase “discriminating against blacks” leaves an unmistakable impression on the reader, showing the disparate treatment received by blacks and whites. How much faith can we put in the particular measure of egalitarian cues that arises from the Newsweek stories? It is, after all, a measure of coverage from a single news outlet over time. One can make a plausible argument, noting especially the prevalence of pack journalism, that Newsweek ought to be a good indicator of what the other mainstream press outlets were covering at the time. Perhaps other sources would produce largely the same results. But perhaps not. And that would damage the generalizability of our findings, for it would indicate that it would be close to impossible to speak of “the media” or “the mainstream press” in a general way. We would have to restrict our conclusions to the less inspiring domain of whether Newsweek influenced public opinion. The broader goal of this research, though, is to use Newsweek as a representative indicator of all of the mainstream press. Because we also have at our disposal several thousand New York Times stories, they can serve as a check on our Newsweek series. Applying the identical dictionary that detects egalitarian cues in Newsweek to the sample of New York Times stories allows for a validation of the Newsweek data. If the same dictionary, applied to two different sources of mainstream news, were to produce wildly divergent results – with the rising and falling of one series over time uncorrelated with the rising and falling of the other – then it would signal that at least one of the media sources is highly idiosyncratic rather than representative. It would say, in effect, that they are more different than they are alike. But if, on the other hand, their trajectories over time have much in common, that would increase our confidence in the Newsweek measure. The stories from the Times, you will recall, were available only from 1981 through 1994. For those years, the stories in Newsweek and the Times correlate at an impressive 0.70. In other words, in years when Newsweek made more references to egalitarianism, so did the Times.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

41

And in years when Newsweek made fewer such references, again, so did the Times. It is not as if these two media outlets presented coverages of race that were completely independent of one another; their themes, over time, echoed one another. This finding is important, particularly because it allows us to proceed with confidence in the longer, more comprehensive Newsweek measure, knowing that the magazine is not completely idiosyncratic. It allows us, in other words, to speak more generally about the mainstream press. Core Values: Individualism Another core value that is relevant for the debate on race is individualism, the sparring partner of egalitarianism. Every time the egalitarian strain of American life jabs with the notion of fair play and governmentenforced equal access, individualism counterpunches with the notion that every person needs to succeed through his or her own efforts. As we shall see in Chapter 3, it is the other key component of the American ethos that influences popular opinion on race. One suspects, as a consequence, that it will appear in the mainstream press in some form. Does it? Individualism is relevant to the politics of race in several ways, which necessitates a strategy for measuring individualistic coverage that is comprehensive in the same way as the egalitarianism measure was. Detection of individualistic value cues in Newsweek focuses on finding the following types of ideas: coverage describing any type of government policy that discriminates against whites in its effort to help blacks; sentences that depict “reverse” discrimination; stories that portray blacks as lazy and undeserving of assistance or equality; stories that mention how well qualified people are when they are in competition (for jobs, college admissions, or the like); and phrases that describe individuals (black or white) as “earning” or “deserving” the benefits or goods that they receive. These stories capture coverage about race that focuses on how individuals are – or how they should be but are not – getting ahead by their own efforts, and about whether people are deserving of assistance. They key issue here is merit: are people, both blacks and whites, getting what they deserve? Deservingness – of economic success, of assistance from the government – is determined by effort. Those who live by the Protestant work ethic are exalted as virtuous, whereas those

0521821711c02

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

42 250

Individualism Cues (Number)

200

150

100

50

94

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

52

19

19

19

50

0

Year

figure 2.4. Individualism Cues in Newsweek, 1950–1994

who are dependent (especially on the government) are seen as lacking in character. Applying a dictionary that captures these principles produces an annual time series from 1950 to 1994 of individualistic references. Figure 2.4 displays the resulting time series of individualistic value cues. References to individualism were exceedingly uncommon before the mid-1960s, although there were occasional stories or portions of stories about black business people that yielded value cues. Between the mid-1960s and early 1970s, the mentions of individualism became somewhat more common. Often such references were made about the Johnson administration’s rather amorphous introduction of the term “affirmative action,” or in questioning the work ethic of blacks in the inner cities during the riots of the period, though it is worth emphasizing that the vast majority of references to black poverty (which I will examine later) did not draw attention to individualistic values in this manner. The spate of federal and state court cases that addressed affirmativeaction programs – including the landmark Bakke v. Board of Regents case of 1978 – produced another burst of growth in individualistic cues that has lasted through the end of the time period. The largest spike in the series, though, came in 1991, which not only included several affirmative-action controversies but also the nomination of Clarence Thomas to the U.S. Supreme Court. The Thomas nomination included

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

43

a good number of references to affirmative action – from which Thomas benefited at various points in his life – and also about his qualifications to be on the high court. After his nomination, coverage of individualism returned roughly to prenomination levels. Were it not for the Thomas nomination, though, it is possible to imagine something of a steady-state equilibrium for the amount of coverage revolving around individualism – a steady state that exists in the post-Bakke era. As with egalitarian cues, some sample paragraphs should help to elucidate the process that produced the individualism time series. For example, the following paragraph appeared in Newsweek’s March 11, 1974, issue, in a story about a Florida Supreme Court case: For almost a decade – a decade of unprecedented racial and social tumult – the issue lay untouched, rather like a bomb that could be removed only if someone set the fuse. Last week, the fuse was finally set. The U.S. Supreme Court heard arguments on one of the most critical civil-rights issues of the 1970s: does the Fourteenth Amendment (which was framed to help relieve discrimination against blacks) prohibit special preferences that discriminate against white people on account of their race? In short, is “reverse discrimination” constitutional?

The paragraph has mixed references. The phrases “discriminate against whites” is an individualistic reference, as is “reverse discrimination,” because both refer to blacks getting ahead because of their skin color rather than through personal effort. However, the parenthetical remark about “discrimination against blacks” counts as an egalitarian reference, for reasons already mentioned. The fact that this was not necessarily intended to convey an egalitarian message is not problematic, because (regardless of the writer’s intent) readers are reminded of discrimination against blacks, and thus the inherent egalitarian message. Articles about so-called reverse discrimination in higher education – in admissions and hiring – were particularly common. The following article, which describes the effects of affirmative-action programs to help both women and blacks, appeared in the December 4, 1972, issue of the magazine: Many professors contend that the inevitable result will be discrimination against male scholars. “If you hire unqualified women,” says Frankel, “qualified white males don’t get jobs.” Adds John Bunzel, president of California

0521821711c02

44

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

State College at San Jose: “I’ve heard cases of people hiring someone just to avoid a hassle with the federal government. I know of people who have gotten letters saying, ‘Your qualifications are excellent, but we are looking for a black or a woman this year.’”

The story included several references to individualism. Note the number of mentions of the word “qualified” (and its variants), and in particular the phrase “qualified white males” in the first quotation. Whatever the accuracy of the claims, they have the effect of bringing the discussion squarely to the issue of qualifications, deservingness, and earning what one gets. The cry of American individualism is heard loud and clear. Finally, a story from November 14, 1994, about Clarence Thomas, three years after his confirmation battle to the Supreme Court, shows how he continues to be a polarizing force in American racial politics: “He [Thomas] was eager to prove his own merit, but doubted himself when he did. At each stage of his education from high school through Yale Law he was admitted through affirmative-action programs he is now famous for disdaining.” Again, the issue of merit is raised directly, and again when it mentions Thomas’s benefiting from affirmative-action policies. Another feature of the coverage of individualism is worth mentioning. In addition to the considerable attention devoted to affirmative action, many articles about education also refer to standardized test scores. These often implicitly – and sometimes more explicitly – raise the issue of individualism, as they compare the achievements and deservingness of black and white students. The implication, of course, is that test scores are valuable indicators of achievement and ought to determine a student’s admission to college. Standardized test scores become the measuring stick of individual achievement. As we did with the egalitarian series, I have compared the Newsweek individualism series to the one obtained from the New York Times. Again, the identical dictionary was used on both sets of text. The similarities turn out to be substantial (the correlation coefficient is r = 0.58), again indicating that we are not measuring something purely idiosyncratic with the Newsweek indicator of individualistic cues. Instead, there seems to be a more general phenomenon at work here, and the Newsweek indicator does a very good job of capturing it.

0521821711c02

0 521 821711

June 2, 2003

8:6

Eras of Media Coverage of Race

45

Black Poverty By comparison with poverty, individualism and egalitarianism are piein-the-sky abstractions. Readily understood by anyone regardless of educational attainment, portrayals of poverty in the mainstream press are politically powerful. This was particularly true in the post-Depression era, when fighting poverty became one of the central jobs of government. Coverage of poverty has the effect of telling the country precisely who these poor people are whom government is supposed to be helping. The connection between poverty and race today is so common as to be almost expected. This, as we shall see, has not always been the case. Describing this evolution is the task at hand. In the mainstream press, the idea of poverty is conveyed in a variety of ways, some of which are straightforward, others less so. Take, for example, the use of the word “ghetto” in the context of a story about black Americans. Clearly the word does not conjure up images of affluence and splendor; but it has not always been a part of the American political lexicon. It is a relatively simple matter, using the Newsweek data base, simply to count the number of mentions of the word. I have done precisely this, and the results appear in Figure 2.5, with annual counts from 1950 through 1994. The figure is striking in several respects. First, notice that, into the early 1960s, Newsweek almost never used the word “ghetto” at all – a grand total of twice through the end of 1960. As late as 1962, there were a paltry six uses of the word in the entire year. In terms of content, the bulk of these early mentions refer 350

Number of Mentions

300

250

200

150

100

50

82 19 84 19 86 19 88 19 90 19 92 19 94

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

52

19

19

19

50

0

Year

figure 2.5. Mentions of the Word “Ghetto” in Newsweek, 1950–1994

0521821711c02

46

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

to cities in the North. As a result, they take on an eerily – and, it turns out, accurately – prophetic quality. Consider the following vignettes – the first from the December 23, 1957, issue: With migration rapidly swelling ranks of the Negro voters in the central cities . . . changes are due. “It’s the white man’s own doing,” says a Negro GOP strategist in Washington. “He’s got them so ghettoed in that he’s got to give them representation.” Los Angeles Negro clergymen have organized to elect a Negro congressman in 1958. “Every churchman has been budgeted with a responsibility for votes, and he will deliver,” says an insider.

– and the second from the November 13, 1961, issue focusing on cities: And what of Harlem itself? The searing fact remains: much of Harlem is a Negro ghetto, there for all the peoples of the world to see. The great bulk of its 400,000 dwellers live in miserable dirty buildings – again at exorbitant rents – on squalid, crime-ridden streets. And increasingly large numbers are coming from the South, often semi-literate and near-primitive in their ways, bewildered and disillusioned by the promised land of the North.

Finally, the following paragraph from the May 27, 1963, issue foreshadows the controversies to come: Everywhere plain signs arose that Birmingham had thrust the whole U.S. racial problem into a new phase – more intense, more flammable, more urgent. New demonstrations marked by occasional violence sprang up in moderate Southern cities that thought they had long since balmed racial tensions with token concessions to the Negro. And in the black ghettos of the North, shock waves from Birmingham stirred up fresh demands for an end to Yankee-style de facto segregation. Among whites, North and South, the crisis not only spurred conscience but touched a nerve of fear.

These references to black poverty come across crystal clear; but they don’t come across loudly, for the key point to remember is that, despite the unmistakable clarity of these paragraphs, as late as 1963 the magazine only rarely helped its readers make the connection between race and poverty. That quickly changed. The word “ghetto” became an indispensable part of the media’s dictionary soon thereafter. Newsweek included 32 references to ghettoes in 1964 – the year of the riots in the New York neighborhoods of Harlem and Bedford-Stuyvesant – and 74 references

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

47

in 1965 – the year of the Watts riots in Los Angeles.29 In the long, hot summers that followed, with civil disturbances spreading throughout the major cities of the industrial North, references to ghettoes proliferated, peaking in 1967, with 310 mentions of the word for an amazing average of almost 6 mentions per issue of Newsweek for that year. Whereas a scant five years earlier, consumers of the mainstream press had heard very little about this facet of black poverty, by 1967 the message was blaring in their ears and could hardly be missed.30 The following sample paragraphs from the riots of 1964 and 1965 highlight the appalling conditions in which many black Americans in the North lived. The very first story on the riots in New York City, from the August 3, 1964, issue, makes the point clearly: For the first time in more than twenty years, the nation’s greatest city flamed with racial strife and bloodshed. Heavily armed helmeted police patrolled the Negro ghettos of Harlem and Brooklyn. Further down in Manhattan, other battalions of police stood ready should the riots threaten to sweep over more of the city. Mayor Robert Wagner broke off a European trip and flew home to take charge. For a few days, large parts of New York City were reminiscent of Paris during the flood tide of Algerian nationalism.

And in August 1965, when riots broke out in the Watts section of Los Angeles, Newsweek (August 30) described the scene in the following way: ‘Black and white’: And therein lay the hard lesson of Los Angeles. In the past decade, mainstream Americans had been painfully educated to the plight of the Negro and had been moved to remedial action – most recently in the momentous Voting Rights Act of 1965. But the landmark battles of the decade – the Birminghams and Little Rocks and Selmas – had been symbolic encounters of black nonviolence with white repression. By contrast, the ghetto riots of the past two summers, from Harlem to Watts, have cast the Negro as the 29

30

This spike is not a function of one or two extensive stories about ghettoes with many mentions. The seventy-four references to the word ghettoes in 1965 come from twentytwo distinct articles. These effects are not a function of a special feature story that might quickly be forgotten; they are the result of language introduced and then used repeatedly, keeping it fresh in the minds of readers. With this and all other computer-generated content analyses, a human coder checked for possible problems with the computer-generated results and found over 95 percent agreement. More important, there were no systematic (that is, repeated over and over) errors in coding.

0521821711c02

48

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

lawbreaker – and compelled whites to respond with massive and sometimes indiscriminate force. There is no other answer to a riot, as the nation’s top Negro leaders were quick to agree. Yet force was not the final answer to the conditions that make riots.

The magazine’s portrayal included dramatic and shocking displays of how quickly the civil rights leadership had lost control of the agenda. Again, from the August 30, 1965, issue: In all history of the Negro revolt, no single leader has moved more men to disciplined, nonviolent action in the name of God and the cause of equality than Martin Luther King. Yet he had barely set foot into riot-shattered Watts last week when he learned, with face-slapping force, the depth of the chasm that divides him no less than white Americans from the angry black ghetto. He came to offer hope to the ghetto, but his welcome was a mocking jeer from an onlooker: “‘I had a dream, I had a dream.’ Hell, we don’t need no damn dreams. We want jobs.”

In all of the subsequent coverage of riots, the media-consuming public was repeatedly reminded of the impoverished conditions in which millions of black Americans lived. These snippets have a similar ring to those from the 1950s and early 1960s. Where they differed was not in their tenor, but in their volume. The American public digested a steady diet of references to ghettoes throughout the rest of the 1960s, although they did not approach the 1967 intensity. Still, by 1971, references to ghettoes were becoming less frequent, reaching a steady state for most of the remainder of the time period, averaging perhaps one use of the word in every two or three issues. Note, however, that the term did not entirely disappear from Newsweek coverage although its use declined still further in the 1990s, and that it still remains a politically meaningful word. References to “ghettoes” since the riots of the 1960s have appeared in several other forms, most notably in coverage of school busing. The reporting of the busing crisis in Boston in the early 1970s captures this well (from the September 23, 1974, issue): After a summer of gathering apprehension, the city of Boston dispatched a fleet of yellow school buses last week to end de facto segregation in the country’s oldest public-school system. On a muggy opening day, just after 8 o’clock, the first bus rolled up to the front gate of South Boston High, filled with twenty black teen-agers from the nearby Roxbury ghetto – and it touched off an

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

49

eruption of ugly passion that made South Boston look like Little Rock in the bad old days.

The reminders of black poverty come in more subtle forms as well. Consider even the following paragraph from an October 16, 1989, article on clothing fashions: Urban blacks have seen their fashion fads move downtown before. Inner-city ghettos have long been a creative crucible for styles that later sweep teenage America. The craze for untied Adidas sneakers and high tops, oversize t-shirts, huge gold earrings and gold chains all began in black neighborhoods. White America soon copied not only the fashions but even the hairdos. The fade, a hairstyle worn by some black men for the last few years, can now be seen on white teenagers. It is trimmed short at the sides and left long on top.

In an article that has no overt connection to politics whatsoever, readers of Newsweek are nonetheless reminded that black Americans form a disproportionately impoverished group of citizens. These reminders, though almost surely without political intentions, can just as surely have political consequences. When we hear the word “ghetto” in the mainstream press today, of course, we think of black poverty. The path of the time series in Figure 2.5 resembles one that is at equilibrium, has a sustained upward shock, then drifts back down again, though not to previous levels. A new steady state has been reached, it seems. The use of “ghetto” in the context of a story about black Americans makes a rather unmistakable reference to the idea that blacks are disproportionately poor. That said, references to black poverty in the mainstream press comprise more than the use of that one word. My more general coding scheme emphasized several combinations of ideas to produce the concept of “blacks as impoverished.” It included references – again, in the context of stories about black Americans – to housing projects, the “inner city,” and the decay of central cities. In addition, references to welfare or welfare programs in the context of black people were included. Finally, references to the inferior economic conditions of blacks (including inferior housing conditions) were counted as references to black poverty. Applying this to the Newsweek stories produces an annual count of the number of references to black poverty. That more complete count is displayed in Figure 2.6, and the sight is a familiar one: the figure

0521821711c02

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

50 800 700

Number of Mentions

600 500 400 300 200 100

92

90

88

86

84

82

80

78

76

74

94 19

19

19

19

19

19

19

19

19

19

70

68

66

64

62

60

58

56

54

52

72

19

19

19

19

19

19

19

19

19

19

19

19

19

50

0

Year

figure 2.6. Mentions of Poverty in Newsweek, 1950–1994

resembles, almost precisely, the graph for mentions of the word “ghetto.”31 The 1950s saw very few references to blacks as impoverished. In 1952 and 1954, in fact, there was not a single reference to black poverty. In the early 1960s, however, the frequency of mentions of black poverty increased to the point where the magazine was averaging close to one such reference per weekly issue. The mid-1960s, however, witnessed a skyrocketing of references to black poverty, with 1967 the high point (680 mentions). In the early 1970s, mentions of black poverty continued at a relatively high rate, but then reached something approximating a steady state from the mid-1970s through the end of the time period under consideration. The variation in this series is striking and confirms the suspicion that the media’s promulgation of the poverty stereotype has not been constant. Rather, there have been distinct eras of coverage emphasizing this theme. The first, covering roughly the years before 1965, exhibits almost no attention paid to black poverty; in effect, in this period, the media did not promote the stereotype of black poverty at all. The second era, covering the mid-1960s through the early 1970s, represents a time when the media suddenly discovered the issue of black poverty as a story frame. Coverage grew extremely rapidly, from only a trivial number of mentions in the early 1960s to an average of over three mentions per issue in 1967. The third era, covering the years since the 31

Of course, there are far more mentions of poverty, as “ghetto” mentions are a subset of the larger category of poverty mentions.

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

51

early 1970s, represents the era of reinforcement of the stereotype, with Newsweek making approximately one mention per issue over this time span. It is worthwhile to note a few other qualitative features of the results. First, upon inspecting the paragraphs where references to black poverty were found, one is struck by their overwhelming sameness. Many or most of these stories did not have as their sole (or even primary) purpose the conveying of the idea that blacks are impoverished. They are an odd mixture of indirectness and directness – indirect in the sense that most of them are transparently “about” something other than black poverty, and direct in the sense that it is impossible to mistake the words “black ghettoes,” for example, for anything but a reference to the fact that blacks are a relatively impoverished group in America. (To wit, a reference to a “white ghetto” or an “Asian ghetto” is almost unthinkable.) Second, the time period of the late 1960s as the peak of references to black poverty is intriguing. What was happening in America that led the media so substantially to increase their references to black poverty at that time? Although several events spring to mind, the most obvious possibility revolves around the urban riots that began in the summer of 1964. A second source was the Johnson administration’s War on Poverty. The administration’s sustained effort, begun in early 1964, also enunciated the connection between race and poverty. In the magazine’s very first words about the War on Poverty, in the February 17, 1964, issue, in a special section entitled “Poverty U.S.A.,” the connection could hardly be more clear: “Summoned now to Lyndon Johnson’s ‘Unconditional war on poverty,’ Americans can find the battlefield on all sides: In a squalid Chicago slum, a Negro mother rages: ‘Why we got to go hungry and naked?’” The proliferation of cues about blacks and poverty arose, it seems on first glance, from a combination of realworld events (such as the riots) and the straightforward reporting of the actions of political leaders. Equally important, though, is the question of what forces caused the decline in references to black poverty in the 1970s. These questions will be addressed more fully in Chapter 5. Interrelationships The issue of the interrelationships among the different types of media framing is an important one. Upon inspecting the figures of the various

0521821711c02

0 521 821711

June 2, 2003

52

8:6

Mass Media and American Racial Attitudes

series representing the different types of framing, it is natural to wonder if they “compete” with one another for magazine space. In other words, is it the case that an increase in egalitarian frames, by definition, comes at the expense of individualistic frames? And, conversely, when individualistic frames come to the fore do egalitarian frames begin to wane? Is it, in effect, one or the other? If indeed they compete for agenda space, we would find them to be negatively correlated: increases in the one would be associated with decreases in the other. The same questions arise concerning the relationship between poverty frames and those of the two values. Will increases in the number of references to poverty coincide with increases in one or the other value? Either scenario is plausible. The linkage between poverty and individualism frames would suggest that a large number of poverty references also contain allusions to effort or laziness in the context of poverty. In contrast, the linkage between poverty and egalitarian frames would suggest that poverty references are more likely to occur in the context of discussions of discrimination and its legacy. Table 2.1 displays simple correlations for the series. In the left-hand column of data, the correlations – which, again, can have absolute values between 0 and 1 – are for the full series from 1950 through 1994. Here, there is no support for the notion that individualism frames occur at the expense of egalitarian frames, for the coefficient of 0.26 is both positive and statistically significant, indicating that increases in one type of framing are associated with increases (not decreases) in the other. However, to some degree this may be a function of the relatively small amount of coverage that race received as a whole before the early 1960s. In the right-hand column of data, then, the same correlation is computed, except it is only for the years 1963 and beyond. The story there is different: The correlation of –0.02 (and no statistical table 2.1. Correlations between Types of Media Framing Pairing Individualism–Egalitarianism Individualism–Poverty Egalitarianism–Poverty a

p < 0.05.

Time Period 1950–1994 0.26a 0.18 0.52a

1963–1994 −0.02 −0.25 0.37a

0521821711c02

0 521 821711

June 2, 2003

Eras of Media Coverage of Race

8:6

53

significance) shows that, for the period when race received more attention, the two types of value frames are all but entirely uncorrelated. In other words, there is no back-and-forth tug in the number of individualism and egalitarianism frames in Newsweek. The correlations between the number of poverty cues and the two value cues in Table 2.1 show that references to poverty in Newsweek are more closely associated with egalitarianism than with individualism. For both the full and the truncated series, references to poverty are more highly correlated with egalitarianism (and they are statistically significant in both cases): the more egalitarian frames there are, the more references to poverty. In contrast, when looking at the full series, references to individualism seem to be positively (though narrowly missing statistical significance) related to poverty cues. For the shorter series, however, the sign of that relationship is reversed; more references to individualism come at the expense of references to poverty; though, again, that relationship narrowly misses statistical significance. However one looks at it, though, the connection between egalitarian cues and references to poverty is relatively strong and positive. Although it is perhaps useful to think of a metaphorical tug-of-war between competing values like individualism and egalitarianism – and, in the next chapter, I shall use this metaphor to describe public opinion on race – when considering media framing on race, the metaphor is not apt. A proliferation of egalitarian value cues does not come at the expense of individualism cues, and vice versa. The finding that these two types of framing are virtually uncorrelated shows that the media, in this case, do not provide the public with a one-sided view at one point in time and an opposite but still one-sided view of race at another point in time.

conclusions Media coverage of race has not been static. Key facets of the system of American political values, such as individualism and egalitarianism, are emphasized to differing degrees through time. Key themes that have received varying attention in the scholarly literature, such as poverty, were not prominent features of mainstream press coverage of race until the mid-1960s. Do the ebbs and flows of media coverage stem from any systematic influences? Are changes in coverage, for example, related

0521821711c02

54

0 521 821711

June 2, 2003

8:6

Mass Media and American Racial Attitudes

to earth-shaking political events that reorient news coverage for years or decades to come? Or, perhaps, are they related to straightforward changes in society itself, which the media merely reflect? Might such changes be influenced by the state of the economy? Or are they a function of feedback from changes in racial policy? These variations through time are understandable on a superficial level; they seem to be related to our political memories, explainable by forces in the social and political world. Having described the evolution of media framing over the past halfcentury, we now turn our attention to describing the evolution of public opinion on race, particularly in the world of policy preferences, the subject of Chapter 3.

0521821711c03

0 521 821711

June 3, 2003

9:35

3 Eras of Racial Liberalism and Conservatism

People think about politics. Not often. Not systematically. But they do. – James A. Stimson (1991)

The disagreements between those who trace the origins of racial policy preferences to ideological commitments and those who trace them to the presence or absence of prejudice are strong and long-standing. But those disagreements are not total. Adherents of both perspectives have traditionally found common ground in the belief that racial policy preferences are stable. To be sure, this agreement stems from varying sources. Those whose explanation for racial attitudes focuses on prejudice may assume that racial policy preferences are stable because prejudice itself is resistant to change, and scholars emphasizing the role of ideology assume that the ideological commitments that drive racial policy preferences are similarly inertial. But despite the fact that they agree for disagreeable reasons, there has been scholarly consensus on the stability of racial policy preferences. In fact, public-opinion researchers of all theoretical stripes have traditionally believed that racial policy preferences (along with politicalparty orientation) are among the most stable elements of a person’s system of political beliefs. No other cluster of attitudes, it has been believed, approaches these two in terms of their lack of volatility. The spring from which all of this research flows is, again, the famous 1964 work of Philip Converse. 55

0521821711c03

0 521 821711

56

June 3, 2003

9:35

Mass Media and American Racial Attitudes

But just how stable are policy preferences on race? A close look at the data in this chapter will reveal considerable fluctuation over the past half-century. Perhaps, considered in the light of the findings from Chapter 2 that press coverage of race has fluctuated so much, this will not strike the reader as surprising. But it is not a finding that has been well documented in the academic study of racial policy preferences. Before turning to the data, however, let us review a relatively new strain of research that suggests a new model of racial policy preferences at the individual level. Later, that model will be aggregated across time and individuals; and in subsequent chapters its implications will be tested.

a new view of racial policy preferences Although Converse’s seminal work has clearly defined the field, beginning with the work of political scientist Robert Lane there has always been a portion of researchers who viewed the nature of public opinion differently. That is, not everyone has conceived of the issue as a constraint-versus-no-constraint dichotomy. Much as closed-ended survey questions do not always do justice to the nature of political attitudes, so many argue that Converse structured the debate about political attitudes artificially. Lane, in particular, thought that Converse’s concept of ideology, or “belief systems,” was too narrow, restrictive, and inflexible. He argued that people have personal ideologies that fit their personal experiences.1 Slowly, other scholars began to build on Lane’s insight. In her study of why there is no strong tradition of mass-based leftism in America, the political scientist Jennifer Hochschild demonstrated that individuals’ political attitudes are not necessarily clearly defined in straightforward “liberal” or “conservative” ideological terms. Rather than conducting a traditional national survey, she spent many hours conducting in-depth interviews with twenty-eight adults, learning about how they formed their own political beliefs. In the process, Hochschild found that most people do not hold simplistic, either-it’s-black-or-it’swhite types of beliefs. To the contrary, she found that many people are 1

See Lane 1962.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

57

ambivalent about important political issues.2 Most of her interviewees saw some merit in both sides of policy debates, refusing to take extreme stands on either side of an issue. She writes: “[G]iven the opportunity, people do not make simple statements; they shade, modulate, deny, retract, or just grind to a halt in frustration. These manifestations of uncertainty are just as meaningful and interesting as the clear, definitive statements of a belief system.”3 In her (admittedly unrepresentative) sample, Hochschild found that expressions of ambivalence were far more common than clear-cut ideological proclamations. Put differently, perhaps Converse found so little ideological thought because most people don’t conceive of their ideology in the way he required them to. But that doesn’t mean that they don’t think about politics in their own personal way. If Hochschild is right, most people do have coherent political opinions. It seems, however, that these opinions are malleable, fuzzy, and sometimes self-contradictory. This idea has been extended into a full-blown, more formalized theory of public opinion by John Zaller and Stanley Feldman, who have abandoned the idea that responses to closed-ended survey questions, in some straightforward fashion, are descriptions of a preexisting preference for a particular policy. Traditional survey research has implicitly assumed that the process of discovering a person’s attitude in a publicopinion survey somewhat resembles the way a computer searches for information on a hard drive: the information resides intact somewhere on the hard drive, and the job of the survey is to scan the hard drive to find the relevant bytes of information. That is, attitudes are already there, perfectly formed, in the minds of respondents. The researcher’s job, so the assumption goes, is merely to locate and retrieve them. Zaller and Feldman clearly reject that view of the nature of public opinion. Instead, they argue that: Most citizens . . . simply do not possess preformed attitudes at the level of specificity demanded in surveys. Rather, they carry around in their heads a mix of only partially consistent ideas and considerations. When questioned, they call to mind a sample of these ideas, including an oversample of ideas made salient by the questionnaire and other recent events, and use them to choose among 2 3

“Ambivalence” is not meant to connote “indifference,” but rather the expression of contradictory opinions about one issue. Hochschild 1981, p. 238.

0521821711c03

58

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

the options offered. But their choices do not, in most cases, reflect anything that can be described as true attitudes; rather, they reflect the thoughts that are most accessible in memory at the moment of response.4

This description of the nature of public opinion is far more specific than Hochschild’s, of course, but the explanations share a core idea. Hochschild’s notion of “ambivalence” is very similar to Zaller and Feldman’s notion of what may be called “competing considerations.” Using open-ended probes – where respondents are asked to elaborate on a just-expressed opinion – after asking a closed-ended question, Zaller and Feldman demonstrate that the closed-ended question does not do full justice to the nature of the individual attitude. Most people, when asked to report “what kind of things come to mind”5 when thinking about an issue, expressed some sentiments that would lead them to favor a policy, but also others that would lead them to oppose the policy. (Some highlights of that work will be described later.) In effect, people made “on the one hand . . . but on the other hand . . . ” sorts of judgments. Most saw at least a nugget of truth in both sides of political arguments. This is quite a different picture from that painted by Converse. These findings do not necessarily discredit his conclusions, but at a minimum, the more recent findings cast the older ones in a different light. Converse’s notion that people are generally “unconstrained” in their political beliefs, to many, is a euphemism for “ignorant,” “unsophisticated,” “unintelligent,” “inattentive,” and “unfit for democratic citizenship.” To be sure, by his definition, most people are unconstrained. However, the normative judgments that tended to follow this conclusion no longer seem appropriate in light of new theory and evidence. Instead of apparently contradictory political opinions resulting from political ignorance, they can now be viewed as the result of ambivalence. People do not have preformed opinions on every issue of the day. This is not to say that they do not have opinions; rather, these opinions are not cast in stone. When given the opportunity, people do express meaningful sentiments about relevant political issues; but they do not see issues in black and white. Any opinion condensed into an answer to a 4 5

Zaller and Feldman 1992, pp. 579–80. See also Zaller 1992 for a more elaborate theoretical statement. These words are those used by National Election Study interviewers.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

59

closed-ended question could better be described as “tentative,” “modulated,” “hesitant,” or “complex.” Using this strain of research as a foundation, it is possible to adapt the competing-considerations model of attitudes in general to the specific case of racial policy preferences. From my perspective, if there has ever been an issue in American politics where the mass public can see some validity to both sides of the debate, it is on the issue of race. Racial policy preferences are not – as they have traditionally been understood to be – noncognitive, gut-level reactions largely driven by sympathy (or the lack thereof) with black people; rather, racial policy preferences are tentative, equivocal judgments that are the result of deep value conflict. Although people may make temporary judgments about the desirability of government intervention on racial issues, these opinions, in the main, are always subject to revision. The picture that emerges is one of a divided electorate, but not in the way traditionally believed. Ever since Gunnar Myrdal’s incisive work An American Dilemma highlighted the inconsistency of American values and our treatment of blacks, most scholars have viewed the American electorate as cleanly divided between racial liberals and racial conservatives.6 In contrast, I contend that the electorate is more profoundly divided, each person against him- or herself. In this sense, race is far from an “easy issue.”7 What are these competing considerations that cause internal conflict? They are, I suggest, a direct result of two core American values, individualism and egalitarianism. Individualism refers to the belief that people should get ahead on their own, pull themselves up by their own bootstraps: a person should get what he or she earns, and earn what he or she gets. Assistance from the government (or anyone else) is not required, nor particularly desirable. Individualism has long been considered the distinguishing American value. Alexis de Tocqueville saw it as the feature that distinguished America from its European counterparts, and recent research confirms its importance in American lives.8 Egalitarianism, on the other hand, asserts the fundamental equal value of all people. Therefore, every person deserves an equal opportunity to 6 7 8

See Myrdal 1944. The phrase is from Carmines and Stimson 1989. See Tocqueville 1945 [1835, 1840]; and, for more recent work on the American context, see Bellah et al. 1985; Lipset 1967; Lipset and Schneider 1978; McCloskey and Zaller 1984.

0521821711c03

60

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

succeed in life. If particular individuals or groups are disadvantaged, the government may take a role in leveling the playing field. Research on American values has convincingly shown that substantial majorities of Americans do not choose one or the other position, but subscribe to the tenets of both of these abstract values.9 Most Americans would readily concede that blacks have suffered a disadvantaged history. Long after slavery was abolished, blacks continued to be the victims of discrimination, particularly (but not exclusively) in the South, where, after Reconstruction, segregation and discrimination were enforced by the power of law. In many ways, blacks were powerless to improve their own situation in life. Certainly, the playing field between blacks and whites has not been level historically. These facts are not in dispute. But what should be done about America’s legacy of slavery and discrimination? Here, the popular consensus ends. For the prescriptions of individualism and egalitarianism – both of which reside in the hearts of most Americans – differ. Individualism, not surprisingly, dictates that blacks, now free from the legal bonds of discrimination, must get ahead on their own. No governmental assistance is desirable in this view. In contrast, egalitarianism prescribes that all people must be given a fair chance, and therefore if blacks have been systematically denied an equal chance, something must be done to rectify past wrongs. Thus, the values that most Americans cherish pull them in different directions on race. At the individual level, there is some evidence to support this claim, although it is of a piecemeal variety. Social psychologists Irwin Katz and R. Glen Hass confirm this quite directly, when they assert that “ . . . Blacks are perceived as deserving help, yet as not doing enough to help themselves; and both attitudes may exist side by side within an individual.”10 In a clever experiment, they found that experimental subjects were more likely to express more “Pro-Black sentiments” when primed with a scale that represented egalitarian values; conversely, being primed with an individualism scale led subjects to express fewer pro-black sentiments.11 That is, merely asking people questions about individualism reminded them of their individualist moorings, which 9 10 11

See McClosky and Zaller 1984; Kinder and Sanders 1996, esp. tables 6.1 and 6.2. See Katz and Hass 1988, p. 894 (emphasis added). See ibid. 1988, esp. p. 899.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

61

subsequently led them to express less sympathy for the plight of blacks. And merely asking people about their feelings about egalitarianism had the opposite effect: bringing their egalitarian values to the forefront of their minds made pro-black sentiments more likely. This result reveals several important things about the nature of racial attitudes. First, racial attitudes are pliable, not permanently fixed, within an individual. Second, exposure to different strains of the core American ethos evokes different reactions in the minds of Americans.12 There are other bits of micro-level evidence consistent with the idea that Americans feel value conflict about race. In a 1987 University of Michigan survey, respondents were asked a fairly typical survey question tapping into whether they thought the federal government should give aid to blacks, or whether they thought blacks should get ahead on their own. In their attempt to penetrate an individual’s thought processes, though, the researchers added an unusual twist. Before the respondent could answer the question, the interviewer interjected, “Before telling me how you feel about this, could you tell me what kinds of things come to mind when you think about aid to blacks?” Responses were transcribed verbatim and later coded for content – mostly, reasons to support and/or to oppose the policy effort. The object of this open-ended probe was to explore the depths of the respondent’s opinions. John Zaller’s analysis of this item shows that only 29 percent of respondents had no “conflicting considerations” – that is, 29 percent either made entirely liberal or entirely conservative comments (consistent with purely ideological thought) or reported no considerations at all. The remaining 71 percent all made at least one comment in support of government aid to blacks as well as at least one comment in opposition to government aid to blacks. And fully 24 percent gave three or more sets of competing considerations in response to the open-ended probes. A sizable majority of Americans, according to this evidence, 12

In addition, Katz and Hass serendipitously discover that most people are neither pure egalitarians nor pure individualists (for similar findings, see also Kinder and Sanders 1996, table 6.4). In all but one of their experimental groups, the individualism and egalitarianism scales were uncorrelated. That is, being an individualist did not make one less egalitarian, and vice versa. This counteracts an important – but, it appears, mistaken – assumption of most of the cross-sectional literature, namely, that people are either individualists or egalitarians, but not both. Katz and Hass’s evidence is consistent with the idea proposed herein that most Americans clearly adhere to both sets of values.

0521821711c03

62

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

sees some merit to both sides of government intervention on behalf of blacks; but this same majority is not entirely comfortable with the idea, for these same people can also think of good reasons to oppose such action.13 Paul Sniderman and Thomas Piazza have also provided individuallevel evidence in support of this perspective. Using an experimental technique that was embedded in a national survey, Sniderman and Piazza began by asking respondents whether they thought the government should “increase spending for programs to help blacks” or whether “blacks should rely only on themselves.” After hearing their responses, Sniderman and Piazza investigated the degree to which respondents could be talked out of their initial positions. Those favoring government activism were asked: “Would you still feel that way even if government help means people get special treatment just because they are black or would that change your mind?” And those initially opposed to government assistance were asked: “Would you still feel that way even if it means that blacks will continue to be poorer and more often out of work than whites or would that change your mind?” In the face of these relatively mild counterarguments, over 40 percent of whites changed their minds and claimed to prefer the other alternative. Sniderman and Piazza replicated this finding with policy questions about government action to prevent job discrimination, enforce fair-housing laws, and impose affirmative action. Even on the last, which is portrayed in the popular press as a hot-button issue on which all Americans have strong opinions, fully 20 percent changed their minds about affirmative action when asked one simple follow-up question that brought a different consideration to mind.14 All of this research is consistent with the proposition that many, perhaps most, Americans feel deeply divided about racial issues and are at least partially sympathetic with both the liberal and conservative sides of the debate. People can, and do, change their minds on matters of race policy. People have core values that pull them in different directions on race. In particular, the work emerging from social psychology that has been synthesized in the work of John Zaller suggests that the field of cues surrounding an individual will tilt him or her in particular, 13 14

See Zaller 1992, pp. 60–2. See Sniderman and Piazza 1993, chap. 6.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

63

predictable directions. As we shall see in Chapter 4, this leads to particular hypotheses about the causes of aggregate shifts in liberalism and conservatism on race. For now, however, it has yet to be seen that such shifts between overall liberalism and conservatism even exist.

the ebbs and flows of racial policy preferences Intuition about political history and eras tells us that the 1960s was a time of liberalism. We note the passage of the Civil Rights Acts of 1964 and 1968, the Voting Rights Act of 1965, and the landslide election in 1964 of Lyndon Johnson, the champion of both pieces of legislation, who, in homage to his assassinated predecessor, promised to make civil rights for blacks a central component of his administration. This era of liberalism on civil rights, however, did not go unchallenged. The 1968 election of Richard Nixon and the candidacy of George Wallace suggest the beginnings of a conservative backlash. Further, the 1980 election of Ronald Reagan represented a basic shift away from continued liberalism on civil rights. Reagan bemoaned the use of affirmative action and “quotas,” opposed the busing of schoolchildren to achieve integration, and attacked the welfare state in general, which some perceived to be a veiled assault on civil rights. Shockingly, despite our wealth of intuition, there has been no scientific evidence to verify, falsify, or modify it. Public-opinion research that focuses on over-time change of necessity requires repeated measures of identical questions over time. But, in the domain of race, the survey record is surprisingly thin.15 In large part, this is because the focus of survey questions has shifted along with the current debate on racial policy. During the 1950s and 1960s, quite understandably, survey questions were more likely to focus on desegregation; in recent years, issues like affirmative action have become more common features of survey research. In addition, in many of the cases where questions have been repeated, there are long gaps between administrations of identical items. This is especially true before the General Social Surveys, which have operated on a nearly annual basis, began in 1973. Prior to 1973, gaps of four years between administrations of an 15

See Schuman and Presser 1981 for a strong argument about the necessity of identically worded survey items.

0521821711c03

64

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

item were not uncommon. Most important, for the 1950s and 1960s, when race was front-page news on a regular basis, there are gaps in the record of public opinion. The compelling nature of the evolution of racial attitudes and policy preferences in America, the answers to the questions about how and why policy preferences on race today differ from those of a half-century ago, demands that we look over the horizon of time. The cross-sectional snapshot of public opinion that dominates the study of racial policy preferences is a valuable approach and has taught the scholarly community much about the forces that shape racial attitudes. But, as with any snapshot, we are left wondering how things looked before the picture was taken and what happened afterward. That part of the story – the dynamic part – has been missing from the study of racial attitudes. To date, no satisfactory remedies to these problems have been developed and applied to the racial attitudes literature. Most researchers proceed to examine all available data and then paint a descriptive picture of the public’s preferences about a variety of policies at various points in time.16 A typical portrait of aggregate racial policy preferences begins with a single issue, say busing, describes the results of several polls that assess support for busing, then moves on to the next issue, repeating the process. But, again due to data inadequacies, none of these portraits can be painted with particularly rich detail. How, then, can we investigate aggregate preferences on race over time? Because of gaps in the data, it is not possible to choose a single indicator as “the” measure of racial policy preferences over time. That is, if one were to analyze preferences about affirmative action, no single time series has enough data points to be of substantial use in a statistical analysis. The hard truth is that there is simply no question of relevance to racial policy that has been asked consistently and regularly for the last half-century. The “regularly” part of the last sentence carries a lot of weight: of evidence for the ebbs and flows of racial policy preferences – on, presumably, an annual basis – the survey record is empty. Therefore some type of summary measure, some type of index of overall racial policy preferences, must be created to combine data from different time series, or else no analysis is possible. Unless we are 16

For examples, see Schuman et al. 1997; Page and Shapiro 1992; Mayer 1992; Sigelman and Welch 1991.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

65

willing to take that step, any comprehensive, overarching analysis is over before it begins. Aside from making an annualized analysis possible, combining opinion data from multiple sources has another advantage. Defining the construct broadly, even of necessity, has benefits. It is more theoretically powerful to talk about “racial policy preferences” as opposed to, say, “preferences about busing.” To form an explanation that more broadly includes preferences about busing, affirmative action, and the other racial policy issues that abound, is scientifically appealing. But it also poses real problems, because no survey items exist at this level of abstraction: survey items uniformly ask about specific issues. Political scientist James Stimson develops, and I adopt here, a methodology that addresses these predicaments.17 Rather than viewing a time series of a survey item concerning busing solely as information about that very specific issue at different points in time, Stimson views items such as these as indicators of a more general concept, which, for his particular case, he calls policy mood. Policy mood – a series that will be put to analytic use later in this chapter – is defined as a global predisposition of the mass public to favor more or less government involvement to solve society’s problems. Stimson demonstrates that multiple time series representing diverse issues – from the environment to intervention in the economy to education – all move together through time, providing evidence that a single “mood” exists. In other words, when the public is more liberal on educational issues, for example, it is simultaneously more liberal on environmental issues as well. Policy mood, then, is a combination of survey marginals from a broad variety of different policy areas. Of course, this methodological technology can be applied to the domain of race. And, for the question at hand, if the broad concept of “racial policy preferences” is meaningful, then various indicators of those preferences should all move in parallel fashion through time. If the concept is not meaningful, then the various series should move independently of one another. I have compiled nineteen diverse time-series items on racial policy preferences. Wording of questions, number of administrations, and the survey organization are listed in the Appendix. The content of the 17

See Stimson 1999.

0521821711c03

66

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

figure 3.1. Four Indicators of Racial Policy Preferences, 1970–1996

items varies, but series I have collected include questions about busing, government spending on blacks, government intervention to prevent race discrimination in jobs, and whether the government is pushing civil rights “too fast,” among others. It represents a nearly complete listing of questions on racial policy that have been asked at least twice by the same survey organization.18 It is perhaps helpful to ask: can any initial evidence of common movement, like that which Stimson discovered, be found in these various series? Figure 3.1 displays four of the series from 1970 to 1996, plus a simple weighted average of the four. In each case, higher values represent more liberal aggregate preferences. Three of the series are taken from the General Social Surveys, the fourth from the University of Michigan’s National Election Studies. The four are: busing, minority aid, spending on blacks, and school segregation. All four have been standardized – that is, shown as deviations from their respective 18

Schuman et al.’s work on racial attitudes (1997) provides an extensive list of questions about racial attitudes (see especially chap. 3). I include most (but not all) of their items that deal with racial policy preferences, but deliberately omit items that are not manifestly about racial policy. For example, because I focus on racial policy preferences, I do not examine questions of social distance or about explanations for black–white inequality. A few items that I include are not found in Schuman et al.’s book. In addition, the nineteen items represented here include an eclectic mixture of absolute items and relative ones; because of the small number of total items, however, using the method described below, it is not possible to separate these items out into discrete indices.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

67

averages – in order to show the truly parallel nature of the different items.19 It appears that these four items share a considerable amount of common movement over this time period – an interesting finding, given that the items cover fairly different issue areas. The items all started out at relatively liberal points in the early 1970s, but simultaneously began drifting in a conservative direction throughout the decade, hitting a conservative peak around 1980. Then, after the election of Ronald Reagan, each began to drift back in a liberal direction, apparently peaking in the early 1990s, since which time they have drifted back toward conservatism. It is worth emphasizing that the movements in these series are substantial: at the conservative zenith around 1980, for example, each of these four series hit a point approaching two standard deviations below its respective mean, representing real shifts in opinion ranging from 12 to 18 points. This parallel movement strongly suggests that there is a “spirit of the times” that influences specific racial policy preferences of all varieties. If the preliminary evidence in Figure 3.1 is not misleading, then it is no longer necessary to examine each time series of racial policy preferences separately. As a result of the visual evidence in Figure 3.1, I have combined the various series into a single annual time series using Stimson’s algorithm, measured as far back in time as the data will allow – from 1950 to 1996.20 The resulting series can be seen in Figure 3.2, where higher values again represent more liberal aggregate preferences. Throughout the 1950s and into the 1960s, the American public slowly, sometimes haltingly, but increasingly favored an activist government to bring about racial equality. Liberalism peaked around 1970, and – consistent with the movement seen in Figure 3.1 – the electorate began a slow but steady move toward conservatism on race throughout the 1970s; but then the trend was reversed around the first election of Ronald Reagan. The public became more liberal throughout the decade of the 1980s, which even eclipsed the liberalism of the 1960s. Data through the mid1990s indicate that this most recent trend has reversed again. This 19 20

Although the series have been standardized for the purposes of graphical presentation, each will remain in its native metric in the analyses to follow. The original algorithm is described in Appendix 1 of Stimson 1999 2nd Ed. The algorithm used here, however, is a slightly different version, and is described in detail in Stimson 1994.

0521821711c03

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

68 80

75

Liberalism

70

65

60

55

50

96

94

19

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

52

19

19

19

50

45

Year

figure 3.2. Racial Policy Preferences, 1950–1996

description lends empirical support to the notion that public opinion varies understandably, not haphazardly or randomly, over time. How well does this index fit the component series? Is it truly an average of the many diverse series, or is it dominated by one or two individual series? In other words, how broad is this composite measure of racial policy preferences? The answers to these questions are found in Table 3.1, where the correlations between each specific indicator of aggregate racial policy preferences and the final index can be found.21 Most of the items load strongly on the factor, with eight of the items having correlations with it above 0.80. Indicators from all subdomains of race politics correlate highly with the index. The fact that aggregate movement in preferences about government guaranteeing equal access to public accommodations runs parallel with preferences about affirmative action, for example, is a novel finding. The striking pattern of correlations, then, suggests that there is an overall, aggregate, policy sentiment on race that encompasses all of the individual racial issues 21

I have presented neither a correlation matrix of the nineteen items nor a principalcomponents analysis. That is because no such analyses are possible due to the volume of missing data. Factor analysis, in particular, cannot handle missing data at all, and a correlation matrix with pairwise deletion of missing data would be misleading, for several of the items do not overlap at all with others, and, more often, there are only a few points in common. Stimson’s methodology provides a way around this problem, in that it is analogous to (dynamic) factor analysis with missing data. The correlations presented in Table 3.1 between each indicator and the index can be interpreted as communality scores.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

69

table 3.1. Correlations of Indicators of Racial Policy Preferences with Overall Index

Item NES/accommodations Harris/affirmative action (Education/employment) Harris/affirmative action (Women/minorities) Trendex/aid minorities GSS/busing Gallup/integration too fast GSS/help blacks NES/aid minorities NES/busing NES/fair treatment in jobs (short) NES/fair treatment in jobs (long) GSS/improving conditions of blacks GSS/assistance to blacks CBS-NYT/affirmative action NES/help blacks NES/residential openness GSS/open housing Roper/ghettos, race, and poverty NES/school segregation

Sample Size

Correlation with Index

4 3

0.880 0.454

9

0.390

46 15 23 11 10 5 3 3 20 11 14 5 5 13 14 9

0.974 0.805 0.985 0.626 0.585 −0.521 0.993 0.405 0.930 0.866 0.358 0.200 0.386 0.875 0.126 0.148

Note: See the Appendix for precise wording of questions.

such as busing, segregation, and the like. When the American public is conservative on affirmative action, in other words, it is simultaneously conservative on other issues of race policy. Similarly, periods of liberalism tend to go hand in hand.22 There are a few exceptions to this overall pattern, however. One item – the National Election Study busing item, which has been asked five times – was actually negatively correlated with the index. Does this mean that busing represents a separate issue dimension, uncorrelated with other domains of race policy? No. To the contrary, the General Social Survey version of the busing item, which has been asked fifteen 22

The index explains 51.8 percent of the variation in all of the nineteen individual items, a figure that compares favorably with that for Stimson’s Policy Mood measure, the first dimension of which explains less than 40 percent of the variation in those survey marginals. (See Erikson, MacKuen, and Stimson 2002.)

0521821711c03

70

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

times, correlates at a very strong 0.805 with the overall factor; so there is nothing about busing per se that leads to this result. The most sensible explanation for these seemingly odd findings, then, is sampling error. Both the NES busing item as well as the other indicators that do not load well on the index – two of the items load below 0.2 – tend to be those with the fewest data points in the series. In the shorter series, one skewed sample can have a significant effect on the correlation between that indicator and the composite series. And the items that are weakly or negatively correlated with the index likely do so for this reason. It is important to emphasize that the movement in Figures 3.1 and 3.2 represents sizable shifts in aggregate sentiment, although the scope of this movement is somewhat obscured by the artificial metric of the algorithm. When examined individually, many of the components of the index of racial policy preferences span ranges of over twenty percentage points in support for liberal policy – and this is especially true of items that cover longer time spans. Conversely, it is rare for an item not to have at least a ten-point difference between its minimum and maximum values.23 A clear picture has emerged: there is sizable and systematic movement in racial policy preferences over time. A single, longitudinal measure captures aggregate movement in racial policy preferences quite well.24 Of course, this does not imply that racial policy preferences are unidimensional at the individual level. Donald Kinder and Lynn Sanders, in particular, suggest that there are three dimensions of racial policy preferences.25 But this is not inconsistent with common 23

24

25

Because the index contains only nineteen items, it is natural to wonder if it is overly sensitive to the particular items used in any given year. If this were the case, the resulting index would differ substantially if any one of the nineteen items were dropped. To test this proposition, I have recalculated the racial policy preferences index nineteen separate times, each time eliminating just one of the original nineteen items (that is, the new nineteen series are unique combinations of eighteen of the nineteen original items). The results (available from the author upon request) are instructive. I correlated each of the nineteen new series with the full racial policy preferences measure. The average correlation between the full measure and each eighteen-item index is 0.975, indicating that the full index is not dependent on any one of the individual items that comprise it. The latest version of Stimson’s algorithm (available at http://www.unc.edu/∼jstimson) allows for the possibility of two-dimensional solutions. Multidimensional solutions did not improve the fit to the data at all, and did not provide an interpretable second dimension. See Kinder and Sanders 1996, chap. 2.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

71

movement through time of these distinct individual-level dimensions; things that are separable at the individual level may still, because of outside forces, drift together through time. And their common movement represents something that is politically meaningful – something at least partially consistent with the intuition that most of us have about the subject. Now, however, our intuition can be checked against actual data of a scientific variety.

parallel publics The analysis presented above is distinct from most of the publicopinion literature on race because of its longitudinal flavor. But it is also unique for one other, more subtle reason: people of all races are analyzed simultaneously, as one superaggregated group. Whereas the standard operating procedure is to parse out survey respondents of different races – with most, but not all, analyses focusing on whites’ attitudes – I make no such effort here, having decided instead to look at all Americans’ opinions together. This is partly (but not entirely) a function of the ways in which such data tend to be available to researchers: survey archives tend to show survey marginals for the population as a whole, and not for subpopulations. My other motivation is substantive. In particular, given the time-series focus of this research – trying to explain ebbs and flows in liberalism over the course of years and decades – there is no particular justification for separating out people by race a priori. Surely, when doing research in the cross-sectional tradition – attempting to explain why some people are liberal and others conservative – it would be foolhardy not to account for a survey respondent’s race when examining their policy preferences on racial matters: differences in black and white opinion, in particular, are well documented.26 For example, black Americans, as we shall see again below, consistently favor busing programs to integrate schools more than do white Americans. But when the focus changes to opinion shifts through time, what is the reason for assuming that black and white opinion will not drift back and forth in parallel fashion, with blacks becoming more liberal 26

For the most recent example, see Kinder and Sanders 1996, esp. chap. 2. See also Sigelman and Welch 1991.

0521821711c03

72

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

at the same time as whites, and more conservative at the same time as whites? My suspicion – and it is really only that – is that most researchers do not believe that black opinion varies all that much.27 That is, most researchers, without having looked at the data, would implicitly assume that black opinion on busing or affirmative action remains relatively consistent over time. It would be assumed that blacks unwaveringly support such liberal policies. Any over-time variation of the sort just discovered, then, must be due to variability in white attitudes. But is this true? In a word, no. The best evidence to date on this front – that of Benjamin Page and Robert Shapiro – indicates that black and white Americans form, in their words, “parallel publics.”28 They acknowledge the unsurprising fact that there exists a substantial gulf between the opinions of whites and blacks on many issues – particularly, of course, issues revolving around race. Blacks are persistently and significantly more liberal than whites on race. But in their examination of opinion shifts over time, Page and Shapiro conclude: “ . . . [O]ur main finding is that – even on issues of special concern to blacks – their opinions have tended to move in parallel with the opinions of whites.”29 When black opinion tends to drift in favor of busing, they conclude, so does white opinion. When black opinion tends to drift back toward opposition to busing, so does white opinion. Page and Shapiro’s conclusions are largely confirmed by a look at some of the opinion time series that are broken down by race. Figure 3.3 presents the percentages of white and black respondents to the General Social Surveys conducted between 1972 and 1996 that favor government busing to achieve integration in schools. The most striking feature of the figure, of course, is that blacks favored busing far more than whites – in some years, the gulf is over 45 points. Looking past that point, however, one is able to detect quite similar year-to-year fluctuations in black and white policy preferences, fluctuations that are largely consistent with the trends discovered in Figures 3.1 and 3.2. From the conservative point for both series in the late 1970s, both 27 28 29

Indeed, in the area of race, I am simply unaware of any single study that aggregates the opinions of whites and blacks into a single group. See Page and Shapiro 1992. The title of this section is borrowed from the title of chapter 7 of their book. See also Stimson 1997. Page and Shapiro 1992, p. 300.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

73

70

60

Percent Liberal

50

Blacks

40

Whites 30

20

10

96

95

19

94

19

93

19

92

19

91

19

90

19

89

19

88

19

87

19

86

19

85

19

84

19

83

19

82

19

81

19

80

19

79

19

78

19

77

19

76

19

75

19

74

19

73

19

19

19

72

0

Year

figure 3.3. Black and White Preferences for Busing, 1972–1996 (Uniform Axis) 65

35

Blacks

30

25 55

20

15

50

Whites

10

Percent Liberal (Whites)

Percent Liberal (Blacks)

60

45 5

96 19

95

94

19

19

93 19

92 19

91 19

90 19

89 19

88 19

87 19

86 19

85

84

19

19

83

82

19

81

19

80

19

19

79

78

19

77

19

19

76 19

75 19

74 19

73

0 19

19

72

40

Year

figure 3.4. Black and White Preferences for Busing, 1972–1996 (Separate Axes)

black and white opinions on busing become steadily more liberal during the 1980s. For both races, these shifts are substantial – a bit over 15 points for whites, a bit under 15 points for blacks. Moreover, support for busing appeared to plateau for both blacks and whites in the early 1990s, and, though there are not enough data points to confirm it definitively, support for busing seems to have eroded among both blacks and whites as the 1990s progressed. Because the gulf between black and white opinion is so visually striking, Figure 3.3, if anything, understates the degree of parallelism in those series. Figure 3.4 allows this common drift to be seen more

0521821711c03

0 521 821711

74

June 3, 2003

9:35

Mass Media and American Racial Attitudes

table 3.2. Correlations beween White and Black Opinion Trends Correlations Survey Item

In Levels of Liberalism

In Year-to-Year Changes

GSS/busing GSS/help blacks GSS/open housing

0.67a 0.79a 0.60a

0.38a 0.41a 0.25a

a

p < 0.05.

easily by plotting the same data on two separate axes. The visual impression there is very strong. The one anomaly is the large drop in black support for busing between 1974 and 1975 – a year in which white support dropped as well, only by a much smaller margin. In other years, however, the degree to which black and white opinion moved in tandem with one another is undeniably large. None of this is unique to preferences about busing. In particular – despite the greater levels of liberalism by blacks – both in absolute levels, as well as in year-to-year fluctuations, black and white opinion on matters of race policy moved in concert. Further evidence of this fact can be seen in Table 3.2, which presents the statistical correlations between black and white opinion on busing as well as preferences on government spending to help blacks and open-housing laws. The correlations can vary from 0 to an absolute value of 1.0. The column on the left shows the correlation in absolute levels, and the column on the right shows the correlation between year-to-year changes. The latter test is particularly stringent, because the yearly black subsamples tend to average only around 150 respondents, which increases random errors associated with sampling fluctuation. The table shows strong and statistically significant correlations between black and white policy preferences on race. The positive coefficients in the left-hand column indicate that, consistent with Figures 3.3 and 3.4, when black opinion is at its most liberal points, so is white opinion; and when black opinion is at its most conservative points, so is white opinion. The smaller (but still positive) coefficients in the right-hand column indicate the slightly different finding that when black opinion becomes more liberal than it was the year before, so does white opinion; and when black opinion

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

75

becomes more conservative than it was the year before, white opinion tends to become more conservative as well. None of this is to say that black and white opinions are in any sense equivalent. The evidence in Figure 3.3 clinches the argument that blacks and whites have markedly different preferences about school busing, and the same is true for the other race-policy areas. But when it comes to over-time analysis, lumping black and white respondents together is easily justifiable. Actually, it would be more difficult to justify separating blacks and whites, for little support for such a procedure can be found in the data. As a result, all subsequent analyses will use the combined survey results for all respondents.

race and the welfare state: evolutionary patterns The over-time dynamics in racial policy preferences that are revealed in the above analyses suggest other questions as well – first among them, how (if at all) do shifts in racial policy preferences coincide with shifts in other policy domains? In particular, it is natural to wonder if shifts in Americans’ policy preferences about race coincide with shifts in their policy preferences on the New Deal/welfare-state cluster of issues. This harks back to the controversy between those scholars who, when analyzing individual-level data, view racial policy preferences as a function of an individual’s ideology, and those who view them as a function of underlying prejudice. Then, I argued that the evidence about the relationship between preferences about activist versus laissez-faire government and policy preferences about race was mixed. What does the view look like when examining this relationship over time? Do fluctuations in Americans’ preferences for activist government generally speaking move in tandem with their preferences about activist government in the domain of race? Are the two unrelated? Or is the relationship more complex? In recent years, the scholarly debate on the relationship between race and the welfare state has been dominated by the pioneering work of political scientists Edward Carmines and James Stimson. Their work on “issue evolutions,” while primarily driven by the theoretical question of how issues compete for prominence in the party system, documents how race came to supersede the welfare state in terms of organizing

0521821711c03

76

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

forces that drive American politics. That is, Carmines and Stimson argue that, beginning with the Great Depression, the New Deal/ welfare-state bundle of issues was the defining fault line in American politics – in effect, the issue that everyone knew divided the Republicans from the Democrats. Cracks in this fault line appeared some quartercentury later, as party elites (and then party activists) began realigning around the issue of race as early as 1958. The mass public, though, did not recognize this new alignment until the presidential contest of 1964 between Democrat Lyndon Johnson and Republican Barry Goldwater. But that campaign made the new fault lines clear, and since that point in time, divisions in American politics have increasingly been over the issue of race.30 Recent work by political scientist Martin Gilens has produced evidence that, at the individual level, whites’ attitudes about welfare policy are largely determined by their attitudes about race. Instead of assuming that racial attitudes and preferences about the welfare state compete with one another, this research examines the possibility that these sets of attitudes are mutually reinforcing. In particular, Gilens investigates the extent to which whites’ attitudes about welfare are a function of stereotypes about blacks. Controlling for other possible influences, such as core values and self-interest, he finds that racial attitudes are the single best predictor of preferences about welfare policy. Gilens clearly shows that assuming that welfare-state issues are independent of attitudes about race is wrong. Attitudes about race and the welfare state are very closely related – more closely related than had been suspected.31 This research is moving the study of public opinion in the right direction, for it recognizes the interconnectedness of Americans’ policy preferences, instead of seeing beliefs about race and beliefs about the welfare state as separate and unrelated. However, it does not answer all of the questions about the relationship. In particular, Gilens paints a clear portrait of the nature of the relationship between racial and welfare-state attitudes in recent years; the surveys he examines are from 1986 and 1994. But we still do not know the process – the how and when of this relationship – of how these sets of attitudes 30 31

See Carmines and Stimson 1989. See Gilens 1995, 1996, 1999; and also Kluegel and Smith 1986.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

77

came to be associated with one another. In sum, what is needed is an examination of this relationship over time. The static part of this relationship is now more clearly understood; the dynamics are still largely unknown. The best recent scholarship on the dynamics of welfare-state attitudes, authored by the political scientist James Stimson, makes a case similar to that of Gilens – namely, that racial policy preferences and what he calls policy mood are inextricably related, really one thing instead of two. Considering the issue of race along with other “social welfare” concerns like education and health care, he writes: “The question at hand is this: Do these represent . . . [separate] areas of controversy, each with its own dynamic of debate, event, controversy, and policy response? Or are they merely separate questions about policy debates that are different in principle but that in fact all turn out to be one?”32 Based on powerful graphical evidence much like that presented earlier in this chapter that shows substantial “parallelism,” Stimson concludes that race and the welfare state, in effect, comprise one set of policy concerns – at least in the minds of the American public. This is a big claim, one that must be addressed seriously given the overwhelming force of data behind it. If he is right, it is a powerful testimony to the power of policy mood, and to the overall unity of public opinion more generally. For one thing to subsume completely the entire domain of racial policy preferences would seem to suggest that the scholarly community has been looking at too many trees and missing the forest. Perhaps a better metaphor would be that, if Stimson’s conclusion is correct, scholars have been standing too close to a pointillist painting, focusing on each one of the multicolored dots. What we need to do, in effect, is take a few steps backward – aggregate across all issues, including race – and we shall see the larger pattern in the dots. Stimson’s conclusion is based on data from the mid-1960s through the 1990s. Therein, as we shall see, lies the problem. For the parallelism that Stimson shows since the mid-1960s largely disappears in the years before that point – years in which the most dramatic acts of the drama of race in contemporary American politics were played out. Extending the analysis back another decade and a half will reveal much about policy preferences, and, as we shall explore in Chapter 5, about 32

See Stimson 1999, p. 79.

0521821711c03

78

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

figure 3.5. Racial and Welfare-State Policy Preferences, 1966–1996

American politics itself. At this point, the best strategy is to allow the data to tell their own story. In order to compare the over-time evolution of attitudes on race and the welfare state, it will be necessary to obtain over-time measures of aggregate public sentiment for each policy domain. The measure of racial policy preferences already described was constructed using Stimson’s methodology. In the case of preferences about the welfare state, Stimson’s measure is readily available. His policy mood measure, as we have already described it, reflects the public’s overall sentiment toward government intervention to solve social problems. It is composed of a wide array of time series on domestic policy concerns, the sum total of which could fairly be called the welfare-state policy agenda. There are items about government involvement in the welfare state, education, environmental protection, and many others. We will use this series as a measure of welfare-state policy preferences.33 How does our series representing racial policy preferences compare with Stimson’s welfare-state series? Figure 3.5, showing the series from 1966 through 1996, essentially replicates Stimson’s findings as described, showing the parallelism that he discovered.34 In fact, the 33

34

One technical note: Stimson’s measure of Policy Mood includes survey items about race, which could potentially confound the analysis here. Therefore, I have recomputed his Policy Mood time series excluding the items on race policy. The two series both use artificial and arbitrary metrics. Hence, the fact that welfarestate policy preferences have higher or lower numbers than Racial Policy Preferences does not indicate that the public is more or less liberal on matters pertaining to the

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

79

figure 3.6. Racial and Welfare-State Policy Preferences, 1950–1965

two series really do appear to be “one thing,” though they are composed of entirely separate survey items. Both series hit high points shortly after Lyndon Johnson’s Great Society was pushed onto the public agenda. Both drift toward a conservative peak in the late 1970s. Both reverse course and become steadily more liberal during the Reagan and Bush (Senior) administrations. And both are heading in a more conservative direction. It is indeed hard to overstate the similarity of aggregate policy sentiment on race and the welfare state during the period after 1965. This is consistent with the individual-level findings that Gilens has produced, and with Stimson’s results. But our data can be extended back in time. And they should be, because of all the important events in race politics that transpired before 1966. Available technology and data allow us to derive annual estimates for both series as early as 1950. Figure 3.6 presents the same series for the years 1950 through 1965. While there is some common movement found there, it is nothing of the sort found in the previous figure for years after 1965. In particular, preferences for the welfare state – represented by the dashed line in the figure – hit a liberal peak around 1961, and began to drift back toward conservatism during a time when racial policy preferences – the solid line in the welfare state than on matters pertaining to race. Because the metrics are arbitrary, no such comparisons between the levels of the respective series can be made. The only meaningful comparisons are for year-to-year shifts.

0521821711c03

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

80

table 3.3. Correlations between Year-to-Year Shifts in Racial and Welfare-State Policy Preferences for Various Time Periods Time Period

Correlation

1950–1996 1950–1965 1966–1996

0.37a 0.03 0.68a

a

p < 0.05.

figure 3.7. Racial and Welfare-State Policy Preferences, 1950–1996

figure – continued their slow-but-steady progress toward liberalism. Year-to-year changes seem to be mostly uncorrelated with one another. Figure 3.7 shows the complete picture, with both series plotted for the full time period, from 1950 through 1996. Some elementary statistical analysis confirms the visual impressions found in the figures. As Table 3.3 shows, year-to-year shifts in racial policy preferences and welfare-state policy preferences correlate, for the full series, at a relatively modest (but statistically significant) 0.37. This, however, surely masks the truth that is evident in the figures. When breaking the series into before-and-after samples, the true nature of the relationship between race and the welfare state becomes apparent. Before 1966, the two series are virtually uncorrelated. But from 1966 forward, the year-to-year correlation leaps to a much stronger 0.68.

0521821711c03

0 521 821711

June 3, 2003

9:35

Eras of Racial Liberalism and Conservatism

81

What this shows, in my view, is that the cross-sectional evidence found by Gilens and the time-series evidence discovered by Stimson about the relationship in the public mind between the two largest issues in twentieth-century America – race and the welfare state – are not so much wrong as incomplete. Gilens shows how intertwined the two sets of attitudes were among individuals in the 1980s and 1990s. Stimson’s time-series data shows how they have drifted together through time since the mid-1960s. But neither of these scholars shows when race and the welfare state came to be associated in the minds of the American public – Gilens because his data is cross-sectional, and Stimson because his analysis does not reach back far enough in time. The evidence presented here answers that question for the first time: the association occurred – as did so many other important things in race politics – in the mid-1960s. Before that time, racial and welfarestate policy preferences meandered independently of one another; since then, they march in lockstep. The cutoff date of 1965 coincides with important events during the civil rights movement, although choosing this year in particular might strike the reader as somewhat arbitrary at this point. However, when we turn to answer the questions of how and why race and the welfare state came to be seen as one issue instead of two, some theoretical flesh will be added to the bare bones presented here.

looking back, looking ahead The analysis presented in this chapter has taken us a long way in terms of description. We now know, for example, that aggregate movement in racial policy preferences can be captured in a single dimension. We have seen back-and-forth movement between liberal peaks and conservative valleys – movement that occurs in a predictable (if not always the predicted) fashion. We have seen shifts that conform to our intuition and historical memory, and at the same time help to cast new light on that memory. For example, the conservative apex that occurred just before 1980, and the subsequent drift toward liberalism, instructs us that indeed the 1980s were a period of conservatism on race, as our intuition would lead us to expect. But it also shows that, because of the timing of the conservative peak, Ronald Reagan did not create a wave of racial conservatism; he rode it. In fact, the tide was beginning to

0521821711c03

82

0 521 821711

June 3, 2003

9:35

Mass Media and American Racial Attitudes

move in a liberal direction even before he took office in 1981. Perhaps it would be comforting for liberals who did not appreciate Reagan’s brand of politics somehow to blame him for the climate of the 1980s, particularly the early 1980s. The findings here suggest that, had Ronald Reagan not been on the political scene, some other similarly conservative politician would have ridden the wave of conservative policy sentiment and surfed on in, so to speak, to presidential victory. This description reveals much about American public opinion in particular and American politics in general. It portrays public opinion as a leading force and belies the myth of a fickle public being led around haphazardly by manipulative politicians. The movements here were too slow, too enduring, to be the product of some Pied Piper scenario. Yet, though we have answered our what and when questions, our findings raise as yet unanswered questions about why and how racial policy preferences have evolved. We do not yet have a sense of why they drift back and forth between liberalism and conservatism as they do. Nor do we yet know how race and the welfare state, which were once seen as separate issues, came to be viewed as one issue in the public mind. Those two questions form the bases of the next two chapters.

0521821711c04

0 521 821711

June 3, 2003

9:42

4 Media Framing and the Dynamics of Racial Policy Preferences

. . . the symbols of revolt are not drawn like musty costumes from a cultural closet and arrayed before the public. Nor are meanings unrolled out of whole cloth. The costumes of revolt are woven from a blend of inherited and invented fibers into collective action frames. . . . – Sidney Tarrow (1998)

After at least a decade of leftward drift, racial policy preferences reached a liberal peak in the mid-1960s. But, as we have seen, the era of liberalism was not sustainable in the long run. After a reversal and a long slow but steady march to the right, the public reached a conservative zenith in policy preferences on racial matters around 1980. That conservatism, too, was short-lived, though few observers have noted this fact: opposition to government activism on race eroded steadily throughout the 1980s. And again, in the early 1990s, the public hit a liberal high, but – in what is by now a familiar pattern – support for liberalism has been waning since then. Why? In search of an explanation, ought we to open the historian’s toolkit and point to particular dates, identify “critical moments” in our history where we, like a sailboat, changed tack? Ought we to search through the records of speeches by influential politicians and societal leaders – for surely the eloquence attained in the pursuit of racial equality, with its rich biblical imagery, is among the pinnacles of rhetorical splendor in the history of the republic – to discover how a particular presidential candidate or movement activist stated something with a new and 83

0521821711c04

84

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

profound kind of power and persuaded plain folk to change their minds? Ought we to pore through the archival records to uncover how a particular Supreme Court decision, congressional roll-call vote, or civil rights march changed everything? Perhaps. Perhaps today’s location on the long arc of American political history is just the sum total of a string of events, each with its own idiosyncratic causes and its own idiosyncratic consequences. Perhaps, in the search for the over-time causes of racial policy preferences, we ought to round up “the usual suspects.” Perhaps we should seek to find the true effects, for example, of the Brown v. Board of Education decision, the March on Washington, the Watts riots, the Bakke v. Board of Regents decision, the Clarence Thomas nomination, and the O. J. Simpson trial, to name just a few. Perhaps there indeed lies the real story of how American politics works, and the way to understand it is to dig deeply into the historical record for the event most plausibly linked to a turnaround in aggregate opinion. But recall something about the evolution of racial policy preferences: notice again, as we did in Chapter 3, that it is a slow mover. Opinion on race is not fickle and does not appear to turn about easily or often. Its periodicity, loosely speaking, is not months or even years but decades. In between a liberal peak and a conservative trough the movement is more or less smooth – a point or two this year, another point or two the next, and so on, until, after a decade or so, we see real movement that has added up to something politically meaningful. What does this suggest about the causal dynamics behind this evolution? It suggests, to me at least, that in our search for causal mechanisms we should venture beyond people, places, and dates, and look instead for systematic causes of the dynamics of racial policy preferences. Instead of historical particularities, we ought to look for dynamic forces behind the back-and-forth shifts in racial policy preferences – forces that themselves move back and forth through time, that might also slowly accumulate and then dissipate. “Critical events” are not like that; they are not slow movers. They are, by their very nature, the very opposite; the world is different today from what it was yesterday.1 My 1

I do not intend to make the grandiose argument that such critical events cannot explain political history. I am merely making the less radical argument that, in the present case of racial policy preferences, historical-idiosyncratic explanations fit the data less

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

85

suspicion is that public opinion, at this level of abstraction and aggregation, is not subject to the influence of such critical events so much as it is a function of other systematic, evolving forces that nudge it both back and forth, in liberal and conservative directions, through time. This chapter represents an effort to build systematic theory toward that end of explaining the over-time dynamics of racial policy preferences.

the causal dynamics of racial policy preferences We saw in Chapter 2 that the way in which the national media frame their coverage of race is not static. The prominence of different core values as portrayed in the American mainstream press ebbs and flows. Egalitarian and individualism cues both find varying degrees of emphasis at different points in time. They do not, strictly speaking, pull and tug at one another, but neither are they constant. And some story lines more naturally lead to one particular kind of framing, whereas other stories have the opposite effect. Stories about segregation rather naturally have had an egalitarian accent. Stories about affirmative action can be framed either way, as we have seen, but are more likely to take on the language of individualism. And we saw in Chapter 3 that American public opinion on race has flowed for years on end in one direction, only to level out, turn around, and begin a long, slow ebb back in the other direction. Small changes in support or opposition of a point or two from one year to the next that, if viewed in isolation, would best be interpreted as “no change,” merely a function of sampling error, are really small waves in an accumulating tide of opinion shift. Over the years, the waves accumulate, altering the shoreline over which they flow, to produce substantial change. We also, in Chapter 3, developed a model of racial attitudes at the individual level based on value conflict and reviewed some evidence that was consistent with that model. And we suggested that the back-and-forth movement of racial policy preferences in a well. In the case of each phenomenon we seek to explain, the question of whether an idiosyncratic or a systematic explanation fits better – and the truth probably lies somewhere between the two extremes most of the time – is an empirical one, to be answered by referring to evidence. In this case, the evidence seems to me to favor the systematic.

0521821711c04

86

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

liberal, then conservative, then again liberal direction is likely to be a function of such value conflict. We have seen that the field of cues does change, and change systematically. We have seen aggregate opinion shift over time, and shift systematically. But as yet the two pieces of the puzzle – the causes and the consequence – have not been linked, either with explicit theory or with evidence in support of it. Almost all Americans would readily concede that blacks have suffered a disadvantaged history, and that, importantly, this disadvantage is largely not of their own making. Slavery provided a rude introduction to America, and, though slavery ended nearly a century and a half ago, it set the stage for race relations in post-slavery America. The Thirteenth, Fourteenth, and Fifteenth Amendments to the Constitution may have ended the Civil War and the official institution of slavery, but the playing field was far from level. First, black Americans were economically handicapped. And second, blacks were still excluded from American life in a myriad ways. They were excluded from the economic mainstream through an economic apartheid that created separate black and white economies. They were excluded from the political realm through poll taxes, grandfather clauses, and literacy tests. These forms of discrimination persisted for nearly another century after the Civil War ended. For most, this look back at the great blemish on American history is uncomfortable; most of us understandably want to remember our triumphs, not our tragedies. Although the backward view may be uncomfortable, it is not particularly controversial. But as soon as the focus turns from the past to the present and the future, the consensus ends. What, if anything, should the government do about this historic burden that has been placed upon the shoulders of black Americans? Here, our values begin to collide with one another. The prescriptions of individualism and egalitarianism differ. Individualism, of course, dictates that blacks, now free from the legal bonds of discrimination, must get ahead on their own without government assistance. In contrast, egalitarianism prescribes that all people must be given a fair chance, and therefore to the extent that blacks have been denied an equal chance, something must be done to rectify past wrongs. The combination of these, in my view, is the state of mind that most Americans possess in their day-to-day lives – one of value conflict. It is a state of mind only

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

87

rarely activated, a portion of long-term memory that is real, but not, for most of us, the product of deep thought on a daily level. Enter the survey researcher, the essence of whose job it is to tap into these very thought structures that lie somewhere beneath the routine surface of our conscious lives. When asked about our social and political attitudes, we reach into the recesses of our memories, often imperfectly. We refer to core values, but, as we have seen, sometimes those values conflict with one another when applied to particular situations like contemporary race politics. We resolve those value conflicts by relying on more recent – and therefore, more accessible – cues in our memories. If, when answering survey questions, respondents fall prey to this so-called recency effect, then this suggests a causal mechanism for the forces that drive public opinion back and forth through time. For if the environment of cues changes systematically over the years and decades – that is, if the balance of individualist and egalitarian cues shifts back and forth – then this, combined with the way in which individuals process information about their conflicted values on race, implies that the likelihood of individuals recalling their individualist and egalitarian moorings will shift. Consequently, opinion will shift as well. Because Americans feel ambivalence about racial policy that is a function of value conflict, their choice between liberal and conservative answers to closed-ended survey questions will depend upon the array of environmental cues that have recently been sent by the national media; their ambivalence is resolved, if only temporarily, by the sampling from the value cues that are easily accessible in recent memory. When that field of cues is disproportionately composed of references to egalitarianism, the balance of competing considerations in the minds of Americans will be tipped in that direction, and, as a result, people will be more likely to express preferences for liberal racial policies. In the 1950s and early 1960s, stories in the media showed blacks being systematically denied their right to vote in most southern states, black children being prevented by the police from attending school with white children, black citizens being denied service at whites-only lunch counters, and nonviolent black marchers being attacked for protesting against these injustices. The media’s message demonstrated the inconsistency between American beliefs in egalitarianism and the ways their society treated blacks. Stories that accentuated egalitarianism in the past remained in the minds of those who listened to, read, or watched

0521821711c04

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

88

them. They became a part of recent memory, easily retrieved for a time. They, in turn, made liberal responses to survey questions more likely. More recent stories that document continued discrimination can have the same effect. When, at other times, the media send messages disproportionately emphasizing the value of individualism, this will tip the balance of competing considerations in the minds of Americans in the opposite direction, tending to make them express more conservative policy preferences. After most of the fundamental legal rights that were the stated goals of the movement had been won, the focus of the movement shifted to thornier issues, namely the means to achieve desegregation (like busing) and affirmative action programs to achieve workplace and educational equality. And this shift generated a shift in the nature of media coverage of civil rights issues toward individualistic themes. Media messages that highlight the inconsistency between the American value of individualism and liberal racial policies lead to more conservative policy preferences.

a first glance at the evidence What evidence can we find to support this theory? We already have over-time measures of all of our key constructs – the media measures from Chapter 2 and the opinion measure from Chapter 3. Having seen them built from the ground up in their respective chapters, we can now take a first glance at them together. Figure 4.1 shows the egalitarian

Egalitarian Cues

Racial Policy Preferences

Egalitarian Cues (Number)

500

75

70 400 65 300 60 200 55 100

50

94

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

58

19

56

19

54

19

19

52

45 19

19

50

0

Racial Policy Preferences (Liberalism)

80

600

Year

figure 4.1. Egalitarian Media Cues and Racial Policy Preferences, 1950–1994

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

89

framing series along with our series of racial policy preferences, both plotted from 1950 through 1994. There is some rather clear visual evidence of a relationship between the two. As predicted by the theory, higher amounts of egalitarian cues are associated with more liberal eras in public opinion on race. Both series begin at relatively low points in their respective time series and achieve a liberal/egalitarian peak in the mid-to-late 1960s, followed by a substantial drift away from those peaks through most of the 1970s. As egalitarian messages in the media began to wane, so too did support for liberal racial policies among the American public. Less easy to spot from the figure is the fact that both series move toward egalitarianism/liberalism during the 1980s. That, of course, is easy to spot in the policy preferences time series; though it is less visually obvious for the media series, the drift is there, and solidly so, with an almost twofold increase in the number of egalitarian cues between the early 1980s and its subsequent peak in the early 1990s. Still harder to detect, because the series end in 1994, is the most recent movement, with both egalitarian media cues and racial policy preferences moving away from egalitarianism/liberalism peaks around 1991. The visual evidence for a connection between individualist media frames and racial policy preferences is presented in Figure 4.2 and is considerably less impressive. In particular, until the early 1970s, when individualist frames became more prominent, there appears to be no relationship – unsurprisingly, given the lack of variation in the

75 Individualist Cues (Number)

200

Racial Policy Preferences

70

150

65

60

100

55 50

Individualist Cues

50

92

90

88

86

84

82

80

78

76

74

94 19

19

19

19

19

19

19

19

19

19

70

68

66

64

62

60

58

56

54

72

19

19

19

19

19

19

19

19

19

19

19

52

45 19

19

50

0

Racial Policy Preferences (Liberalism)

80

250

Year

figure 4.2. Individualist Media Framing and Racial Policy Preferences, 1950– 1994

0521821711c04

90

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

media series. Since the mid-1970s, though, there are hints that the series might be related. Recall that our theory predicts that higher numbers of individualism cues will lead to more conservative policy preferences; therefore we expect that the two should be inversely related. Indeed, we see some evidence of that in Figure 4.2 in the late 1970s, with the rapid growth of individualism cues coinciding with a movement toward conservatism around 1980. The spurt of individualistic cues, though, did not last at those levels into the 1980s, and neither did the conservatism of public opinion, which, as we have noticed, began to become more liberal again. Finally, the recent spike in individualistic frames in the early 1990s may coincide with the recent reversal of racial policy preferences after the liberal peak in 1991. Those first, visual impressions of the data are just that – first impressions. They ought not to sway our opinions in any ultimate sense, and they surely do not clinch the case for or against our theory in any scientific sense. We can, and will, go beyond visual confirmation to a more rigorous quantitative evaluation of the data. This evidence comes in two forms, the first a test for causal ordering and the second to estimate the strength of the discovered relationships. Before moving into a detailed examination of how strong (or weak) the effects of media framing upon public opinion are, we must clarify an important counterpossibility, namely that the direction of influence runs the other way, from opinion to the media. This possibility is both serious and reasonable. It is serious because the theory outlined above predicts that egalitarian and individualistic media framing will drive racial policy preferences through time, rather than create either reverse or reciprocal causal scenarios. If the true causal influence runs from public opinion to the media, but we estimate the effects of the media on opinion, we will quite obviously mischaracterize the nature of the relationship. And the reverse-causality model is reasonable as well, with quite different implications about the nature of public opinion. This causal scenario would begin with the truism that the American media exist in a market situation, where media outlets compete with one another for an audience of readers, viewers, or listeners. Profits, of course, are largely determined by advertising revenues, which are almost completely a function of audience size. Therefore, newspapers, magazines, and television and radio stations that march out of step with their audiences do so at their financial peril. Consumers of media information

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

91

who sense that a particular media outlet is out of touch can always shift their attention to another information source. According to this perspective, then, producers of news information have a strong incentive – lucrative market share – to be sensitive to current public opinion. Time-series analysts have a particular tool that is uniquely suited to resolving this type of chicken-and-egg dilemma – namely, a statistical test known as Granger causality.2 Intuitively, the test begins with the proposition that a time series’s current values are likely to be influenced by its immediately preceding ones. Thus, for example, it assumes that racial policy preferences in 1980 will not likely be radically different from racial policy preferences in 1979. The value of the series in 1979, therefore, serves as a rather good predictor for what its value will be in 1980. And so on. Once we “filter out” the effects of a series’s own past values, can another series help to predict how the first will move through time? If so, it is evidence that, in this particular sense, the second series “causes” the first. This is a stringent test, and it is precisely what the Granger test tells us.3 What we can do, then, is to run the Granger causality tests both ways. That is, we can resolve our “Do eggs cause chickens or do chickens cause eggs?” query and see if we can find evidence that the media causes public opinion to shift and also to see if public-opinion shifts affect the media at all. We might find evidence that one causes the other, or that the other causes the one, or both, or neither. That is what we do in Table 4.1, where we look at Granger test results for the relationships between both types of media framing and racial policy preferences.4 The columns of the table show the dependent variable, and the rows show the presumed cause of it. (The cells on the diagonal, understandably, are empty, because 2

3 4

In fact, in a rather rare display of academic humor, W. N. Thurman and M. E. Fisher (1988) ran Granger-causality tests to resolve precisely this age-old chickens-versus-eggs “debate.” They used aggregate time series of egg production and chicken populations to see whether, over a fifty-four year period, chickens “cause” eggs or eggs “cause” chickens. They concluded that the egg came first. For an excellent introduction to the details of Granger-causality tests, see Freeman 1983 and Freeman, Williams, and Lin 1989. Granger tests can be sensitive to various “trend” properties in the data, the tests for which are rather esoteric. To avoid further muddying the waters with more methodological details, I don’t consider that possibility here, but refer the interested reader to previous work I have done on the subject (Kellstedt 2000), which largely allays the reasonable fear that trends in the data might confound the statistical tests presented here.

0521821711c04

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

92

table 4.1. The Media and Public Opinion: Granger Test Results Dependent Variables Independent Variables Racial Policy Preferences Egalitarian Media Framing Individualist Media Framing

Racial Policy Preferences

Egalitarian Media Framing

Individualist Media Framing

–

0.57

0.81

0.01a

–

0.90

0.66

0.61

–

Note: Each cell represents a different equation that includes two lags of all variables. The cell entries are p-values from block F-tests, which represent the probability that the independent variable does not cause the dependent variable. N = 45 for all equations. a p < 0.05.

we know that each variable predicts itself quite well.) The cell entries represent the probability that the independent variable does not cause the dependent variable. In the left-hand column, for example, the table shows that racial policy preferences are caused by shifts in egalitarian media framing – or, more properly, but very awkwardly, the probability that egalitarian framing does not cause racial policy preferences is only 0.01, easily surpassing the 0.05 cut-off point for statistical significance. This is consistent both with our theory and with the visual evidence from Figure 4.1. On the other hand, the theory receives no support from the cell representing the effect of individualistic framing on racial policy preferences. There, the cell entry of 0.66 signifies that there is about a two-in-three chance that individualistic framing does not cause racial policy preferences. Importantly, the reverse-causal scenario that posited that public opinion might cause media framing to shift finds no support. The top row of the table shows that racial policy preferences cause neither egalitarian framing (p = 0.57) nor individualistic framing (p = 0.81). Therefore, we need not be concerned about either reverse causality or mutual causality in subsequent analyses. The causal flow, where we have found it to be present, runs exclusively from media toward public opinion. This is an important piece of evidence in support of our theory, because it largely rules out the threat of reverse-causality. But although

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

93

table 4.2. The Media and Racial Policy Preferences: Preliminary Regression Analyses Equation Independent Variables Dynamics Egalitarian media framing Individualist media framing Constant R2

(1)

(2) a

0.89 (0.06) 0.01a (0.00) –

5.40 (3.57) 0.86

(3) a

0.89 (0.07) –

0.01 (0.01) 6.98 (3.88) 0.85

0.88a (0.06) 0.01a (0.00) 0.00 (0.01) 5.99 (3.81) 0.86

Note: Each column represents a different equation. The cell entries are slope coefficients, and their standard errors are in parentheses below. N = 45 all equations. a p < 0.05.

it clears up our concerns about causal ordering, Table 4.1 does not give us any information about the strength of the relationship between the media and public opinion. It is possible that the effects found in Table 4.1 are statistically meaningful but practically and politically insubstantial, and, for all intents and purposes, have zero influence. To get a better sense of the strength of the relationships, we turn to more conventional regression analysis. Those results are presented in Table 4.2, which shows three separate regression analyses to estimate the strength of the effects of both types of media framing on racial policy preferences. In column 1 of the table, the egalitarian framing variable alone predicts the movement in racial policy preferences, and it does quite well. The effect is statistically significant, and in the expected (positive) direction, meaning higher doses of egalitarian coverage translate to a more liberal public. The coefficient for egalitarian framing, 0.01, is deceptively small and is a function of the large difference in the variances of the variables. To translate the effect to a more easily interpretable metric, a shift of one standard deviation in egalitarian framing produces an immediate impact of about one-sixth of a standard deviation in racial policy preferences. And this understates the total true effect of egalitarian framing, because that immediate effect is carried forward into subsequent years by the strong dynamics in the model. The row in the table labeled

0521821711c04

94

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

“Dynamics,” which is simply the dependent variable lagged one time period, represents how far into the future the shifts in the independent variables persist. Under most circumstances, a variable like this will range between 0, meaning the shocks in the independent variable die out immediately in the current time period, to 1, indicating that the shocks quite literally never die out. Because the dependent variable in this case is a measure of public opinion, the dynamic coefficients in a sense represent how quickly the public “forget” the media messages it receives. In this case, the rather strong dynamic coefficient of 0.89 means that about nine-tenths of the effect of a shift in media coverage remain a year later, and about eight-tenths remain two years later – because 0.892 = 0.79 – and so on, until the effect eventually decays geometrically to zero. This is a rather long-winded way of saying that the effects of a shift in egalitarian media coverage linger for several years. The second column of Table 4.2 provides confirmation of another finding of the Granger test results of Table 4.1 – namely, that there is little preliminary evidence connecting individualist media cues to racial policy preferences. In fact, the coefficient there, 0.01, is in the opposite direction from what our theory predicted. That is, we suspected that increases in individualistic coverage would nudge public opinion in a conservative direction; what we find in column 2 of the table is exactly the opposite – that increases in individualistic coverage, just as in egalitarian coverage, leads to a more liberal public. The effect, though, is not statistically significant, which is to say that the true effect might be zero. The third column combines the first two, estimating a regression equation with both types of media effects estimated simultaneously. The story there is largely a composite of the first two columns: egalitarian framing shows rather strong and statistically significant effects, magnified by strong dynamics, whereas there is no noticeable effect for individualist framing. Our theory, then, has thus far received partial confirmation.

a multivariate model of racial policy preferences The analyses above showing the connection between media messages, especially of the egalitarian variety, and racial policy preferences may be seriously misleading unless we are careful to control for other forces

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

95

that also might contribute to the evolution of public opinion on race. As sensible as the theory of media influence may be, and as straightforward as the evidence appears to be on the face of it, we cannot be certain that we have discovered a causal connection between media frames and public opinion unless we try simultaneously to account for the other forces that might cause opinion to become more liberal or conservative through time. In technical terms, we risk what statisticians call “omitted-variables bias” in the coefficient estimates reported in Table 4.2. That is, the estimated effect of the media variables may not reflect their “true” values, but instead are polluted with the unaccounted-for effects of the variables that have been (for now) omitted from the equation. In short, the rather strong suspicion that there are other relevant forces which also cause racial policy preferences to move through time should lead us to be somewhat skeptical of the coefficients in Table 4.2. This bias could, in theory, produce estimated effects that are either too large or too small – and this latter possibility is particularly relevant for our analysis of the effects of individualistic framing, for which we have yet to find evidence. Still, they are a good benchmark, and we can see how much they are affected once we “control for” the effects of these other causes. Before we can control for them, however, we must first specify what they are – or at least what they might plausibly be. Six possibly confounding (and possibly complementary) forces come to mind, some of which are suggested by existing research on racial attitudes and public opinion more generally. First, it is reasonable to speculate that other types of media messages might influence racial policy preferences. In particular, the emphasis on the issue of states’ rights might be important. This could be particularly true for the 1950s and 1960s, when it (rather than individualism) was the central “value conflict” with egalitarianism. (Recall the relative paucity of individualistic cues before the 1970s as portrayed in Figure 3.4.) In this earlier period, one of the primary arguments against federal government intrusion on behalf of blacks was that this policy domain was an area more properly left to the states. Because of this argument, which rests on the long history of American federalism, we expect that an increase in the number of references to states’ rights will make the public less sympathetic to federal governmental efforts to bring about equality for blacks. Measuring the number of states’ rights frames with the Newsweek data base is

0521821711c04

0 521 821711

96

June 3, 2003

9:42

Mass Media and American Racial Attitudes

straightforward, including the literal combination of the words “states” and “rights,” most prominently, and other less common (and now largely forgotten) words like “interposition” and “nullification.” Second, it is possible that racial policy preferences do not respond to a particular type of coverage about blacks, but to coverage of any sort about blacks. Perhaps it is not stories that highlight the importance of values like individualism or egalitarianism in the context of race that influence public opinion, but rather the sheer volume of stories on race. In this case, our theory about the sensitivity of opinion to various subtleties of media coverage would be ascribing too much sophistication to opinion – assuming, in a sense, that it responds to gentle nudges, to themes within coverage, when in reality it only responds to the blunt amount of coverage as a whole, ignoring the finer-grained details of the content. In a sense, if this were true, it would be as if the particular content of coverage of race did not matter all that much, so long as something was said. And it is possible to see the effects of the amount of coverage going in either direction; increases in the amount of coverage might generate sympathy for or antipathy toward blacks, either creating or eroding support for liberal policies. The measure used here is simply the numbers of stories about race per year in Newsweek magazine that are shown in Figure 3.1. Other nonmediated sources might reasonably influence racial policy preferences as well. Our third factor draws from the work of political scientist Robert Durr, who theorized that aggregate policy preferences for a liberal or conservative government would be a function of aggregate optimism or pessimism about the future of the economy.5 Durr postulates that periods in which the public anticipates future economic prosperity are associated with an aggregate willingness to pay for liberal (and often expensive) government policies. In contrast, when the outlook for the economic future is less rosy, the public becomes less enthusiastic about bankrolling a liberal government agenda. His theory has an intuitive ring to it that harkens back to Maslow’s hierarchy of needs. In a sense, when we are relatively secure about our individual prosperity in the future, we are more generous with others and more willing to spend our tax dollars on programs that benefit the less well-off. But when we are less confident in the economic future, 5

See Durr 1993.

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

97

including our own prosperity, we become more selfish and less willing to spend tax dollars on programs that benefit others. Durr found that shifts in aggregate optimism and pessimism about the economy strongly predicted shifts in a measure of policy preferences that we have become familiar with – Stimson’s measure of policy mood. Like Durr, I will measure economic optimism and pessimism with the University of Michigan’s time series on consumer expectations for business conditions in the next five years.6 A fourth, and related, alternative explanation for shifts in racial policy preferences is that public willingness to accept government intervention on matters of race is a function of public willingness to accept government intervention more generally. That is, people are willing to have the government actively help black Americans to the same degree that they are willing to have government be an active player in other social concerns. In this sense, policy preferences on race are viewed as a subset of a broader set of policy concerns. Changes in racial policy preferences, then, reflect changes in policy concerns generally. We are, by now, familiar with Stimson’s measure of welfare-state policy preferences, which he calls policy mood. It represents the global predisposition of the mass public to endorse or reject government activism.7 A fifth alternative for explaining movements in racial policy preferences is the generational-replacement hypothesis, which arises from the work of Howard Schuman and his colleagues.8 To put it simply, the authors hypothesize that earlier generations of Americans were socialized into a society where bigotry and racism were expected, both from individuals and from governments. Generations socialized since the civil

6

7

8

There is one difference between Durr’s measure of expectations and mine. Durr purges his measure of economic expectations of the influence of political events (such as presidential approval and honeymoon periods), whereas I do not. This difference should have minimal effects, if any, because Durr’s aggregation interval is quarters, whereas mine is years. The universe of items used in my measure of policy mood is slightly different from Stimson’s. He includes questions about race in his measure of policy mood; I recompute his index excluding the items that have racial content, for the obvious reason that including the same indicators of racial policy preferences in two separate measures would artificially produce some association between them. In this particular case, it turns out not to affect the results (which are available upon request from the author). See Schuman et al. 1997, especially chap. 4.

0521821711c04

98

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

rights movement grew up in a country that, while still imperfect, began to include blacks as full partners in the American Dream. These effects are important to capture. In a society where open bigotry was tolerated, perhaps expected, the reasoning goes, a fair percentage of Americans opposed any government intervention on behalf of blacks merely because they disliked them. The generational-replacement hypothesis straightforwardly asserts that, as earlier generations are replaced by those less exposed to overt prejudice, a smaller share of the public will be immediately opposed to liberal racial policies. Therefore, with the slow passing of generations, the public should become slowly more liberal on race, other things being equal. I will measure these effects in each year by counting the proportion of the adult population in that year that turned eighteen during or after 1963, which was the beginning of a brief period when civil rights dominated the American agenda.9 For years before 1963, the series consists of zeros. Finally, it is possible that racial policy preferences are somehow influenced by shifts in actual racial policy – in a policy-feedback type of relationship. Perhaps as racial policy becomes more liberal, the public finds itself less enchanted with such policies; or perhaps they simply stop desiring ever more liberal policies. Conversely, as racial policy becomes more conservative, perhaps the public increasingly sees the need for liberal policies, as continuing discrimination becomes more and more evident. This model of representation has been described by political scientist Christopher Wlezien using the apt metaphor of a thermostat.10 When the public wants more “heat,” the policy “furnace,” fueled by the mechanism of representation, kicks in, and the public gets what it wants. But the very act of getting what it wants makes the public want “more heat” less and less, lest they get too hot. Although the idea of a thermostatic model of policy feedback is intuitive, it will require a brief measurement detour to operationalize and test for its effects. 9

10

It could be argued that this series should start in 1964, when the Civil Rights Act was passed. Without debating the relative merits of different measurement strategies, it should suffice to note that none of the subsequent results are at all affected by using different measures of generational replacement. See Wlezien 1995, and also Durr 1993.

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

99

Measuring Racial Policy Conceptually, what does it mean to say that the government is becoming “more liberal” or “more conservative” on racial policy? The question has no obvious answer, and there are concomitant difficulties in quantifying this somewhat elusive concept. In the abstract, the focus here is the government’s tangible commitment to fight the discrimination that is proscribed in the various pieces of civil rights legislation. If store owners, realtors, bankers, and voter registration officials want to deny equal rights to blacks, mere declarations from Washington that such activities are illegal will not completely stop the practice of discrimination. The government, if it is committed to creating a level playing field for all races, must back up the laws with enforcement. It must establish a bureaucracy to investigate charges of discrimination in its various forms and prosecute individuals and companies that violate the laws. Without enforcement efforts, the civil rights legislation would be oft ignored. There are several possible strategies to measure racial policy over time. We might adopt a strategy to measure congressional liberalism on race by examining roll-call votes over time, comparing levels of liberalism from year to year. Such a procedure would have the benefit of capturing the massive and difficult legislative efforts of 1964 and 1965. But this also highlights the fatal weakness of such a measurement procedure: how can we compare roll-call votes on race from year to year? The content of the legislation varies widely from year to year, and numerous deceptive “killer amendments” are offered at various stages in the process before a final roll-call vote is taken. From a measurement perspective, we would risk creating an invalid measure of racial policy if we relied on roll-call votes in Congress. What to do? It is often said that budgets are the most sincere expressions of governmental priorities, as the government has the opportunity to “put its money where its mouth is.” Unlike roll-call measures, budgetary measures do not suffer from a validity problem from year to year. Dollars spent in the 1960s are comparable to dollars spent in the 1990s, after controlling for inflation. In the case of race, six key agencies or divisions of departments are involved in enforcing existing civil rights legislation: the Office of Civil Rights in the Department of Education; the Office of Civil Rights in the Department of Health and Human

0521821711c04

0 521 821711

June 3, 2003

100

9:42

Mass Media and American Racial Attitudes

Services (and, in earlier years, the Office of Civil Rights in the Department of Health, Education, and Welfare); the Civil Rights Division of the Department of Justice; the Office of Federal Contract Compliance Programs (OFCCP) in the Department of Labor; the Office of Fair Housing and Equal Opportunity in Housing and Urban Development (HUD); and the independent Equal Employment Opportunity Commission (EEOC).11 Each of these agencies requires a substantial sum of money to carry out its mission. They all employ investigators to look into complaints and lawyers to pursue those complaints if warranted. It is possible, then, to measure the federal government’s commitment to the enforcement of civil rights laws relative to other governmental priorities by simply examining the size of each agency’s annual budget. From here on, the measurement is straightforward. For each fiscal year in which these agencies existed, I sum all dollars in their budgets (giving each program zero dollars for years before they existed or after they expired) and divide that subtotal by the total government budget, yielding a percentage of the total federal budget devoted to antidiscrimination efforts. By this construction, the more liberal the government is on issues of civil rights, the more money it will spend on those programs relative to others in the budget. Conversely, when the government wants to scale back efforts on the civil rights front, it will spend less on those programs relative to others. The thermostatic theory proposed by Wlezien predicts that, as government policy on race grows more liberal, the public will come to prefer a less active government on race; and, as government policy drifts a bit too far in a conservative direction, the public will come to favor a more liberal government. Therefore, we expect to find a negative relationship between our measure of racial policy and racial policy preferences.

multivariate results We are now fully equipped to present a complete model of the forces that drive racial policy preferences in both liberal and conservative directions through time. A traditional dynamic regression is presented 11

See United States Commission on Civil Rights 1982.

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

101

table 4.3. A Multivariate Model of Racial Policy Preferences Independent Variable

Effect

Dynamics

0.52c (0.10) 0.02c (0.00) −0.01a (0.00) −0.31b (0.15) −0.02 (0.02) −8.28a (4.17) −0.01 (0.02) 0.24b (0.11) 0.15c (0.04) 13.19b (5.52) 0.92 45

Egalitarian media cues Individualist media cues States’ rights media cues Volume of media coverage of race Policy feedback Economic expectations Welfare-state policy preferences Generational replacement Constant R2 N

Note: The cell entries are slope coefficients, and their standard errors are in parentheses below. a −p < 0.10. b −p < 0.05. c −p < 0.01.

in Table 4.3. The dependent variable, as before, is racial policy preferences. As in Table 4.2, a lagged dependent variable is included to control for the effects of previous lags of the independent variables that are excluded from the model. The fully specified model contains all three types of media framing, plus the total volume of coverage on race, as well as policy feedback, economic expectations, policy mood, and generational replacement as causal variables. Consistent with all previous findings, egalitarian media framing emerges as a significant predictor of racial policy preferences, and in this model, the effects are enhanced. The standardized β = 0.26, meaning a one-standarddeviation shift in egalitarian frames translates into an immediate shift

0521821711c04

102

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

of one-fourth of a standard-deviation shift in policy preferences. Again, the dynamic coefficient in the model indicates that some of this immediate impact lingers for several subsequent years, making the true effect even stronger. To translate these effects into more interpretable numbers, consider that this indicates that, between 1982 and 1991, during which time the General Social Surveys item on busing showed an increase in liberalism of fifteen points, about four of those fifteen points are due to an increase in egalitarian framing over the same time period. In this multivariate context, individualistic media framing narrowly misses the conventional 0.05 cutoff representing statistical significance, but it does cross the somewhat lower hurdle of significance at the 0.10 level. Importantly, unlike the bivariate results in Table 4.2, this time the effects are in the expected (negative) direction. That is, consistent with our theoretical expectations, we now see evidence that higher amounts of individualistic cues edge the public in a conservative direction on race. These results, because the statistical model is now fully specified, should be viewed as the best estimates that we can produce about the effects of individualistic cues, more reliable than the evidence seen in Tables 4.1 and 4.2, which in retrospect appear to be influenced by omitted-variables bias. The individualistic-cues coefficient of −0.01, when standardized, indicates that the effects of individualistic cues are less than half as large as the effects of egalitarian cues (standardized β = −0.12). It is possible that the public simply is not as sensitive to their individualistic heritage as they are to that of egalitarianism. Another possibility, though, is that the effects of individualism are somewhat contaminated by the difficulties associated with measurement. It is well known that measurement error in independent variables tends to attenuate statistical relationships, and the Newsweek measures used here, though the best available to date, surely are imperfect indicators of overall media coverage. If that is the case, these findings likely understate the size of the effects of media coverage on national opinion on race. Table 4.3 also shows the effects of the other variables in the model. In particular, there is, as expected, a negative and significant effect of coverage of states’ rights on racial policy preferences, though by comparison this effect is smaller than that for both egalitarian and individualistic cues (standardized β = −0.11). Historically, an increased focus on states’ rights lessens support for liberal racial policies. Our

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

103

final media variable, though, finds no connection between it and racial policy preferences. The total volume of coverage of race in Newsweek is unrelated to racial policy preferences; although the relationship is in the expected direction, it does not approach statistical significance. Hence we see more sensitivity to the tenor of media coverage than to its raw volume. The policy-feedback variable also emerges as a (marginally) significant predictor of racial policy preferences. Consistent with the thermostatic model, liberal shifts in policy produce disenchantment with those policies, and opinion becomes more conservative. In addition, both welfare-state policy preferences and generational replacement are significant predictors of racial policy preferences. That is, as the public becomes more accepting of government activism generally, it also becomes more accepting of government intervention to bring about racial equality. This represents testimony to the power of a “national mood,” where preferences across a wide range of issues (including race) move in tandem through time. In addition, the effects of generational replacement are evident and strong (standardized β = 0.53); as older generations that were socialized into an American society which expected discrimination against blacks become a smaller portion of the population, society becomes more liberal on racial policy. Finally, Table 4.3 shows that economic expectations are not a significant predictor of racial policy preferences, and what effect can be found runs in the wrong direction. Compared to public opinion on the welfare state, preferences for racial policy seem to be much less affected by forecasts about the health of the nation’s economy.

implications Clearly, several forces shape racial policy preferences and have contributed to their evolution over time. Unsurprisingly, generational replacement has been powerfully important in the liberalization of American opinion on race. But the media have also played a role, as have opinion on the size of the welfare state and racial policy itself. Although it might be tempting to reduce the main finding of this chapter to a statement of the type “The media do influence racial policy preferences,” a more complete rendition of the findings here would recognize that the nature of media influence is extremely subtle. The national

0521821711c04

104

0 521 821711

June 3, 2003

9:42

Mass Media and American Racial Attitudes

media do not make blanket statements about race, such as “Blacks are good” or “Blacks are bad,” which the public blindly accepts as truth and incorporates into its policy preferences. This is in contrast to other areas of coverage such as the economy, where media messages are far more direct (“The economy continues to show signs of sluggishness”). In those cases, it is easy to portray media influence in a simple way: the media say something; the people believe it. Race, as usual, presents a more complex scenario. The types of messages that influence public opinion are those that resonate with values to which most Americans subscribe. The question, then, becomes one of emphasis, whether the media tend to emphasize one core value or another. The choices are subtle, but their impact is substantial, at least with respect to public opinion. Perhaps the most significant contribution of this analysis is the effort to explain both liberal and conservative shifts in racial policy preferences. Instead of limiting the focus to why Americans have become more liberal (or more conservative) on race, I have provided a framework that can explain movements in either direction. And this, in my view, is the key contribution to a systematic understanding of politics. Instead of trying to explain every bump and wiggle in a half-centurylong time series, we have sought, and found, several forces that can help to explain the waves of both liberalism and conservatism that seem to ebb and flow with the historical tides of American political history. As new issues arise in racial politics, they present opportunities to define and redefine the debate afresh, for leaders to “frame” the debate in the way they like. A recent example is the spate of revelations that many law-enforcement agencies have been engaging in what has come to be called “racial profiling.”12 What does this suggest to the average American media consumer? I expect that such stories are rife with egalitarian resonances, because they show how members of racial minorities – especially blacks, in many cases – are treated differently from whites. When the police target a black person for a stop-and-frisk when they would not have stopped a white person, it shows patently unequal, nonegalitarian treatment. Moreover, unlike the case of, say, affirmative action, there is no readily available competing interpretation. Nor are 12

This example will have to remain purely hypothetical, as the issue did not arise until the later part of the 1990s, after my media time series ends.

0521821711c04

0 521 821711

June 3, 2003

9:42

Media Framing and Racial Policy Preferences

105

there prominent conservatives eager to defend the practice and offer such an interpretation. This leaves the egalitarian reference essentially unchallenged and, I suspect, leads to a proliferation of egalitarian cues any time that racial profiling is covered. All of this will be ingested and filed away in the minds of Americans, and will represent a portion of the cues they recall when prompted to articulate an attitude about racial policy.

0521821711c05

0 521 821711

June 3, 2003

9:46

5 The Fusion of Race and the Welfare State in the Public Mind

The fact is that the black middle class of 1974, like that of earlier years, is a minority within the black community. In 1974, as in 1964, 1954 and in the decades stretching into the distant past, the social and economic reality of the majority of black people has been poverty and marginal status in the wings of our society. – Newsweek, July 8, 1974

Observers of American politics speak in terms of eras. Perhaps this intellectual habit comes from our collective fascination with realignment theories, which help us to organize American history. We think of American political history in chunks, like the Jacksonian era or the Progressive era, reducing the essential political divisions of the moment to bite-sized pieces. The two most recent overarching issues that have shaped all of American politics are the New Deal and race. The New Deal era – or, as I will call it, the welfare state – concerns the proper role of government involvement in securing the well-being of a society. And the issue of race revolves around the treatment of blacks in American society. Although such shorthand is surely convenient for arranging our thoughts about politics, it is surprising how little we actually know about the relationship between the welfare state and race, how that relationship has evolved, and the causes and consequences of this evolution. We do have some sense that, at the mass level, the issue of race 106

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

107

and the cluster of welfare-state issues used to be separate.1 But we also have the sense that they are not separate today, and in Chapter 3, we have seen some evidence of this fact. We remember, for example, that the New Deal could not have been passed by Congress without support from southern senators and representatives, many of whom were chairs of powerful committees. Surely, at that time, if race and the welfare state had been seen as synonymous by the mass public, legislators from the South would not have supported Franklin Roosevelt’s New Deal legislative agenda. But they were not considered the same thing back in the 1930s, and the New Deal ultimately won passage. Today, on the other hand, we see race and the welfare state as close to inseparable. Indeed, the bulk of our welfare-state terminology has been tainted with the issue of race. The notion of “poverty” itself has somehow shifted in the public eye from the condition of rural, poor, uneducated white people to that of urban blacks. The individual image of “welfare” has become “the welfare queen.” It almost goes without saying that the welfare queen is envisioned as a young black woman, not a young white (or Hispanic or Asian) woman. In today’s political arena, it is patently obvious to state that government assistance to the less privileged in America is thoroughly imbued with the politics of race. And politicians in recent decades have used the veiled politics of race to create opposition to welfare-state policies.2 Figure 5.1 reminds us about the over-time relationship between attitudes toward race and the welfare state. The solid line represents the public’s liberalism on matters of race, and the dotted line represents liberalism on the cluster of welfare-state issues. As we discussed in Chapter 3, it is easy to see how the two series appear virtually uncorrelated before the middle portion of the 1960s, and how they seem virtually indistinguishable since that point. The idea that race and the welfare state have somehow merged is not particularly novel. However, we know little about how or when this state of affairs came into being. At one point in time, they were separate issues; now they seem inextricably related. How did this happen? What

1

2

I emphasize “at the mass level,” because there is reason to believe that at the elite/policy level, the politics of race had profound effects on the actual New Deal policies. See Quadagno 1994 for a detailed treatment of this idea. See Huckfeldt and Kohfeld 1989.

0521821711c05

108

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

figure 5.1. Racial and Welfare-State Policy Preferences, 1950–1996

was the causal mechanism involved? What led to the fusion of race and the welfare state? The aim of this chapter is to answer these questions. Consistent with the major theme of this book, the answer suggested is that race and the welfare state became fused because of the ways in which the mainstream media framed their coverage of race, and how that coverage changed over time. Specifically, in the 1950s and early 1960s, the picture that Americans had of blacks was not dominated by references to poverty or dependence on welfare. But in the middle to late 1960s, several major shifts in the tone of media coverage of race emphasized that blacks were disproportionately poor and disproportionately dependent on government assistance. That is how race and the welfare state became one issue.

a literature in confusion There has been much academic speculation about the relationship between race and the welfare state, and most of it emerged from work on the shifts in party systems and realignments. In particular, Edward Carmines and James Stimson use time-series data to argue that, during the 1960s, race eclipsed the welfare state in terms of its centrality in organizing the party system. That is to say, before the 1960s, disagreements between Republicans and Democrats were effectively over the cluster of New Deal issues; after the mid-1960s, those disagreements revolved around the question of the role of blacks in American society,

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

109

and what (if anything) the federal government ought to do about it. Their theory of “issue evolutions,” which in many ways is the reigning conventional wisdom, posits that this change occurred first at the level of candidates for office, then at the level of party activists, and finally at the mass level.3 Some scholars – most prominently the political scientist Alan Abramowitz – have criticized Carmines and Stimson’s conclusion that the party system revolves around racial issues today. His mode of analysis tends to differ from the over-time analyses of Carmines and Stimson, focusing instead on cross-sectional surveys constructed to carefully document the causes of partisan attitudes at a single point in time. In particular, using data from the 1988 National Election Study, Abramowitz conducts multiple-regression analyses which, he claims, support the notion that welfare-state policy preferences trump racial policy preferences in terms of predicting partisanship. Abramowitz shows that once welfare-state policy preferences and certain demographics are controlled for, racial liberalism or conservatism is not a statistically significant predictor of partisanship.4 Partisanship, in his view, is still overwhelmingly a function of an individual’s attitudes about the welfare state. This is seemingly straightforward evidence that racial issues did not eclipse the welfare state in the universe of American politics. For if, as late as 1988, a person’s choice about whether to identify as a Democrat or a Republican is a function of his or her attitudes about the welfare state rather than race, how can race be the central organizing force behind contemporary American politics? It is worth noting that both of these scholarly arguments are somewhat at odds with the popular wisdom that racial issues and welfarestate issues have become inseparable. Carmines and Stimson claim that race has replaced the New Deal as the nexus of mass attitudes; Abramowitz says it has not. Neither of these opinions squares with our intuitive sense that the two issues became one. And it is also worth noting that, although their work is somewhat related to the question at hand, neither the work of Carmines and Stimson nor that of Abramowitz directly addresses the question that I pose. They are primarily concerned about the focal point of the party system, whereas my 3 4

See Carmines and Stimson 1980, 1989. See Abramowitz 1994, esp. tables 8 and 9.

0521821711c05

110

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

question revolves around how race and the welfare state are perceived in the mind of the public. Obviously there is some overlap between the questions, but that overlap is not complete. Public perceptions about racial policy and the welfare-state cluster of issues can exist outside of consideration of where Republicans and Democrats stand on those issues. There is a possibility that neither set of researchers has considered. The current debate seems to be obsessed with which issue domain determines the shape of the party system, and there is an underlying assumption that either race or the welfare state is dominant. Why must this be so? Rather than conceiving of the situation in such fashion, it is possible to speculate that race and the welfare state, at some point in time, became melded together, and that they now represent one set of issues rather than two. This possibility has not been entertained either by Carmines and Stimson or by Abramowitz. Moreover, if race and the welfare state became one and the same thing in the mind of the public, then analyses like Abramowitz’s, which purport to sort out the effects of race over and against the welfare state, would be misleading, both substantively and methodologically.5 Recently, however, some individual-level research has emerged that explores this relationship between racial attitudes and welfare-state policy preferences. Instead of assuming that racial attitudes and preferences about the welfare state compete with one another, this research examines the possibility that these sets of attitudes are mutually reinforcing. In particular, political scientist Martin Gilens has investigated the extent to which whites’ attitudes about welfare are a function of stereotypes about blacks.6 Controlling for other possible influences, such as core values and self-interest, he finds that racial stereotyping is the single best predictor of preferences about welfare policy. That is, attitudes about race and the welfare state are inextricably entwined. This highlights the just-mentioned problems with Abramowitz’s 5

6

One statistical implication of this is that including both sets of issues in a multipleregression setup would lead to high multicollinearity (because both independent variables represent the same underlying set of attitudes) and hence inflated standard errors and lower t-ratios. More important, such an analysis would also be plagued with endogeneity problems. Abramowitz’s analysis assumes that racial attitudes and welfarestate policy preferences are both exogenous (to partisanship), when in fact attitudes about race and the welfare state might be mutually endogenous. See Gilens 1995, 1996, and 1999; also Kluegel and Smith 1986.

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

111

analysis, and how sorting out the effects of racial and welfare-state attitudes is far from a simple matter. Gilens clearly shows that assuming that welfare-state issues are independent of attitudes about race is wrong. Attitudes about race and the welfare state are very closely related – more closely related than has been suspected.7 This research is useful, for it recognizes the interconnectedness of Americans’ policy preferences rather than regarding beliefs about race and beliefs about the welfare state as separate and independent. However, it does not answer all the questions about the relationship. Although it details the relationship between racial stereotyping and welfare-state attitudes, Gilens’s research does not consider the role played by policy preferences about race – a critical difference upon which I will elaborate. Moreover, the surveys that Gilens examines are, by the standards of research in race politics, relatively recent ones, conducted in 1986 and 1994. We still do not understand the process, or the timing, of how these sets of attitudes came to be associated with one another. What is needed is an examination of this relationship over time. I now turn to that task.

race and the welfare state in the public mind Why did race and the welfare state become related? The answer that I propose involves a theory of political learning as well as a theory of political attitudes. In effect, the main question of interest can be answered by first answering two other questions: how do people learn about what blacks are like; and how do people learn about what poor people are like? People do not normally experience national politics firsthand. That is, the normal routine of most people’s lives does not bring them into contact with salient political issues on a regular basis. Most people rarely, if ever, go to protests, rallies, or campaign events. Most people do not directly experience the effect of newsworthy public policies (except, perhaps, for taxation) on a daily basis. This is particularly true in the case of racial policy today, where, for example, most parents do not have children who are bused to another district to achieve integration, and most people are not denied or granted jobs because of 7

See Gilens 1995, esp. fig. 1 and Gilens 1996, esp. table 1.

0521821711c05

112

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

affirmative-action requirements. This was perhaps also true even when racial policy took center stage in everyday news during the 1960s – or, at least, it was true of most northerners at the time: civil rights demonstrations and protests were overwhelmingly viewed as a southern problem. Because most people do not experience politics firsthand, they do not learn about politics primarily from their own experiences but rather through those of other people. These experiences can be communicated in a variety of ways: through conversations with friends, family members, fellow worshipers at religious services, and colleagues at work, or, perhaps more often, through the mass media. Because American residential communities are relatively homogeneous, especially economically, people who are wealthy do not normally learn about poor people – who they are, what they are like – through daily, sustained interactions. More probably they learn about poverty through the media. This is especially likely to be true when the subject is how the composition of the poor has (or has not) changed over time. This is, in effect, a small-sample problem: even though many people may have a few acquaintances who are impoverished, every individual’s sample is too small to give an accurate portrayal of the overall makeup of the poor, and how the distribution of poverty has changed over time. And yet people have real and politically meaningful ideas about which groups are disproportionately impoverished in American society. Where do these ideas come from? They come from the national media. The same is likely to be true of learning about groups in society with which a person has little personal contact. How do most whites learn about blacks, Asians, or Hispanics? As our society becomes more diverse and less segregated than it has been historically, individuals will clearly learn more about other societal groups through firsthand contact – on the job, in the neighborhood, and in recreational activities. However, even today it is surely the case that a large portion of such information comes from the media. In decades gone by, this was probably even more true than it is today. As individuals learn about other groups at least in part through the media, they will develop a reservoir of associations with those group. Stories communicate messages about the character of their subjects, and the recipients of those messages fall prey to the socialpsychological phenomenon of generalizing these characteristics to the

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

113

relevant groups.8 For example, when a story appears in the news about a particular black person – perhaps one who happens to be impoverished – an individual who is exposed to that story might be tempted to think that the characteristics that apply to that person also apply to blacks in general. As these stories accumulate in the memories of media consumers, an aggregate, collective picture of different groups in society emerges. Previous images are reinforced by multiple stories with consistent themes, thus making the process of generalization and stereotyping more likely. These collective pictures are surely complex and multifaceted. Multiple images probably arise in the aggregate because of the diversity of separate stories relevant to any group. And these multiple images can be thought of as “competing considerations.” Associations develop between characteristics and groups. Sometimes these considerations may conflict with one another; at other times they may complement or reinforce one another. But considerations need not stay in our minds forever. The effect of any particular story is not likely to be a permanent association between a group and a characteristic. Memories fade. Production values and story lines change. New considerations replace old, outdated ones as media consumers encounter new information. And this is the key point for understanding the relationship between race and the welfare state. Over the last several decades, the images of blacks that Americans have received through the mass media may have changed substantially. As a result of these changes, the public’s image of blacks as a group has been altered, and this has led to a fusion of race and the welfare state in the minds of the American public. Specifically, I am hypothesizing that the portrayal of blacks in the American media became increasingly infused with the theme of blacks as impoverished and dependent on government assistance. Such coverage, as we saw in Chapter 2, was not always prominent; recall that, in fact, it was almost nonexistent as late as 1960. But as this theme emerged in the media and came to be common in the media’s coverage of blacks, it had a substantial effect: what Americans had once thought of as two separate problems – race and poverty – came increasingly to be viewed as one. Media coverage that portrayed blacks 8

See Iyengar and Kinder 1987.

0521821711c05

114

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

as poverty-stricken caused the fusion of race and the welfare state. In other words, in earlier eras, when race and poverty were presented as separate problems, American beliefs about these two issues were more or less independent of one another. In more recent years, as the two problems have been linked in stories in the mass media, Americans as a consequence have come to think of race and poverty as one problem, not two. The dynamic connection I am positing between policy attitudes on race and the welfare state, though, differs in some subtle yet important ways from the cross-sectional portrayal by Gilens. In his framework, welfare attitudes are a function of prejudicial stereotyping of black Americans, not of policy attitudes toward race. The question at issue here is the over-time dynamics of policy attitudes about race, not racial prejudice.9

timing: when two issues became one Although the recent cross-sectional research by scholars like Gilens has brought the relationship between race and the welfare state into focus, it has not yet been demonstrated how that relationship has evolved over time.10 How can we subject our dynamic theory to empirical scrutiny? We need, of course, measures of the key concepts involved: media framing of black poverty and the merging of racial and welfarestate policy preferences. In Chapter 2, we developed precisely this type of media measure using the Newsweek data set. Because we have all the stories printed on black Americans, we can easily track references to black poverty.11 9

10

11

Because over-time measures of bigotry have consistently shown a steady decline in the portion of the American public that harbors prejudice, it seems an unlikely causal agent in explaining back-and-forth movement in policy attitudes. Moreover, Gilens is positing a causal connection between the two sets of attitudes, whereas I am intent on showing how the two streams merged to form a single river without attributing causal priority to one or the other; instead, I explore the causal primacy of cues from the national media. The model posited and tested here, for these reasons, is not a simple extension of Gilens’s model into the domain of time. Unlike in previous chapters, all analyses in this chapter will be conducted on data beginning in 1958 instead of 1950. This is because the equality measure used below did not exist before 1958, which limits the period of analysis here. Surely, to make the test more exacting, it would be desirable to have all stories about poverty, and to inquire whether, over time, stories about poverty increasingly came

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

115

Tracking the Merger But how can we develop a measure of the relationship between welfarestate and racial policy preferences? This analysis is not perfectly straightforward because it involves an unusual type of dependent variable. The dependent variable here is not a simple time series of policy preferences of one type or another, but rather the similarities (or differences) between two separate types of policy preferences. The first step toward a solution is to realize that the dependent variable of interest is some measure of distance, of disequilibrium, between racial policy preferences and welfare-state preferences. This can be accomplished in several ways, but because the series are roughly comparable in variances, the simplest solution is to subtract one from the other. In such a case, larger absolute values represent greater disequilibrium, and smaller ones indicate greater synchronization, with a value of zero indicating perfect equilibrium.12 Figure 5.2 presents our measure of reductions in disequilibrium between attitudes toward race and the welfare state. The solid line in the figure represents the gap between our two measures of policy preferences. Note first the presence of the zero point on the left-handed y-axis. Next, see how the solid line begins substantially away from that zero line; in the late 1950s and early 1960s, the two opinion measures were out of equilibrium with one another. That changed in the middle portion of the 1960s, and the gap between race and the welfare state closed steadily and rapidly, reaching near-perfect equilibrium – indicated in the figure by the solid line hovering around the zero line – in the very early 1970s.

12

to be dominated by references to blacks. This complementary data set would help clinch the case about the nature of the sea change in media coverage; it would show that, not only did stories about blacks change to increasingly emphasize poverty, but stories about poverty also changed to increasingly emphasize black Americans. You will recall, however, that our population of stories from Newsweek does not contain all stories about poverty, but rather all stories about blacks, making such idealized comparisons impossible. Nor would such a data set be easily created, for “poverty,” unlike race, is not something that is indexed in Newsweek. Because it is intuitively desirable to have a value of zero represent perfect equilibrium, it was necessary to recalibrate the means of the series so that the subtraction would produce values close to zero for the most recent years of the resulting series. This transformation, of course, is perfectly innocuous from a statistical standpoint, and has no bearing at all on the results presented below. It is done for graphical purposes only, and the details are available from the author.

0521821711c05

0 521 821711

116

June 3, 2003

9:46

Mass Media and American Racial Attitudes

figure 5.2. Merged Public Opinion and Media Cues on “Black Poverty,” 1958–1994

The figure shows precisely what one would expect, given the results of Figure 5.1. There, the two public-opinion time series seemed to be independent of one another before the mid-1960s, and perfectly synchronous thereafter. In Figure 5.2, we now have a single time series that represents this very phenomenon. This series, which represents the gap between racial and welfare-state policy preferences, becomes our key dependent variable in the analyses below. The Newsweek series representing the amount of coverage of black poverty, first shown in Chapter 2, also appears in Figure 5.2. The visual evidence is striking and suggestive. Consistent with our theory, the rapid growth in black poverty cues coincides with the movement toward equilibrium in our public-opinion time series. Increasing portrayals of black poverty seem to lead to a closing of the gap between race and the welfare state in the collective mind of the American public. Modeling the Merger Thus far, the evidence about the relationship between media portrayals of poverty and the dynamics of public opinion on race and the welfare state has all been visual, and sometimes our eyes – with the help of graphics packages – can deceive us. Making statistical connections will be more convincing. We begin that process in Table 5.1, in which

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

117

table 5.1. The Media and the Relationship between Welfare-State and Racial Policy Preferences Independent Variable

Effect

Dynamics

0.34a (0.16) 0.04a (0.02) −0.08 (0.36) 0.30 37

“Black poverty” media frames Constant R2 N

Note: The cell entries are slope coefficients, and their standard errors are in parentheses below. a −p < 0.05.

year-to-year shifts in the gap between racial and welfare-state preferences are modeled as a function of the yearly amount of media references to black poverty. As in all other analyses in this book, a dynamic term, which is a lagged value of the dependent variable, is included to control for previous lags of the causal variables, which presumably have lingering effects in subsequent years.13 In the table, we see that our eyes did not, in fact, deceive us; an increase in the number of messages sent by the national media connecting race and poverty leads the public to link race and the welfare state more closely in their minds. The coefficient for the media variable of 0.04 translates into a standardized coefficient of 0.30 – meaning that a one-standard-deviation shift in media frames produces an immediate change in dependent variable of roughly one-third of a standard deviation. These sizable immediate effects are, as they have been throughout this book, magnified by the dynamics that carry these effects forward for several time periods. This initial evidence, at least, suggests rather strongly that the American public increasingly 13

It is worth noting that the actual dependent variable here is not the level of disequilibrium between the two opinion time series, but the changes in that disequilibrium. It is a more theoretically appropriate test, as the theory presented here speaks to the ways in which the media produced moves toward equilibrium. One side effect of this is that the goodness-of-fit measures in the table, the R2 , are substantially smaller than in previous chapters. This is the typical and expected result of using a dependent variable in first differences instead of levels.

0521821711c05

118

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

connected race and the welfare state because the mass media increasingly made the connection for them. The startling evidence that we discovered in Chapter 2 about the explosion of references to black poverty in the mid-to-late 1960s – jumping from only 10 in 1960 to 680 merely seven years later – had political consequences. It changed the way people thought about race, and also the way they thought about poverty. But, as we did in the previous chapter, again we must grapple with the complicating issue of alternative explanations for the dynamics of public opinion, lest we misattribute to the news media some causal power that actually belongs elsewhere. That is, public opinion may have been responding to other forces in society that caused people to connect race and the welfare state in their minds. Two such forces come to mind as possibilities, each of which will be dealt with in the next two sections.

perception versus reality The above theory, for which we have now found empirical support, postulates that public opinion on race and the welfare state merged in the middle and late 1960s as a result of a dramatic increase in the number of messages the mass media conveyed during that period of time explicitly linking the two issues. That is, people came to think of race and the welfare state as essentially one issue because they were told that the two problems were, for all intents and purposes, one problem. Since then, their over-time dynamics have become almost perfectly synchronous. But perhaps it is possible that the American public is not responding to media messages about poverty – to their perceived reality – but to reality itself. Perhaps public opinion actually responds to shifting conditions in society – to changes in socioeconomic realities – instead of media coverage of black poverty. That is, maybe the real reason that attitudes about race and the welfare state have merged has less to do with what people are told in the mass media and more to do with the actual levels of equality between black and white Americans. As poverty becomes more of a problem that disproportionately affects black Americans, so the reasoning goes, people will be more likely to see the issue of government assistance to poor people and government

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

119

'Black Poverty' Cues in the Media

Merger of Attitudes on Race and the Welfare State

Economic Racial Equality

figure 5.3. Economic Equality and Public Opinion: An Alternative Causal Model

assistance to black people as one and the same thing. If this were the case, the media effects reported above would represent intermediary links in a longer chain of causality. This alternative scenario is represented in Figure 5.3. The lighter, dotted line represents the theory of media influence as outlined earlier in this chapter, and the heavier, darker arrows represent the alternative causal scenario. Therefore, in this different conceptualization, both the evolution of public opinion and media coverage of black poverty are portrayed as a function of actual economic equality between blacks and whites. This theory is the most straightforward and benign portrayal of media influence possible. It says, in effect, that although the media presentations of race and poverty fluctuate, they do so because of realworld shifts, not for some other reason. Media coverage is epiphenomenal. That is, if there were underlying shifts in the economic makeup of society, the media would simply report that change. The mainstream press would respond to a change in the economic structure of society, report it, and public opinion would adjust accordingly. The media would be reflecting a reality, not creating one. The press would be reporting what it saw, holding up, as it were, a mirror to the face of American society and merely showing us who and what we are. Importantly, too, the story presented in the figure implies that public opinion responds to this reality, and not primarily to what is portrayed in the media. If this were true, what would the observable implications be? If opinion were responding to the reality of racial inequality instead

0521821711c05

120

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

of to media reports of it, then the statistical linkage between media framing and the gap between welfare-state and racial policy preferences that we saw in Table 5.1 would vanish once we introduce a statistical control for the dynamics of equality. Before we turn to this analysis, however, we need a measure of economic racial equality. I now turn briefly to a strategy to measure it. Measuring Economic Equality What does it mean to say that black and white Americans are, in an economic sense, more or less equal? It is important to state that we are conceiving equality as something that is inseparable from the concept of time, something that must be seen in the context of time to gain a proper perspective. Surely the concept of equality has many significant meanings aside from the economic – equality before the law and equality of opportunity being the most obvious – but here we are focusing on the economic aspects of equality because that conceptualization is most relevant to the theory at hand, which involves race and poverty. In earlier work, I have argued that equality in contemporary American society has several components.14 Income is the most obvious indicator of economic equality and is something of a gateway to other forms of economic equality. Were we forced to pick a single indicator of economic equality, income would be it. But we are not forced to choose just one. Economic equality has other facets. Particularly as health-care costs skyrocket, access to quality health care is a component of equality. So is the ability to own a home as compared to renting, which provides means to accumulate wealth over the long term. For the years 1958 through 1994, I have compiled data for five indicators of economic equality between blacks and whites that correspond to these aspects of economic equality – income, health, and housing. They are: the black–white family-income ratio; the black-white infantmortality ratio; the black–white maternal-mortality ratio; the ratio of physician visits per year between blacks and whites; and the ratio of home ownership between blacks and whites. All of these can be considered imperfect but nonetheless useful indicators of economic equality. 14

See Kellstedt, McAvoy, and Stimson 1996, esp. pp. 136–41.

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

121

table 5.2. Correlations of Indicators of Racial Equality with Overall Index

Item

Number of Observations

Correlation with Index

Black-white income ratio Infant mortality ratio Maternal mortality ratio Physician visits ratio Home ownership ratio

34 37 37 25 20

0.82 0.69 0.53 0.83 0.72

Source: Updated by the author from Kellstedt, McAvoy, and Stimson 1996.

And most of them – the infant- and maternal-mortality indicators being the exceptions – have years where the government did not report statistics for them. But, even with the problem of missing observations, we can combine these five indicators into a single index using the technology described in Chapter 3 to create our index of racial policy preferences. The correlations of each of these indicators with the resulting equality index can be seen in Table 5.2. The table reveals that each of the indicators serves as a useful measure of economic equality, with the income ratio and physician-visits ratio – correlated with the index at 0.82 and 0.83, respectively – most strongly related to it. But each of the five indicators provides useful information to the index, contributing to a reduction in measurement error for the equality concept as a whole. The index is displayed in Figure 5.4, where higher numbers represent greater equality between blacks and whites, and lower numbers indicate more inequality. Between 1958 and the middle of the 1960s, the series is relatively flat, but around 1965 it witnesses a sharp, sustained upturn toward greater equality. That movement lasts until the early 1970s, where another plateau is reached. Toward the end of the series, coinciding with the recession in the George H. W. Bush administration, equality dips a bit, and seems to recover at the end of the series. As noted, if the causal model of Figure 5.3 is right, then, once we control our new measures of economic equality, the connection between the media and the dynamics of opinion will largely disappear. To test for this possibility, Table 5.3 presents models of the dynamics of opinion much like the one in Table 5.1, now adding controls for

0521821711c05

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

122

table 5.3. The Media and the Relationship between Welfare-State and Racial Policy Preferences: Controlling for Economic Equality Equation Independent Variable

(1)

(2) b

Dynamics

0.30a (0.16) 0.04b (0.02) –

0.33 (0.16) 0.04b (0.02) –10.14a (5.69) –

“Black poverty” media frames Black-white income ratio Black-white equality index

–12.01b (5.54) 7.17b (3.37) 0.39 37

5.98a (3.42) 0.36 37

Constant R2 N

Note: The cell entries are slope coefficients, and their standard errors are in parentheses below. a −p < 0.10. b −p < 0.05. 0.68 0.66

Economic Racial Equality

0.64 0.62 0.60 0.58 0.56 0.54 0.52

94

92

19

90

19

88

19

86

19

84

19

82

19

80

19

78

19

76

19

74

19

72

19

70

19

68

19

66

19

64

19

62

19

60

19

19

19

58

0.50

Year

figure 5.4. Economic Racial Equality, 1958–1994

economic equality. In column 1 of the table, we use the simple income ratio as a baseline indicator of equality. The results there show that adding this statistical control does not reduce the impact of the media on public opinion at all. The coefficient of 0.04, in fact, is exactly the

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

123

size of its counterpart in Table 5.1. The black-white income ratio, too, approaches statistical significance, but critically, the effects are in the opposite direction than was hypothesized. That is, the results in column 1 suggest that the public came to associate race and welfare most closely as blacks became more equal in economic terms. Perhaps this anomolous finding is a function of using the rather crude income-ratio variable as our measure of economic equality; if so, that would help to explain the inexplicably wrong-signed coefficient for the income-ratio variable. So, in column 2 of Table 5.3, we replace the income-ratio variable with the broader index of economic racial equality described above. The results, though, are nearly identical to those in column 1. In fact, the size of the media effect on the evolution of opinion on race and the welfare state remains unchanged at a statistically significant 0.04. As for the equality index, again the effect is in the opposite direction than was hypothesized, but this time it crosses the hurdle of statistical significance. What should we make of this? Explaining this perverse result requires either some mental gymnastics of almost Olympic caliber, or it requires us to discard the notion that the public responds to economic reality. More important, though, is the finding that public opinion responds over time, not to the evolving social reality, but to portrayals of that reality in the national media. The suggested revisions to our theory as presented in the alternative causal model of Figure 5.3 cannot be sustained.

closing the policy loop Still another alternative theory exists to explain the merger between racial and welfare-state attitudes. Perhaps, instead of being sensitive to shifts in the media, public opinion on race and the welfare state is responding to shifts in government policy on race. Public opinion on race and the welfare state might have merged because the federal government took a more active role in securing the rights of black Americans. As the government took steps toward protecting, and then enforcing, the rights of its black citizens in the marketplace, this may have suggested to the American public that blacks, because of discrimination, were disadvantaged in economic competition. In protecting and fighting on behalf of the victims of discrimination, the government may have produced the unintended side effect of reminding the public that

0521821711c05

0 521 821711

124

June 3, 2003

9:46

Mass Media and American Racial Attitudes

'Black Poverty' Cues in the Media

Merger of Attitudes on Race and the Welfare State

Racial Policy Liberalism

figure 5.5. Policy Liberalism and Public Opinion: An Alternative Causal Model

blacks were poor relative to whites, which in turn led policy preferences on race and the welfare state to merge. Moreover, shifts in government antidiscrimination policy may drive media portrayals of black poverty. As the press responds to the implementation of such policies, they report what they see the government doing, and thus government efforts to intervene in the marketplace to protect the rights of blacks. This reporting naturally contains signals about the economic position of the beneficiaries of liberal government policies. These suggested causal scenarios are presented in Figure 5.5. There, the connection between media coverage and public opinion is again represented by a dotted line, and the effects of government policies on the media, and directly on opinion, are presented in bold, solid lines. Again, if this were true, then adding a variable for government antidiscrimination policy to the equations estimated in Table 5.1 would have the effect of dampening the effect of the media on public opinion. This possibility is investigated in Table 5.4, where we add the government antidiscrimination policy variable, developed in Chapter 4, to the regression equations from earlier. There, in the face of another challenge, we find the media-cues variable remaining strong, positive, and statistically significant. The coefficient remains perfectly unperturbed at 0.04, as it was in all previous analyses. Here, though, we see some evidence that opinion is indeed responding to shifts in government policy – and this time, the effects are in the expected direction. The more liberal federal antidiscrimination policy became over the years,

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

125

table 5.4. The Media and the Relationship between Welfare-State and Racial Policy Preferences: Controlling for Policy Liberalism Independent Variable

Effect

Dynamics

0.28a (0.16) 0.04b (0.02) –3.06a (1.80) 0.58 (0.52) 0.36 37

“Black poverty” media frames Racial policy liberalism Constant R2 N

Note: The cell entries are slope coefficients, and their standard errors are in parentheses below. a −p < 0.10. b −p < 0.05.

the more similar public opinion became on race and the welfare state. The coefficient of –3.06 is not statistically significant at conventional levels, but it is at the looser 0.10 standard, and its effects are probably real. Upon glancing at the respective sizes of the coefficients, one is immediately tempted to think that the effects of the policy variable are substantially larger than the media variable. In fact, the opposite is true. The discrepancy exists because of the differences in the scales of the two independent variables. When comparing their standardized effects, which filter out the differences in their metrics, we see that the media variable exerts noticeably more causal influence than the policy variable. The standardized coefficient for the media variable is again 0.30, and the government policy variable registers a standardized effect of –0.22. Clearly, though its effect is not as large as that of the media, we do see a noticeable impact of government policy on public opinion. We have seen two challenges to the theory of media influence proposed in this chapter. The first, that opinion responds to economic reality rather than media cues, found no support. The second, that opinion responds to government policy liberalism, turns out to be something of a complementary, not a contradictory, explanation. The merging of

0521821711c05

0 521 821711

126

June 3, 2003

9:46

Mass Media and American Racial Attitudes

attitudes toward race and the welfare state, as we have seen, was primarily a function of a dramatic increase in the number of messages the national media communicated connecting race and poverty. It was also, to a lesser extent, a function of the government’s own antidiscrimination efforts. The enactment of, and then the increased enforcement of, laws that forbade marketplace discrimination against black Americans had some role in connecting race and poverty in the public’s mind. It is worth noting that the two challenges to our theory of media influence are actually quite closely related on an empirical level. Our variables representing the government’s antidiscrimination efforts and economic racial equality are correlated at a striking r = 0.84. Surely, this clears up the perverse findings from Table 5.3 that seemed to show that increases in equality were associated with the merging of attitudes about race and the welfare state. The results from Table 5.4, combined with the close correlation between policy and equality, suggest that the results from Table 5.3 were spurious: those misleading effects were merely discovered because we had not yet controlled for the effects of the closely related policy variable. The more theoretically sensible results from Table 5.4 help to resolve the anomalous findings. Although it is really the subject for another book, this finding about the relationship between the government’s antidiscrimination policy efforts and economic equality between blacks and whites deserves some attention. Amid all the furor about the unintended political consequences of policies like affirmative action, or the concerns about creating immense government bureaucracies to end poverty, we have here some evidence that the less controversial efforts of the government to fight discrimination produce movement toward economic equality. To be sure, this simple two-variable correlation does not control for the important effects of economic cycles, or for other policy instruments such as affirmative action or a broader antipoverty effort, and therefore the finding should be regarded as tentative. But a tentative finding is a finding nonetheless, and should guide future research to investigate more fully the dynamics of economic equality.

conclusions A few clear conclusions can be drawn from this analysis. The most significant is that the relationship between attitudes about race and

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

127

attitudes about the welfare state is more subtle – and, in my view, more interesting – than had previously been thought. The evidence presented here points to the fact that racial policy preferences and welfare-state policy preferences were once unrelated at the aggregate level, but that within a short period of time their over-time dynamics became virtually indistinguishable, and have remained so ever since. The mass media were the major force that led to this merger. But this raises new questions. The epigraph that introduces this chapter, taken from an issue of Newsweek in 1974, has a particular bite to it, a bite that contains more than a hint of self-criticism. The writer tries to paint a picture of where blacks stand economically in comparison to whites and notes that, despite some recent gains, “the social and economic reality of the majority of black people has been poverty.” The quotation is not poignant because it is a profound revelation; it is poignant because it was written in 1974, and yet the author notes that these realities were present “in 1964, 1954 and in the decades stretching into the distant past.” Despite the fact that this social situation existed for decades before the article appeared, the article was not written in 1964, 1954, or in any of the decades past. In 1954, there were exactly zero mentions of black poverty in all of Newsweek – this in a year not lacking for news on the civil rights front. A decade later, when the civil rights movement was in full swing, the mainstream press was beginning to report this reality, and report it loudly. We have seen that this enormous shift in media content had political consequences – in this case, it caused two separate streams of public opinion to merge into a single river. But that leaves open the question of where this surge in media portrayals of black poverty originated. Why did the mainstream press ignore this reality for so long, then cover it with a sudden vengeance? We have seen that it cannot be the case that shifts in economic standing caused the media increasingly to portray blacks as poor. The simple “mirror of reality” theory has to be shattered. In fact, coverage of black poverty was increasing exponentially at the precise time when blacks were making their most substantial economic progress. This raises the questions: If the economic reality of the dynamics of poverty did not influence the dynamics of media coverage, then what did? Why did the media more frequently mention black poverty when blacks were becoming less poor? The analyses in Table 5.4 (and the

0521821711c05

128

0 521 821711

June 3, 2003

9:46

Mass Media and American Racial Attitudes

associated causal model in Figure 5.5) begin to point to an answer: media coverage was not responding to economic and social realities but to the dynamics of politics and policy in Washington. Media coverage of poverty was increasing during the same years in which the government enacted and began serious enforcement of antidiscrimination legislation. These were the same years in which President Lyndon Johnson declared the War on Poverty – a war that was to be fought in the largely black inner cities, not the largely white trailer parks of Appalachia. And the Johnson administration, as we saw in the quotations in Chapter 2 from Newsweek introducing the War on Poverty in early 1964, made that connection explicit. The news media were given a new story line from the chief newsmaker of them all: the president of the United States. Because poverty – and, by extension, black poverty – became a particular and sustained focus of the administration, it also became a natural focus of a news media that, by and large, reports on what politicians do and say. Note, too, the decline in references to black poverty that occurred in every year of the Nixon administration, which did not make antipoverty the centerpiece of its policy attention. When black poverty fell off the president’s agenda, it fell off the media’s agenda as well. The War on Poverty might have been a mere blip on the radar screen – overwhelmed, say, by the Vietnam War – had it not been for some real-world events that served to reinforce the importance of the administration’s efforts. The riots that began in 1965 and continued for several subsequent summers provided the mainstream press with a new lexicon, a new vocabulary to use in their stories of blacks. This frame differed markedly from the prevailing story lines before the Watts riots, which emphasized the civil rights struggle, voting rights, and nonviolent protest. In a way, these frames were no longer “news,” even if the events were. The “black poverty” frame that became available to journalists after Watts represented news, not only in the immediate sense, but also in the sense that they had found a new way to report the news. And the frame persisted far longer than the riots did; indeed, we still hear of “urban ghettoes” in today’s coverage. This frame did not even exist before the mid-1960s. The urban ghettoes themselves, of course, did exist before the mid-1960s, but not as a focus of the nation’s policy apparatus. This goes a long way, in my view, toward explaining media dynamics.

0521821711c05

0 521 821711

June 3, 2003

9:46

Race and the Welfare State in the Public Mind

129

Another reinforcing cause also likely played a role in the proliferation of the poverty frame. After the legislative victories of 1964 and 1965, the civil rights movement’s agenda changed, and open-housing demands (largely in cities) and fighting poverty became rallying calls for the leadership. (Martin Luther King, Jr., when he was assassinated in Memphis, was actually visiting the city to coordinate a Poor People’s Campaign.) To the extent that the mainstream press took its cues from the civil rights leadership, references to poverty would naturally have increased. The president, after all, is not the only newsmaker in the country, even if he is the only one who has a bully pulpit by virtue of his position. King, in particular, had a rather substantial figurative pulpit of his own (in addition to the actual one in his congregation), which he had earned through years of leading nonviolent protest. The press, then, was surely responding to King and his fellow leaders as their agenda shifted to urban issues after 1965. This combination of forces – the leadership of the president, the effects of major events like the riots, and the leadership of nonpresidential civil rights movement figures – produced the dynamics of media coverage we have witnessed. The media, it seems, responded not so much to the social reality being enacted before them but to the unfolding political drama they cover as a matter of course. And this seems very much like an accurate definition of what is “news.”

0521821711c06

0 521 821711

May 30, 2003

7:59

6 A New American Dilemma for a New Millennium?

The latent causes of faction are thus sown in the nature of man; and we see them every where brought into different degrees of activity, according to the different circumstances of civil society. – James Madison, The Federalist Papers, Number 10

My analysis, like all others, leaves several questions unanswered. Although we have uncovered much about the causes of the ebbs and flows of racial policy preferences, we have not begun to explore the consequences of these shifts. To turn our key dependent variable into an independent variable, we can ask: Do these shifts in public opinion on race have policy consequences? Do they have electoral consequences, affecting the fortunes of candidates for public office? If they affect policy, how do they do it? Do they affect various types of racial policy, like antidiscrimination enforcement or affirmative action, or (and?) do they affect broader policies like welfare and the government’s redistributive posture? The answer to some of the above questions, presumably, is yes, though there is no evidence for it in this book. Perhaps I say “presumably” out of a sense of hope or optimism that our policy-making system is at least partly a function of inputs from public opinion. Or perhaps I say it out of fear that, if opinion has no effects, we have studied the causes of something that is itself impotent. Probably both. The questions I have raised here strike me as important. Yet it also strikes me that we know precious little about the over-time dynamics 130

0521821711c06

0 521 821711

May 30, 2003

7:59

A New American Dilemma for a New Millennium?

131

of policy-making in America.1 Perhaps this is the foremost example of the size of the gulf that exists between the empirical study of American politics and the normative concerns of democratic theory. Indeed, the very notion of “empirical democratic theory” is something of an oxymoron in academia these days, because the questions seem too abstract, the measurement problems too intractable, or the connection to normative concerns too tenuous. It is my hope that this can begin to change.

summary of key findings The study of racial policy preferences has heretofore been almost exclusively understood as the study of individual racial attitudes. And we, as a scholarly community, have searched long and hard to uncover the forces that shape individuals’ racial attitudes – that is, why a particular person is a racial liberal or a racial conservative. This book has taken the exact opposite starting point, focusing instead on aggregates and on why collective support for and opposition to particular racial policies will wax and wane over time. We have discovered that racial policy preferences can be coherently captured by three related geological metaphors. First, racial policy preferences are global, in the sense that sentiment on all facets of racial policy tends to drift in parallel fashion. When the public is liberal on busing, it is liberal on spending for blacks, on affirmative action, and the rest; and when it is conservative on one, it is conservative on all. Second, racial policy preferences are glacial, in the sense that they are slow-moving, stable in the short term but never stable in the long term. In any given year, a good predictor of this year’s level of support for a particular policy would be last year’s level of support; but that prediction would fail in the longer term. The slow accumulation of one- and two-point shifts eventually leads to a collective sentiment that is different from what it was, say, a decade ago. And third, we have seen that racial policy preferences are tidal, in the sense that they have the tendency to reach a peak of liberalism, then begin a slow ebb to conservatism, only to reverse themselves again. Like the ocean, where every high tide is inevitably followed by 1

There is one prominent exception to this generalization. See Stimson, MacKuen and Erikson 1995, and Erikson, MacKuen, and Stimson 2002.

0521821711c06

132

0 521 821711

May 30, 2003

7:59

Mass Media and American Racial Attitudes

low tide (and vice versa), every high level of liberalism on racial policy preferences is followed by a drift toward conservatism. And in this study we have learned much about the evolutionary forces that have shaped our geological observations. We have seen that the national media, through their coverage that emphasizes core American values to varying degrees over time, shape the policy preferences of a divided public over time. Growth in egalitarian coverage helps tilt the public toward preferring liberal policies, whereas growth in individualistic coverage reduces support for those very policies. We have also seen how the press’s surge in coverage of poverty in the mid1960s had profound implications for the relationship between attitudes on race and the welfare state, causing two separate streams of public opinion to unite to form a single river. And we have seen that the media themselves are far from static. Their portrayals of race are dynamic, not in the “exciting” sense of the word – though sometimes they are that as well – but in the sense that they are linked to time by their very nature. Their coverage varies in emphasis over the decades, and seems to depend on political leadership for setting their agenda.

implications for the mass media We have, in this book, found several instances where the national media exerted important influences on the trajectory of American public opinion. These findings differ from those of many older examinations of the media–opinion connection, which often found “minimal effects” of media on opinion. The findings here are more consistent with Shanto Iyengar and Donald Kinder’s “not-so-minimal effects,” which they discovered in experimental work involving television news. Accordingly, as with any study that attributes causal power to the mass media, we must address the question of media dominance. Ought we to conclude from this influence that the media are somehow guilty of manipulating public opinion, that somehow they decide what messages to send, and the public in some automatic reaction accepts what the media tell it to believe? Is the public merely some kind of echo chamber that repeats whatever the media tell them? To put it mildly, such a conclusion would, in the domain of race, be novel. And such a conclusion would require that the public have, at best, unformed and excessively malleable attitudes about race. But

0521821711c06

0 521 821711

May 30, 2003

7:59

A New American Dilemma for a New Millennium?

133

never has any scholar alleged that the American public does not have meaningful attitudes on race and mindlessly spouts the opinions suggested to them by the press. The manipulation hypothesis requires a public without real and substantive feelings and beliefs on race. But even Phillip Converse, pessimist that he was about the degree of sophistication of American public opinion, concluded that Americans’ opinions on matters of race had consistency, both in cross-section and over time. Paul Sniderman and Thomas Piazza summarize the sentiment of every scholar in the field when they write that, on matters of race, Americans “have their feet, if not exactly set in cement, then at any rate firmly planted.”2 Such a conclusion of media dominance would also, in my view, be unwarranted – an extreme interpretation of the data presented in this volume. In particular, such a conclusion would ignore two other prominent features of our findings. First would be the fact that we have consistently found other systematic forces in addition to the media to be relevant shapers of public opinion. Along with media coverage, in Chapter 4 we discovered that racial policy preferences are influenced by feedback from the policy system in a kind of thermostatic relationship. More liberal policies lead the public to decrease their demand for those policies. And welfare-state preferences also exert an influence on racial policy preferences. Further, in Chapter 5 we discovered that the merging of attitudes on race and the welfare state was, in addition to the media, caused by this same liberalization of racial policy. These causal influences are, quite simply, not consistent with the characterization of a puppet-like public. The media-manipulation hypothesis also ignores what we have learned about where media coverage comes from. After all, in order for the media to be manipulative in any meaningful sense of the word press coverage itself could not be the result of other forces; it would have to be, in other words, that unmoved first mover. But it is not. The media seem to respond to the agendas of the most prominent American newsmakers – particularly the president of the United States, but also prominent movement leaders with their own agendas. This was particularly true, we have seen, in the case of the coverage of black poverty. But it is also true of coverage of egalitarian and individualist values. In 2

See Converse 1964; Sniderman and Piazza 1993, p. 137.

0521821711c06

0 521 821711

May 30, 2003

134

7:59

Mass Media and American Racial Attitudes

other work, I have shown those types of coverage to be responsive to shifts in the economy and, again, to feedback from the policy system.3 What all of this rather clearly indicates is that the mass media are not, as some aficionados of conspiracy theories would like to suggest, the grand manipulators of public opinion, at least with respect to race politics. The media, to be sure, do play an important role in determining the shape and trajectory of public attitudes on race; but so, too, do other forces. And press coverage itself is a function of other, systematic forces. The media, in the end, are an important part of a system of influences that determine public opinion on race in America; but they are far from manipulators.

implications for public opinion Is public opinion rational? This is a question that public-opinion scholars have grappled with for years. And, you will notice, it is something of a loaded one. To answer in the negative is to cast serious doubt on the value of the democratic experiment. For to say that public opinion is not rational is to imply something unflattering – “irrational,” “emotional,” and “wrong” are adjectives that come to mind as antonyms. Normatively, that is a shaky foundation. But it also seems to me to be the wrong way to evaluate public opinion. If we know that opinion on race is global, glacial, and tidal, we also know that opinion is not haphazardly moody. It is not subject to simple manipulations of slick politicians that cause abrupt reversals of opinion, lead to sudden reversals of public policy, and therefore undermine the stability of government. Because we know that public opinion is both slow-moving and a function of sensible, explicable forces, it is not random at its core. It is explainable. Perhaps that is rational enough. A New American Dilemma? The Swedish sociologist Gunnar Myrdal’s famous work An American Dilemma (1944) set the stage, intellectually speaking, for the civil rights movement that was to begin a decade or so later. His treatise served as 3

See Kellstedt 2000, esp. table 4.

0521821711c06

0 521 821711

May 30, 2003

7:59

A New American Dilemma for a New Millennium?

135

the backdrop for the forthcoming social movement, and it clearly filtered down to the level of the popular debate. His thesis was that there existed a great chasm between America’s egalitarian rhetoric and our clearly nonegalitarian treatment of blacks. On the one hand, almost all Americans were familiar with and proud of the country’s egalitarian principles, which are embodied in the words and writings of their country’s political heroes. “We hold these truths to be self-evident, that all men are created equal,” proclaims the Declaration of Independence. Equally self-evident was the fact that blacks were not considered to be a part of “all men” in the “all men are created equal.” White Americans spoke the language of equality as eloquently as anyone, and these principles were for the most part uttered sincerely. Myrdal did not doubt this. But they were simply not thought to extend to blacks; and he noted that our prejudice was inconsistent with our liberal tradition. Myrdal was not exaggerating when he argued that it was hypocritical to subscribe to the abstract notion that all men are created equal but deny that blacks are a part of the equation. This was Myrdal’s American Dilemma, and no one had identified the problem quite so clearly before he did. But as this American dilemma has faded into the past, a new American dilemma has emerged to replace the old. Should American society endorse policies that attempt to compensate for the injustices of the past, or do such policies create problems bigger than the ones they purport to solve? Should society, either literally or in effect, make amends to its black members, or should society tell its black members to make it on their own? When faced with such a decision, most individuals will reflect on the options and pick the one most consistent with their core values. And that is precisely why this situation is so vexing. For in this case, one of America’s core values – egalitarianism – prescribes activist government policies, while another – individualism – forbids them. This is the new American Dilemma. A critical distinction between the old American Dilemma and the new is that the lines of division have changed. Before, America was divided into two groups of people, some of whom thought blacks deserved the full benefits of American citizenship, others of whom thought that blacks were inherently inferior beings. Today, the fault lines are not between individuals but within them. One part of most Americans cries out for compensatory justice, that our society must

0521821711c06

0 521 821711

May 30, 2003

136

7:59

Mass Media and American Racial Attitudes

atone for its sins, right the wrongs of the past; therefore, government must be involved. But another part objects, drawing on the different (but no less central) core societal value maintaining that, once the legal barriers between blacks and whites have been eliminated, government ought to stay out of the picture and let individuals of any color compete for society’s riches. America now has trouble making up its mind on what to do about race largely because most individuals have difficulty making up their minds on what to do about race. What about Prejudice? The academic literature on racial attitudes has been nearly obsessed with the degree to which prejudice does (or does not) determine racial policy attitudes.4 Our focus here, in contrast, has been, not on the forces that make some individuals either racial liberals or conservatives, but instead on the forces that make the American public as a whole drift in a liberal or conservative direction over time. This concentration on over-time dynamics has naturally taken some of the focus away from bigotry and prejudice as potential causes of opinion. Why? Because, as a hypothetical time series, prejudice has some different characteristics. Unlike most of the time series presented in this book, prejudice is presumably the least likely to move back and forth through time, with increases and decreases in aggregate levels of prejudice through time. But this does not mean that shifts in prejudice do not contribute to the trends in racial policy preferences that we have witnessed. In fact, by controlling for generational replacement in our models of the dynamics of racial policy preferences, we are, in an admittedly crude way, controlling for trends in overt bigotry. Recall from Chapter 4 that generational replacement strongly influenced racial policy preferences. The notion of controlling for the replacement of generations arose from the work of Howard Schuman and his colleagues, and my findings are quite consistent with theirs.5 The view here, though, is that generational replacement has largely, if not quite fully, run its course. The proportion of the voting-age 4

5

A recent edited volume features essays from protagonists on both sides of the debate, focusing squarely on the role of prejudice in contemporary American politics. See Sears, Sidanius, and Bobo 2000. See Schuman et al. 1997.

0521821711c06

0 521 821711

May 30, 2003

7:59

A New American Dilemma for a New Millennium?

137

electorate that was socialized before the civil rights movement hit the peak of its influence in the 1960s is, at the dawn of a new century, about one adult in five, and it shrinks with each passing year. That means that the effects of generational turnover have probably been close to fully realized by now, and any future liberalization in public opinion on race will not be the result of it. Looked at from a different perspective, the measures of what is commonly called biological prejudice – such as the notion that blacks are inherently inferior to whites – have almost hit the ceiling, with close to 100 percent of Americans disagreeing with such sentiments. That is not to say that prejudice is a thing of the past. Not at all. It is merely to say that this particular form of prejudice, which was quite common among previous generations of Americans, has become less common. Based on this rather limited evidence, it is my conclusion that reductions in bigotry, which have coincided almost perfectly with the very slow turnover in generations, have contributed to shifts in racial policy preferences. We have seen the rather large effects of this turnover, but at this point, most of what we see will be in the rear-view mirror. In other words, when looking into the future for the likely dynamic causes of shifts in racial attitudes, it is my suspicion that we need to look at causes other than fluctuations in bigotry. I suspect this because we have no particular reason to conclude that continued generational turnover will lead to large reductions in prejudice.

implications for american democracy and the politics of race Anything approaching a complete, normative evaluation of democratic politics would surely require the consideration of forces beyond those examined in this volume. For, although we have undertaken a significant task in describing the dynamics of public opinion and how the mass media, and other forces, play a role in shaping opinion over time, this is not a complete picture of democratic institutions in America. In particular, we have not taken the role of elections seriously, nor have we considered the determinants of public policy – though we have seen its impact in the policy-feedback loop. Nevertheless, with a fresh set of empirical findings in hand, we are in a position to comment upon how this portion of our democracy is working, in a normative sense.

0521821711c06

138

0 521 821711

May 30, 2003

7:59

Mass Media and American Racial Attitudes

Over two hundred years ago, the elites of colonial America who would become our nation’s Founding Fathers expressed considerable fears about public opinion. The basis of the Founders’ fears, to the extent we can infer it from their writings, was the danger of what they called “faction.” James Madison, one of the authors of The Federalist Papers, begins Federalist 10 with the following words: Among the numerous advantages promised by a well-constructed union, none deserves to be more accurately developed than its tendency to break and control the violence of faction. . . . By a faction I understand a number of citizens, whether amounting to a majority or minority of the whole, who are united and actuated by some common impulse of passion, or of interest, adverse to the rights of other citizens or to the permanent and aggregate interests of the community.

Madison argues that, because factionalism is inherent in human nature, its causes cannot be controlled without resorting to tyranny. Consequently, “relief is only to be sought in the means of controlling its effects.” The Founders sought to do precisely this in their design of the American system of government. In erecting America’s democratic institutions, they placed more than a few institutional barriers between the people and policy outcomes, thus making it more difficult for a fickle or moody public to capture control of the levers of power and make public policy that would be harmful to the best interests of the country. These institutional barriers include such elements as a bicameral legislature, a federal system of government that reserves many powers for the states, and the separation of powers between the executive and legislative branches of government, among others. Their goal was to make permanent factions impossible. They knew that conflict is inevitable in politics and hoped, through the dispersal of power, to create a system where today’s enemies would be tomorrow’s allies. In other words, they wanted to ensure that public opinion would never be permanently aligned to benefit one group to the detriment of another (and, by implication, of all). If they were to base their judgment solely on the findings from this book, it is my belief that the Founders would proclaim their experiment to be a resounding success, that the potential evils of faction have been successfully averted. In the case of race politics, the tide

0521821711c06

0 521 821711

May 30, 2003

7:59

A New American Dilemma for a New Millennium?

139

of public opinion is constantly shifting. Importantly, it is not shifting overnight, but slowly, gradually, over the span of a decade or more. This, as we have noted repeatedly, is not a “moody” public, in the sense that policy moods do not change as our emotional moods do – haphazardly, at times overnight. There are, to be sure, passionate advocates of and opponents to government policies on race, but public opinion as a whole drifts back and forth between support for and opposition to these policies. This creates a system where there are no permanent winners or permanent losers – precisely what the Founders had envisioned. Losers of this year’s legislative or legal battle always have hope that things will go their way next year – perhaps when public opinion is more supportive of their position. And protagonists on both sides of the debate have at their disposal one of America’s cherished values: the opportunity to frame their appeals to a divided public. In this sense, it is my conjecture that the Founders would be very happy indeed. What does all of this portend for the future of race politics? While it would be dramatic, perhaps melodramatic, to conclude with a grand proclamation of optimism or pessimism about the future of race politics, it would also be disingenuous. In short, what I believe we can expect from the future is a good deal more of what we have experienced in the recent past. By this, I mean that we will see a back-and-forth struggle between those pushing for more liberal racial policies and those resisting them. That struggle will take the form of a competition for space on the political agenda, and hence space on the media’s agenda. It will also take the form of competing efforts to frame the debate in the guise of the cherished values that most benefit the particular advocate’s point of view. These back-and-forth movements are unlikely to be perfectly predictable, as the prominence of different values in the media surely depends on the relative skill of the leaders of competing social movements at getting their message out, in “spinning” news stories effectively to the advantage of their respective flocks. In the case of the racial liberals, we will see efforts in the legal and legislative arenas that are directed at emphasizing the egalitarian strain of the American ethos. They will file lawsuits and push for legislation on issues that make their cases resonate easily with our heritage of equality. The conservatives will use similar tactics, but on different points of controversy over race. They will contend in courts of law and the court of public opinion that

0521821711c06

140

0 521 821711

May 30, 2003

7:59

Mass Media and American Racial Attitudes

liberal racial policies are inconsistent with America’s cherished value of individualism. The most recent data presented in this book about the direction of racial policy preferences indicate that they are headed in a conservative direction. That is to say, the conservatives, for the moment, are winning the rhetorical battle. This should not be surprising. The Clinton administration provided plenty of fodder for the conservatives’ cannon. Consider the controversial (and eventually abandoned) nomination of Lani Guineer – who has praised the merits of race-based methods of legislative apportionment and representation – to be Assistant Attorney General for Civil Rights. Conservatives had ample opportunity, and a very public forum, to present their view of the implications of liberal race policies – and bear in mind that President Clinton never endorsed Guineer’s vision for apportioning representation. The public, though, got a heavy dose of individualism during those weeks in early 1993. They are also fed generous portions of individualism when conservatives file case after case in federal courts against affirmative-action programs in places like colleges and universities. Such lawsuits continue to this day and, because of contradictory rulings in lower courts, face what seems like an inevitable showdown in the United States Supreme Court. Affirmative action will be on the government’s agenda well past the day on which the Court decides whatever it decides, as both the winners and the losers cope with the fallout of the decision. But, demonstrating the back-and-forth nature of this struggle perfectly, the liberals hold the high cards in different suits of race politics. After the civil rights movement seemed to end government-based discrimination against racial minorities with its legislative and legal victories of the 1950s and 1960s, the focus of discrimination shifted to the private sector. But recently the issue of public-sector discrimination has returned in the form of racial profiling cases. As the documentation of the extent of racial profiling widens – in Rhode Island, for example, the state government has launched a statewide “driving while black” study to document systematically whether racial minorities are targeted by the state police in highway stops – it gives liberals an opportunity to expose the persistence of discrimination, of inegalitarian treatment. What could be a more blatant example of unequal treatment by the government than revelations that the police systematically pull over drivers

0521821711c06

0 521 821711

May 30, 2003

7:59

A New American Dilemma for a New Millennium?

141

with dark skin more often than they pull over drivers with white skin? I suspect we have not heard the last, or perhaps even the most atrocious, instances of racial profiling. Coverage in the press that documents this practice will resonate with the egalitarian strain of our values. It will remind the public of the everyday reality of unequal treatment, and how this violates our beliefs. This is what I mean when I suggest that we should expect “more of the same.” Compelling dramas about the politics of race sometimes remind us of our individualistic heritage, and other, equally compelling dramas remind us of our egalitarian values. We will be pulled in different directions at the same time, with one side temporarily gaining ascendance and the other responding when opportunity presents itself. This drama will be further complicated by our finding that the politics of race and the politics of the welfare state have increasingly become interconnected. Discussions of poverty inevitably lead to discussions of the “root causes” of poverty. Are individuals to blame for their own situation? Did they not work hard enough? Were they too lazy to take advantage of the opportunities presented to them? Or were there no real opportunities to take advantage of? Were they, through no real fault of their own, dealt an unfair hand by life? Protagonists on both sides of this debate, which is now attached to race politics, will surely cloak their positions in the familiar language of American values.

0521821711c9

0 521 821711

April 11, 2003

12:59

Appendix Survey Items Used in Index of Racial Policy Preferences

Survey Organization: National Election Study (NES) Subject: Accommodations Question Wording: As you may know, Congress passed a bill that says that black people should have the right to go to any hotel or restaurant they can afford, just like anyone else. Some people feel that this is something the government in Washington should support. Others feel that the government should stay out of this matter. Have you been interested enough in this to favor one side over another? [If yes] Should the government support the right of black people to go to any hotel or restaurant they can afford, or should it stay out of this matter? Survey Organization: Trendex Subject: Aid to minorities Question Wording: . . . [Would you] like to see the government do more, less or do about the same amount as they have been on . . . helping minority groups? Survey Organization: General Social Survey (GSS) Subject: Busing Question Wording: In general, do you favor or oppose the busing of (Negro/ black) and white children from one district to another? Survey Organization: Harris Subject: Affirmative action 142

0521821711c9

0 521 821711

April 11, 2003

Appendix

12:59

143

Question Wording: Now let me read you some statements about affirmativeaction programs in education and employment. For each, tell me if you tend to agree or disagree. After years of discrimination, it is only fair to set up special programs to make sure that women and minorities are given every chance to have equal opportunities in employment and education. Survey Organization: Harris Subject: Affirmative action Question Wording: Do you favor or oppose federal laws requiring affirmativeaction programs for women and minorities in employment, provided there are no rigid quotas? Survey Organization: Gallup Subject: Integration Question Wording: Do you think the ——- Administration is pushing racial integration too fast, or not fast enough? Survey Organization: GSS Subject: Help to blacks Question Wording: Some think that (blacks/Negroes) have been discriminated against for so long that the government has a special obligation to improve their living standards. Others believe that the government should not be giving special treatment. Where would you place yourself on this scale? Survey Organization: NES Subject: Help to blacks Question Wording: Some people feel that the government in Washington should make every possible effort to improve the social and economic position of Negroes and other minority groups. Others feel that the government should not make any effort to help minorities because they should be expected to help themselves. Where would you place yourself on this scale? Survey Organization: NES Subject: Busing Question Wording: There is much discussion about the best way to deal with racial problems. Some people think that achieving racial integration of schools is so important that it justifies busing children to schools out of their own neighborhoods. Others think that letting children go to their neighborhood schools

0521821711c9

0 521 821711

April 11, 2003

12:59

144

Appendix

is so important that they oppose busing. Where would you place yourself on this scale? Survey Organization: NES Subject: Fair treatment in jobs Question Wording: If Negroes are not getting fair treatment in jobs and housing, the government should see to it that they do. (Agree/Disagree) Survey Organization: NES Subject: Fair treatment in jobs Question Wording: Some feel that if Negroes are not getting fair treatment in jobs the government in Washington ought to see to it that they do. Others feel that this is not the federal government’s business. Have you had enough interest in this to favor one side or the other? [If yes] How do you feel? Should the government in Washington see to it that black people get fair treatment in jobs or is this not the federal government’s business? Survey Organization: GSS Subject: Spending on blacks Question Wording: . . . Are we spending too much, too little, or about the right amount on . . . improving the conditions of blacks? Survey Organization: GSS Subject: Spending on blacks Question Wording: Are we spending too much, too little, or about the right amount on . . . assistance to blacks? Survey Organization: CBS-New York Times Subject: Affirmative action Question Wording: Do you believe that where there has been job discrimination against blacks in the past, preference in hiring or promotion should be given to blacks today? Survey Organization: GSS Subject: Open housing Question Wording: Suppose there is a community-wide vote on the general housing issue. There are two possible laws to vote on. Which would you vote for? One law says that a homeowner can decide for himself whom to sell his

0521821711c9

0 521 821711

April 11, 2003

Appendix

12:59

145

house to even if he prefers not to sell to (Negroes/blacks). The second law says that a homeowner cannot refuse to sell to someone because of their race or color. Survey Organization: NES Subject: Aid to blacks Question Wording: If you had a say in making up the federal budget this year, for which programs would you like to see spending increased and for which would you like to see spending decreased? Should federal spending on programs that assist blacks be increased, decreased, or kept about the same? Survey Organization: NES Subject: Open housing Question Wording: Some people say that Negroes should be allowed to live in any part of town they want to. How do you feel? Should Negroes be allowed to live in any part of town they want to or not? Survey Organization: Roper Subject: Race and poverty Question Wording: I’m going to show you a list of problems, and I’d like you to tell me if each is something the government should be making a major effort on now, or something the government should be making some effort on now, or something not needing any particular government effort now . . . trying to solve the problems caused by ghettos, race, and poverty. Survey Organization: NES Subject: School segregation Question Wording: Some people say that the government in Washington should see to it that white and Negro children are allowed to go to the same schools. Others claim that this is not the government’s business. Have you been concerned enough about this question to favor one side over the other? [If yes] Do you think the government in Washington should see to it that white and black children go to the same schools, or stay out of this area as it is not its business?

0521821711c9

0 521 821711

April 11, 2003

12:59

146

0521821711ref

0 521 821711

June 2, 2003

12:46

References

Abramowitz, Alan I. 1994. Issue Evolution Reconsidered: Racial Attitudes and Partisanship in the U.S. Electorate. American Journal of Political Science 38:1–24. Alvarez, R. Michael, and John Brehm. 1997. Are Americans Ambivalent towards Racial Policies? American Journal of Political Science 41:345– 74. Bellah, Robert N., Richard Madsen, William M. Sullivan, Ann Swindler, and Steven M. Tipton. 1985. Habits of the Heart: Individualism and Commitment in American Life. Berkeley: University of California Press. Bobo, Lawrence. 1983. Whites’ Opposition to Busing: Symbolic Racism or Realistic Group Conflict? Journal of Personality and Social Psychology 45:1196–1210. Burstein, Paul. 1998. Discrimination, Jobs, and Politics: The Struggle for Equal Employment Opportunity in the United States since the New Deal. Chicago: University of Chicago Press. Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald Stokes. 1960. The American Voter. New York: Wiley. Cantril, Hadley, with Mildred Strunk. 1951. Public Opinion, 1935–1946. Princeton, N.J.: Princeton University Press. Carmines, Edward G., and James A. Stimson. 1980. The Two Faces of Issue Voting. American Political Science Review 74(1):78–91. 1989. Issue Evolution: Race and the Transformation of American Politics. Princeton, N.J.: Princeton University Press. Chong, Dennis. 1991. Collective Action and the Civil Rights Movement. Chicago: University of Chicago Press. Cohodas, Nadine. 1993. Strom Thurmond and the Politics of Southern Change. New York: Simon and Schuster.

147

0521821711ref

148

0 521 821711

June 2, 2003

12:46

References

Colfax, J. David, and Susan Frankel Sternberg. 1972. The Perpetuation of Racial Stereotypes: Blacks in Mass Circulation Magazine Advertisements. Public Opinion Quarterly 36:8–18. Columbia Journalism Review. 1968. Journalism and the Kerner Report. Special Section 7(Fall):42–65. Converse, Philip E. 1964. The Nature of Belief Systems in Mass Publics. In David E. Apter, ed., Ideology and Discontent. New York: Free Press. Cox, Keith K. 1969. Changes in Stereotyping of Negroes and Whites in Magazine Advertisements. Public Opinion Quarterly 33:603–6. Delli Carpini, Michael X., and Scott Keeter. 1996. What Americans Know about Politics and Why It Matters. New Haven, Conn.: Yale University Press. Durr, Robert H. 1993. What Moves Policy Sentiment? American Political Science Review 87:158–70. Entman, Robert M. 1994. Representation and Reality in the Portrayal of Blacks on Network Television News. Journalism Quarterly 71 (Autumn):509– 20. Erikson, Robert S., Michael B. MacKuen, and James A. Stimson. 2002. The Macro Polity. New York: Cambridge University Press. Fisher, Paul L., and Ralph L. Lowenstein, eds. 1967. Race and the News Media. New York: Frederick A. Praeger. Freeman, John R. 1983. Granger Causality and the Time Series Analysis of Political Relationships. American Journal of Political Science 27:327–58. Freeman, John R., John T. Williams, and Tse-min Lin. 1989. Vector Autoregression and the Study of Politics. American Journal of Political Science 33:842–77. Gans, Herbert J. 1979. Deciding What’s News: A Study of CBS Evening News, NBC Nightly News, Newsweek, and Time. New York: Vintage Books. Gamson, William, and Andre Modigliani. 1966. “Knowledge and Foreign Policy Opinion.” Public Opinion Quarterly 30:187–99. Gilens, Martin. 1995. Racial Attitudes and Opposition to Welfare. Journal of Politics 57:994–1014. 1996. ‘Race Coding’ and White Opposition to Welfare. American Political Science Review 90:593–604. 1999. Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy. Chicago: Universtiy of Chicago Press. Giles, Michael W., and Arthur Evans. 1986. The Power Approach to Intergroup Hostility. Journal of Conflict Resolution 30:469–86. Gitlin, Todd. 1980. The Whole World Is Watching: Mass Media in the Making and Unmaking of the New Left. Berkeley: University of California Press. Gordon, Randall A., Jennifer L. Michels, and Caroline L. Nelson. 1996. “Group perceptions of criminal behavior: The accuracy of race-related crime stereotypes.” Journal of Applied Social Psychology 26:148–59. Hochschild, Jennifer L. 1981. What’s Fair? American Beliefs about Distributive Justice. Cambridge, Mass.: Harvard University Press. Huckfeldt, R. Robert, and Carol Kohfeld. 1989. Race and the Decline of Class in America. Urbana: University of Illinois Press.

0521821711ref

0 521 821711

References

June 2, 2003

12:46

149

Hyman, Herbert H., and Paul B. Sheatsley. 1956. Attitudes Toward Desegregation. Scientific American 195:35–9. 1964. Attitudes toward Desegregation. Scientific American 211:16–23. Iyengar, Shanto, and Donald R. Kinder. 1987. News That Matters. Chicago: University of Chicago Press. Johnson, Paula B., David O. Sears, and John B. McConahay. 1971. Black Invisibility, the Press, and the Los Angeles Riot. American Journal of Sociology 76 (January):698–721. Kassarjian, Harold H. 1969. The Negro and American Advertising, 1946– 1965. Journal of Marketing Research 6 (February):29–39. Katz, Irwin, and R. Glen Hass. 1988. Racial Ambivalence and American Value Conflict: Correlational and Priming Studies of Dual Cognitive Structures. Journal of Personality and Social Psychology 55(6):893–905. Kellstedt, Paul M. 2000. Media Framing and the Dynamics of Racial Policy Preferences. American Journal of Political Science 44(2):245–60. Kellstedt, Paul M., Gregory E. McAvoy, and James A. Stimson. 1996. Dynamic Analysis with Latent Constructs. Political Analysis 5:113–50. Key, V. O., Jr. 1949. Southern Politics in State and Nation. New York: Vintage Books. Kinder, Donald R. 1986. The Continuing American Dilemma: White Resistance to Racial Change 40 Years after Myrdal. Journal of Social Issues 42(2):151–71. Kinder, Donald R., and Tali Mendelerg. 1995. Cracks in American Apartheid: The Political Impact of Prejudice among Desegregated Whites. Journal of Politics 57(2):402–24. Kinder, Donald R., and Lynn M. Sanders. 1996. Divided by Color: Racial Politics and Democratic Ideals. Chicago: University of Chicago Press. Kinder, Donald R., and David O. Sears. 1981. Prejudice and Politics: Symbolic Racism versus Racial Threats to the Good Life. Journal of Personality and Social Psychology 40:414–31. Klandermans, Bert. 1997. The Social Psychology of Protest. Oxford: Blackwell. Kluegel, James R., and Eliot R. Smith. 1983. Affirmative Action Attitudes: Effects of Self-Interest, Racial Affect, and Stratification Beliefs on Whites’ Views. Social Forces 61:797–824. 1986. Beliefs about Inequality: Americans’ Views of What Is and What Ought to Be. New York: Aldine de Gruyter. Krosnick, Jon A., and Donald R. Kinder. 1990. Altering the Foundations of Support for the President through Priming. American Political Science Review 84:497–512. Kuklinski, James H., Paul M. Sniderman, Kathleen Knight, Thomas Piazza, Philip E. Tetlock, Gordon R. Lawrence, and Barbara Mellers. 1997. Racial Prejudice and Attitudes toward Affirmative Action. American Journal of Political Science 41:402–19. Lane, Robert E. 1962. Political Ideology: Why the American Common Man Believes What He Does. New York: Free Press.

0521821711ref

150

0 521 821711

June 2, 2003

12:46

References

Lester, Paul, and Ron Smith. 1990. African-American Photo Coverage in Life, Newsweek, and Time, 1937–1988. Journalism Quarterly 67 (Spring): 128– 36. Lipset, Seymour Martin. 1967. The First New Nation. Garden City, N.Y.: Doubleday. Lipset, Seymour Martin, and William Schneider. 1978. The Bakke Case: How Would it be Decided at the Bar of Public Opinion? Public Opinion 1:38–44. Martindale, Carolyn. 1986. The White Press and Black America. New York: Greenwood. Mayer, William G. 1992. The Changing American Mind. Ann Arbor: University of Michigan Press. McAdam, Doug. 1982. Political Processes and the Development of Black Insurgency, 1930–1970. Chicago: University of Chicago Press. McClosky, Herbert, and John R. Zaller. 1984. The American Ethos. Cambridge, Mass.: Harvard University Press. Myrdal, Gunnar. 1944. An American Dilemma. New York: Harper and Row. Page, Benjamin I., and Robert Y. Shapiro. 1992. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago: University of Chicago Press. Piazza, Thomas, Paul M. Sniderman, and Philip E. Tetlock. 1989. Analysis of the Dynamics of Political Reasoning: A General-Purpose ComputerAssisted Methodology. Political Analysis 1:99–119. Poindexter, Paula M., and Carolyn A. Stroman. 1981. Blacks and Television: A Review of the Research Literature. Journal of Broadcasting 25 (Spring):103– 22. Popkin, Samuel L. 1994. The Reasoning Voter: Communication and Persuasion in Presidential Campaigns. 2d ed. Chicago: University of Chicago Press. Quadagno, Jill. 1994. The Color of Welfare: How Racism Undermined the War on Poverty. New York: Oxford University Press. Rahn, Wendy W., John Transue. 1996. “The Political Significance of Fear of Crime.” A Report to the National Election Studies Board of Overseers. Rainville, Raymond E., and Edward McCormick. 1977. Extent of Covert Racial Prejudice in Pro Football Announcers’ Speech. Journalism Quarterly 54 (Spring):20–6. Report of the National Advisory Commission on Civil Disorders. 1968. New York: E. P. Dutton. Schuman, Howard, and Stanley Presser. 1981. Questions and Answers in Attitude Surveys: Experiments on Question Form, Wording, and Context. New York: Academic Press. Schuman, Howard, Charlotte Steeh, Lawrence Bobo, and Maria Krysan. 1997. Racial Attitudes in America: Trends and Interpretations. Rev. ed. Cambridge, Mass.: Harvard University Press. Sears, David O. 1988. Symbolic Racism. In P. A. Katz and D. A. Taylor, eds., Eliminating Racism: Profiles in Controversy. New York: Plenum.

0521821711ref

0 521 821711

References

June 2, 2003

12:46

151

Sears, David O., and Jack Citrin. 1985. Tax Revolt: Something for Nothing in California. Cambridge, Mass.: Harvard University Press. Sears, David O., and Donald R. Kinder. 1971. Racial Tensions and Voting in Los Angeles. In W. Z. Hirsch, ed., Los Angeles: Viability and Prospects for Metropolitan Leadership. New York: Praeger. Sears, David O., James Sidanius, and Lawrence Bobo, eds. 2000. Racialized Politics: The Debate about Racism in America. Chicago: University of Chicago Press. Shuey, Audrey M., Nancy King, and Barbara Griffith. 1953. Stereotyping of Negroes and Whites: An Analysis of Magazine Pictures. Public Opinion Quarterly 17 (Summer):281–7. Sidanius, James, Pam Singh, John J. Hetts, and Chris Federico. 2000. It’s Not Affirmative Action, It’s the Blacks: The Continuing Relevance of Race in American Politics. In David O. Sears, James Sidanius, and Lawrence Bobo, eds., Racialized Politics: The Debate about Racism in America. Chicago: University of Chicago Press. Sigelman, Lee, and Susan Welch. 1991. Black Americans’ Views of Racial Inequality: The Dream Deferred. New York: Cambridge University Press. Skogan, Wesley G. 1995. “Crime and the racial fears of white Americans.” The Annals of the American Academy of Political and Social Science 539 (May): 59–71. Sniderman, Paul M., and Edward G. Carmines. 1997. Reaching beyond Race. Cambridge, Mass.: Harvard University Press. Sniderman, Paul M., and Thomas Piazza. 1993. The Scar of Race. Cambridge, Mass.: The Belknap Press of Harvard University Press. Sniderman, Paul M., Thomas Piazza, Philip E. Tetlock, and Ann Kendrick. 1991. The New Racism. American Journal of Political Science 35:423–47. Sniderman, Paul M., and Philip E. Tetlock. 1986a. Symbolic Racism: Problems of Motive Attribution in Political Analysis. Journal of Social Issues 42(2):129–50. 1986b. Reflections on American Racism. Journal of Social Issues 42(2):173– 87. Stimson, James A. 1994. Domestic Policy Mood: An Update. The Political Methodologist 6(1):20–2. 1997. The Micro Foundations of Mood. In James Kuklinski, ed., Citizens and Politics: A Political Psychology Perspective. New York: Cambridge University Press. 1999 [1991]. Public Opinion in America: Moods, Cycles, and Swings. 2d ed. Boulder, Colo.: Westview Press. Stimson, James A., Michael B. MacKuen, and Robert S. Erikson. 1995. Dynamic Representation. American Political Science Review 89(3):543– 65. Tarrow, Sidney. 1998. Power in Movement: Social Movements and Contentious Politics. 2d ed. New York: Cambridge University Press.

0521821711ref

152

0 521 821711

June 2, 2003

12:46

References

Thurman, W. N., and M. E. Fisher. 1988. Chickens, Eggs, and Causality, or Which Came First? American Journal of Agricultural Economics 70(2): 237–8. United States Commission on Civil Rights. 1982. The Federal Civil Rights Enforcement Budget: Fiscal Year 1983. Washington, D.C.: Government Printing Office. de Tocqueville, Alexis. 1945 [1835, 1840]. Democracy in America. New York: Vintage Books. Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1996. Voice and Equality. Cambridge, Mass.: Harvard University Press. White, Sylvia E., and Tania Fuentes. 1997. Analysis of Black Images in Comic Strips. Newspaper Research Journal 18:72–85. Wlezien, Christopher. 1995. The Public as Thermostat: Dynamics of Preferences for Spending. American Journal of Political Science 39:981–1000. Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge University Press. Zaller, John, and Stanley Feldman. 1992. A Simple Theory of the Survey Response. American Journal of Political Science 36:579–616.

CY252-Ind

0 521 821711

May 29, 2003

14:45

Index

Abramowitz, Alan, 109–10 affirmative action, 5, 42–4, 140; media stories about, 33 Bakke v. Board of Regents, 42, 84 bigotry, see prejudice Brown v. Board of Education, 5, 31–2, 35–6, 84 busing, media stories about, 48–9; public opinion about, 66–7, 69–70, 72–5 Carmines, Edward, 75–6, 108–10 content analysis, examples of, 37, 39–40, 43–4, 46–9 Converse, Philip, 9–10, 55–6, 58, 133 discrimination, media stories emphasizing, 35; public opinion on, 2, 4; reverse, media stories emphasizing, 41–2 Durr, Robert, 96–7 egalitarianism (see also value conflict), 23, 59–60, 139–41; media stories emphasizing, 34–41, 52–3, 85–8, 88–94, 101–2

equality (see also egalitarianism), 118–23 Feldman, Stanley, 57–8 framing, 23, 34, 104; correlations between various types of, 51–3; and its effect on racial policy preferences, 85–94 generational replacement, and its effect on racial policy preferences, 97–8, 101, 103 ghetto, 45–9, 128 Gilens, Martin, 14–5, 76–7, 81, 110–1, 114 Hass, R. Glen, 60–1 Hochschild, Jennifer, 56–7 ideology, as a possible cause of individuals’ racial attitudes, 6 individualism (see also value conflict), 23, 59–60, 139–41; media stories emphasizing, 41–4, 52–3, 85–94, 101–2 integration, see segregation Iyengar, Shanto, 132 153

CY252-Ind

0 521 821711

May 29, 2003

14:45

154 Jackson, Jesse, 32 Johnson, Lyndon, 76 Katz, Irwin, 60–1 Kennedy, John F., 32, 37, 38 Kinder, Donald, 7, 70, 132 King, Martin Luther, Jr., 38, 48, 129 Klandermans, Bert, 25 Lane, Robert, 56 Madison, James, 130, 138–9 mass media (see also framing), and its connection to public opinion, 5, 14–5, 85–8; and race, existing studies on, 15–22; emphasis on egalitarianism, 34–41, 85–8; emphasis on individualism, 41–4, 85–8; emphasis on black poverty, 45–51, 113–4; emphasis on segregation, 35–7; volume of coverage of race in, 30–4, 96 mood, see policy mood Myrdal, Gunnar, 59, 134–5 Newsweek (see also mass media), 106, 127; data base of stories, 27–30 New York Times (see also mass media); comparisons with Newsweek series, 40–1, 44; data base of stories, 29–30 Nixon, Richard, 32, 63 Page, Benjamin, 72 Piazza, Thomas, 62, 133 policy mood, 65, 78–81, 97 poverty, 107; stereotypical portrayals of race and, 18–9, 23; media stories emphasizing black, 45–51, 113–4, 116–8, 126–9

Index prejudice, 136–7; as a possible cause of racial attitudes, 6–7 press, see mass media public opinion (see also racial policy preferences), 134–7; of blacks and whites compared, 71–5; individual-level theories of, 6–7, 56–63; on race relations, 2, 4; on race and the welfare state, 75–81 racial policy, as a cause of racial policy preferences, 98, 123–6; a longitudinal measure of, 99–101, 103 racial policy preferences, aggregate change in, 7–11, 131–2; individual-level theories of, 6–7, 9–10; longitudinal measure of, 63–71; media coverage and its effects on, 85–103; over-time causes of, 83–8, 94–103; valueconflict theory of, 59–60, 134–5; and welfare-state attitudes, 75–81, 106–8, 111–8 racism, see prejudice Reagan, Ronald, 38, 63, 67, 81–2 riots, 84, 128; media stories about, 32, 46–8 Sanders, Lynn, 70 segregation, public opinion on, 2, 4, 66–7, 69; media stories about, 31, 33, 35–7, 48 Shapiro, Robert, 72 Simpson, O.J., 33, 84 Sniderman, Paul, 6, 62, 133 states’ rights, media coverage emphasizing, 95–6, 101–2 Stimson, James, 55, 65, 67, 75–81, 97, 108–10

CY252-Ind

0 521 821711

May 29, 2003

14:45

Index Tarrow, Sidney, 25, 83 Thomas, Clarence, 33, 42–3, 44, 84 Thurmond, J. Strom, 1 de Toqueville, Alexis, 59 value conflict, 10, 23, 59–62, 85–8, 104, 135

155 Washington, Harold, 33 welfare state, 75; public opinion on (see also policy mood), 75–81, 106–8, 111–8, 122, 124–5 Wlezien, Christopher, 98, 100 Zaller, John, 57–8, 61–2