The End of Race? 9780300183597

Citation preview

THE END OF RACE?

This page intentionally left blank

The End of Race? Obama, 2008, and Racial Politics in America Donald R. Kinder and Allison Dale-Riddle

New Haven & London

Copyright © 2012 by Donald R. Kinder and Allison Dale-Riddle. All rights reserved. This book may not be reproduced, in whole or in part, including illustrations, in any form (beyond that copying permitted by Sections 107 and 108 of the U.S. Copyright Law and except by reviewers for the public press), without written permission from the publishers. Yale University Press books may be purchased in quantity for educational, business, or promotional use. For information, please e-mail [email protected] (U.S. office) or [email protected] (U.K. office). Library of Congress Control Number: 2011933857 ISBN 978-0-300-17519-6 (paperback : alk. paper) A cata logue record for this book is available from the British Library. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 10 9 8 7 6 5 4 3 2 1

CONTENTS

Preface vii Introduction: Elections as Revelations 1 Social Groups and the Vote

1

8

2 Divided by Race—and by Gender: The 2008 Democratic Nomination Contest

26

3 Triumph! 67 4 Phantom Landslide

98

5 The Reverend and the General

118

6 President Obama 137 7 The End of Race?

158

Appendix: Scales, Codes, and Auxiliary Results Notes 237 References 275 Index 301

187

v

This page intentionally left blank

PREFACE

This wasn’t supposed to be a book. Several years ago, one of us (Kinder) was about to fi nish a book. That’s what he kept telling himself and others: Almost finished. Nearly done. The subject of that book was the role of prejudice in contemporary American politics. It was organized around Gunnar Myrdal’s famous prediction, spelled out in his masterwork of social science, An American Dilemma: The Negro Problem and Modern Democracy (1944). According to Myrdal, white Americans were caught in a dilemma, suspended between their commitment to democratic principles, on the one side, and their belief in the superiority of the white race, on the other. Myrdal was certain that democratic principles would prevail. Prejudice and discrimination were about to disappear. Kinder referred to the book he was about to finish, with apologies to Spalding Gray, as the “Monster in a Box.” As Gray did with his incomplete novel, Kinder carried his book with him wherever he went, adding what seemed to him to be indispensable insights, scratching vital notes in the margins, imagining additional crucial analyses, never quite coming to the end. The box was large—but the book (Kinder kept saying) was nearly done. Then, out of the blue appeared Barack Hussein Obama. Kinder first learned of him from reading his name on a T-shirt worn by his son Jacob, then age twelve. At the time, Obama was an Illinois state senator; Jacob thought Obama should be president. Kinder began to read about him; soon enough, we all did.

vii

viii

PREFACE

Obama posed a problem for the book in the box, which argued that Myrdal was wrong. Obama’s remarkable rise to prominence and his historic victory in 2008 did not, by itself, prove Myrdal right. But Kinder clearly had some explaining to do. With this in mind, we began in the winter of 2009 to collaborate on what we assumed would be a chapter for the book in the box. We started out by comparing the role of race in the 2008 presidential election to the role played by religion in the Kennedy-Nixon contest of 1960. The resulting paper was long and illuminating (to us at least), but it only told part of the story we needed to tell if we were to get at the real meaning of 2008 for racial politics in America. At some point, we came to the realization that we were writing a book. And so we wrote it. We have many people to thank. Portions of the argument and evidence were presented in Chicago at the annual meeting of the Mid-West Political Science Association and in seminars at Michigan, MIT, and Vanderbilt. On those occasions, we received hard questions, good advice, and welcome encouragement. We especially want to thank Adam Berinsky, Marc Hetherington, Cindy Kam, Tim Ryan, and David Sears. On technical matters concerning the Voting Rights Act and racial polarized voting, we asked for and received fine and speedy counsel from Bernie Grofman, Richard Pildes, David Lublin, and Richard Engstrom. Matt Riddle provided patient and excellent advice on statistical matters. Our understanding of gender and politics owes much to a happy, long-running, and continuing collaboration with Nancy Burns. We thank Reviewer A and Reviewer B for their tremendously helpful reports. Yale University Press has withheld their identities, but we came to know them well, as we grappled with their challenging questions and argued with their alternative interpretations. In the process, we materially modified the manuscript; we would have been foolish not to. We are also grateful to Cindy Kam, Hanes Walton, and Janet Weiss, who plowed through the entire manuscript and presented us with a bushel full of excellent suggestions. In ways large and small, this book is much the better for their efforts. Kinder’s good friend Lance Sandelands read not a word—nevertheless, he bears some responsibility for what follows. Sandelands is not just a close friend but also a perceptive, sympathetic, and wise critic; he seemed happy to talk, again

PREFACE

and again, about our project. In the end, have we succeeded in persuading Sandelands? Not for us to say. But we can say that the book is much clearer and stronger than it would have been without all those conversations. Finally, Samuel Weiss is a fi ne writer and a splendid editor. We know this because, having read all of our sentences, he proceeded to eliminate quite a few, fi x others, and convert awkward constructions into felicitous phrases—and he did all of this cheerfully and diplomatically, despite the Oedipal temptation. Thanks to one and all.

ix

This page intentionally left blank

INTRODUCTION: ELECTIONS AS REVELATIONS

Each election year is a revelation—in the way the electorate is consulted, wooed, or baffled; in the way issues are chosen, presented, or evaded; in the demands and promises made, compromises struck, strains felt tacitly or voiced. The nation at once celebrates and mourns itself. —Garry Wills

On a bitterly cold, sun-splashed Tuesday in Washington, before an enormous and delighted crowd, Barack Obama is about to be sworn in as the forty-fourth president of the United States. Among the guests of honor on the inaugural platform sits John Lewis, a veteran of the civil rights movement and longtime congressman from Georgia. Some forty years before, on a quiet Sunday afternoon, Lewis led a dignified, doublefile procession of some six hundred American citizens from Brown’s Chapel up onto the Edmund Pettus Bridge in Selma, Alabama. The march was headed to Montgomery; its purpose was to convince Governor Wallace (and the nation) of the moral necessity of extending the vote to black Americans—as a legal right and a practical reality. For their impertinence, the marchers were sprayed with tear gas, knocked to the pavement, and savagely beaten. Lewis slumped to the ground, his skull split open by a trooper’s truncheon.1 As Obama makes his way to the front of the inaugural stand, he glances to his right, notices Lewis, moves down an aisle, and bends to embrace Wills’s quote from Nixon Agonistes (1969, p. 46).

1

2

INTRODUCTION

him. On a remarkable day, it is a remarkable moment. In Lewis’s time, African Americans were denied the vote, shut out of politics, and scorned and dishonored by their government. In 2008, African Americans had the vote, and they voted in overwhelming numbers for a black man for president. More astonishing still, millions of white Americans did the same. The country had changed.2 In his inaugural address, the new president did not dwell on this transformation, but he acknowledged it. Toward the end of his speech, Obama invoked the majesty of democratic ideals. Commitment to the American Creed, he said, is “why men and women and children of every race and every faith can join in celebration across this magnificent mall. And why a man whose father less than 60 years ago might not have been served at a local restaurant can now stand before you to take a most sacred oath.” The words were fine, but it was enough to marvel at the sight itself: silhouetted against a background of gleaming white marble, a tall, thin, black man, the president of the United States. Following Garry Wills’s lead, we treat the 2008 election as a revelation—in particular, as a revelation about the place of race in national politics, now and into the future. On the face of it, Obama’s victory seemed to announce that the United States had overcome its past; that at long last, “the most painful of all American struggles” had come to an end.3 Interpreted this way, the 2008 election confirms Gunnar Myrdal’s famous prediction. In 1937, Myrdal, a brilliant young economist and an influential adviser to the Swedish government, was invited by the Carnegie Corporation of New York to undertake a comprehensive study of the “Negro problem” in the United States. Myrdal accepted Carnegie’s offer, moved his family to the United States, threw himself into the project, and eventually produced a masterwork of social science. In An American Dilemma: The Negro Problem and Modern Democracy, Myrdal (1944) put forward a conceptual framework and presented a mountain of evidence that would influence intellectuals, courts, activists, and policy makers for more than a generation. The dilemma that Myrdal claimed to find at the core of American race relations arose out of the glaring contradiction between democratic ideals

INTRODUCTION

and racial discrimination. According to Myrdal, white Americans were caught in a dilemma, suspended between their commitment to noble democratic principles—what Myrdal called the American Creed—on the one side, and their belief in the superiority of their race, on the other. In the struggle between democratic principles and race prejudice, Myrdal was certain that the former was stronger, that the Creed would prevail. Taught in school, preached in church, written into the Declaration of Independence and the Bill of Rights, the American Creed constituted, in Myrdal’s judgment, the “the glory of the nation” (p. lxix). Supported by mere tradition or interest, prejudice and discrimination would be swept away. This prediction was the centerpiece of Myrdal’s influential book. Myrdal concluded that the American Creed’s advance was inexorable and that prejudice’s days were numbered. Indeed, in Myrdal’s view, by the middle of the twentieth century, prejudice was already in full retreat. The theory that blacks were inferior to whites in intelligence and character, and that such inferiorities were inherent and permanent, was now rejected by many educated whites. Such ideas were harder to find in books, journals, and public speeches, and white supremacy was not so regularly nourished by society’s leaders. As a respectable theory, racism was in decline, a casualty of advances in education and transformations in science. In his final chapter, Myrdal wrote that the “gradual destruction of the popular theory behind race prejudice is the most important of all social trends in the field of interracial relations” (p. 1003). To be sure, stray cases of prejudice remained. But according to Myrdal, prejudice based on race had been driven underground and to the margins of society, consigned to “a surreptitious life.” Racism was an American aberration, and it was being stamped out. The nation’s democratic heritage was about to be redeemed. Had he lived to see it, Myrdal no doubt would have interpreted Obama’s victory as a resounding and uplifting display of principle over prejudice. Myrdal would not have been alone in doing so. Indeed, well before Obama’s victory, a gathering tide of scholarship on racial politics in the United States had been accentuating the positive, emphasizing postwar improvements in American race relations, and expressing skepticism about the present-day political power of racism. An excellent and influential

3

4

INTRODUCTION

example is supplied by Stephan and Abigail Thernstrom. In America in Black and White (1997), the Thernstroms offer an analysis that, like Myrdal’s, is optimistic on American race relations. Progress is the most important lesson of the last fifty years, they say, and “revolutionary” changes in whites’ racial attitudes have transformed American politics. “In the years before the civil rights revolution, hard-core white racism was ubiquitous; in the 1990s, it is largely a thing of the past” (p. 499). Racists can still be found occasionally, but according to the Thernstroms, they have “become a tiny remnant with no influence in any important sphere on American life.” As a result, “disagreements over affirmative action, fair housing, and other racial questions are primarily political” (p. 500).4 If Myrdal and the Thernstroms are right, then Obama’s victory is simply an outsized and dramatic example of American redemption, the decisive and final triumph of democratic principles over racial discord. Perhaps this is right, but we don’t think so. As we show here, race was deeply implicated in the 2008 contest. Obama’s candidacy provoked a huge racial divide in the vote. Obama’s support among black Americans was driven significantly by racial group solidarity. Opposition to Obama among white Americans was driven significantly by resentments rooted in race. All things considered, Barack Obama became president in spite of his race. Under conditions less debilitating to the Republicans—John McCain was handed the task of defending an unpopular war and either papering over or explaining away a looming economic catastrophe—Obama would not have been elected at all.

A Look Ahead Does Obama’s victory in 2008 mean the end of race? The short answer is “No.” Saying so is, of course, one thing; establishing the claim is quite another. The latter requires evidence and argument. Supplying both is the object of our book. As for evidence, we go right to the source: to American voters from all walks of life. We analyze high-quality national surveys, relying principally on the National Election Studies (NES), but also the 2006 Cooperative

INTRODUCTION

Congressional Election Study (CCES), the 2007–2008 Cooperative Campaign Analysis Project (CCAP), the Gallup Poll, and surveys undertaken jointly by CBS News and the New York Times. We have more to say about all these studies later on, as needed. We are interested first and foremost in the role of race in American politics today, but we begin with a more general framework, set out in chapter 1, one that will allow us to make comparisons to social cleavages other than race, and to times and settings other than the immediate present. We introduce and develop two essential concepts: racial solidarity among African Americans and racial resentment among white Americans.5 Chapter 2 takes up the extraordinary 2008 Democratic nomination contest. Heading into the race, Senator Clinton was the clear favorite. She enjoyed the backing of her party, endorsements from prominent African Americans, money to burn, and what appeared to be a commanding lead over her rivals. But as we know, in a tight and fiercely contested race, Obama eventually secured the nomination. The object of chapter 2 is to scrutinize Obama’s surprising nomination victory, and as we do so, to supply a comparison between race and gender as principles of political organization and allegiance. Chapter 3 turns to the general election and the grand prize of American politics. Here we compare the role of race in 2008 to the role of religion in 1960. Like Barack Obama, John Kennedy was young, relatively inexperienced, charismatic, and facing long odds. Neither ran as representatives of their “tribes,” but each became entangled by tribal associations—Kennedy by religion, Obama by race. In the end, both prevailed. What lessons can we find in Kennedy’s struggle with religion for Obama’s attempt to manage race in 2008? Over these two chapters, we find race still to be a powerful force in presidential politics. Race matters, but it does so in offsetting ways. Obama’s racial identity generally helped his cause among African Americans and hurt him among white Americans. Determining the net effect of race for the electorate as a whole is the business of chapter 4. There we conclude that Obama should have won the 2008 election in a landslide. He did not. The gap between the vote he received and the vote that we calculate he should have received is due substantially to race.

5

6

INTRODUCTION

In chapter 5, we consider both the progress the nation has made on race and the obstacles that still stand in the way. We wrestle with the question of progress through a close comparison of Barack Obama’s experience in the spring of 2008 with the experience of Jesse Jackson twenty years before. A generation apart, Obama and Jackson campaigned for the same purpose: to win the presidential nomination of the Democratic Party. Obama prevailed, Jackson did not—but (we sometimes forget) Jackson mounted a strong campaign and became the principal challenger to Governor Dukakis, the eventual nominee. Comparing Obama to Jackson opens a window onto how much has changed over the last twenty years in U.S. racial politics. In the second half of chapter 5, we turn to the possibility of color blindness. We do so through an analysis of the public’s reaction to General Colin Powell, back when Powell was thinking about jumping into the Republican presidential race and when serious people believed he could win. Powell, it was said at the time, transcended race. Did he—and did his remarkable popularity with Americans, black and white—anticipate Obama’s rise to prominence and power? Chapter 6 provides a first look at Obama as president, starting with his inauguration in January 2009 and ending with the national midterm elections in November 2010. During this period, the new president was busy. Among other things large and small, Obama appointed Hillary Clinton secretary of state, signed an $800 billion stimulus package, bailed out Chrysler and General Motors, went to Cairo to address the Muslim world, placed Sonia Sotomayor onto the Supreme Court, and engineered a major reform of the U.S. health-care system. In the avalanche of news spilling out of Washington, did Americans begin to forget that their new president was black? In the concluding chapter, we summarize our fi ndings and grapple with their implications. On matters of race and politics, we try to determine how far we have come as a society; to locate where we are now; and to suggest how far we have yet to go. In this respect, we follow in Myrdal’s footsteps. Like Myrdal in An American Dilemma (1944), we take what ordinary Americans think and do about race to be a cardinal test of democracy, a pointed way to gauge the extent to which the United States lives up to its democratic aspirations. If we depart from Myrdal’s optimistic

INTRODUCTION

conclusion—that prejudice was about to be forever swept aside—we certainly share his view on the importance of the test. It is from this perspective that we offer our analysis of Obama’s victory. Properly decoded, the 2008 election can serve as a revelation about race, democracy, and the promise of America.

7

1 SOCIAL GROUPS AND THE VOTE

A man with a ballot in his hand is the master of the situation. He defi nes all his other rights. What is not already given him, he takes. The ballot is opportunity, education, fair play, right to office, and elbow room. —Wendell Phillips, abolitionist

We are primarily interested in understanding one thing in particular: the role of race in the 2008 American presidential election. But we start by putting forward a general framework that enables us to make instructive comparisons to social cleavages other than race and to times and places other than the contemporary United States. We begin by defining two basic terms: politics and groups. Next, we argue that groups become relevant to politics insofar as they are sites of persistent inequality, and we document the fact that race in the United States fulfi lls this condition all too well. With these points established, we turn in the heart of the chapter to a theory of voter choice. The theory advances two principal claims: First, social groups enter the voter’s decision either through identification with the in-group (for example, solidarity among African Americans) or through attitude toward out-groups (for example, racial resentment among white Americans). Second, the aspects of group identity and group attitude that become important in voters’ choices—which aspects are activated—depend on political circumstances.

Phillips is quoted in Gillette (1979, p. 23).

8

SOCIAL GROUPS AND THE VOTE

Politics Politics, according to Charles Lindblom, is a process whereby “people who want authority struggle to get it while others try to control those who hold it.” It is authority, Lindblom says, that provides the “bedrock on which government is erected.”1 Over the long sweep of human history, the struggle over authority has often been chaotic and bloody, won more frequently by force than by reason. The constitutional movement in the West was an attempt, in Lindblom’s estimation, to convert the deadly struggle for authority into more peaceful and orderly procedures. New forms of participation in politics were invented. Grain seizures, collective invasions of forbidden fields and forests, attacks on machines, sacking of private houses, and turnouts were once the established forms of contention. With the development of capitalism and the rise of the nation-state, however, “the interests and organizations of ordinary people shifted away from local affairs and powerful patrons to national affairs and major concentrations of power and capital” (Tilly 1986, p. 395). A new repertoire of collective action began to take shape. No longer so parochial in scope, forms of contention were now addressed to national authorities. No longer so dependent on patrons, collective action was now autonomous and versatile. In place of the grain seizure and the sacking of private homes came the demonstration, the strike, the social movement, and most notably the election campaign. In modern liberal democratic societies like the United States, authority is won primarily through elections.2 Indeed, in the United States, elections are widely regarded as the democratic moment: elections as the linchpin of the democratic machine and voting as “the central act of democracy” (Riker 1982, p. 5). In theory at least, elections are the “critical technique,” as Robert Dahl once put it, for motivating leaders to be responsive to the aspirations and interests of the voters. Do American elections actually work this way? By and large, they do. The two major parties differ on important matters, and when one party wins, it generally pursues policies broadly consistent with the interests and preferences of its core constituents: on taxes, unemployment, inflation, foreign affairs, the size and scope of government in general, and not least, race.3

9

10

SOCIAL GROUPS AND THE VOTE

Of course, American elections have their share of problems. For one thing, they are blunt instruments of influence: insofar as elections shape policy, they do so partially and often retrospectively, sometimes well after the damage has been done. Moreover, voters are to some degree captive of the campaigns they are presented, which rarely meet the standards set by those who place sober deliberation and thoughtful discussion at the center of democratic politics. Nevertheless, elections do perform an instrumental function, if imperfectly. By taking part in elections, American voters can register their pleasure or displeasure with the governing party and return or replace leaders accordingly, thereby setting in motion alterations in government policy. Policy is at stake in elections, but so too are pride and recognition. This is one clear lesson to be taken from the African Americans’ long struggle for the vote. White Americans resisted extending the franchise to blacks so fiercely and blacks demanded the vote so steadfastly not just because the vote would give blacks the power to protect their interests and have some say on matters of policy, but also because the right to vote was understood on both sides of the color line to convey symbolic authority, a special kind of democratic recognition. To whites, the exclusive right to vote was a public and prominent sign of superiority, not to be easily relinquished. According to Gunnar Myrdal (1944): Already in the ante-bellum elections, political campaigning and voting had acquired a ceremonial significance as marking off a distinct sphere of power and responsibility for the free citizen. From Reconstruction on, voting remained to the white Southerner more than a mere action: it was, and still is, a symbol of superiority. Partly because it is a public activity and does not lend itself to privacy or segregation, it becomes so hard for the white Southerner to admit the Negro to full participation in it. To African Americans, the symbolic weight attached to being denied the vote was no less. Frederick Douglass argued that emancipation was not real, that slavery was not abolished, until African Americans had the ballot: “Men are so constituted that they derive their conviction of their own possibilities largely from the estimate formed of them by others. If

SOCIAL GROUPS AND THE VOTE

nothing is expected of a people, that people will fi nd it difficult to contradict that expectation. By depriving us of suffrage, you affi rm our incapacity to form intelligent judgments respecting public measures” (Foner 1955, pp. 159–160). When individuals are denied the vote, they feel scorned and dishonored; with it, they are invested with democratic responsibility and dignity. From this perspective, participation in elections is “an affirmation of belonging.”4 Much more could be said about voters and elections. Some of it we say in subsequent chapters, as our analysis requires. For now, we are satisfied if we have established two basic points: First, politics is a struggle for authority, and in democratic systems, that struggle is carried out importantly through elections. Second, elections can deliver or withhold two kinds of prizes: the instrumental prize of policy and the expressive prize of recognition.

Social Group Any aggregation of individuals can be a group, if the aggregation is experienced as such. Women, college professors, the neighborhood bridge club: all “are groups in so far as they are social categories or regions in an individual’s social outlook—objects of opinions, attitudes, affect, and striving.” Groups do not require institutional sponsors, formal membership, or face-to-face interaction—though they might have all three. Any collection of people that constitutes a psychological entity for any individual becomes, thereby, a group.5 Defined this way, groups can be enormous (women) or tiny (the neighborhood bridge club). Because of our preoccupation with politics on a national scale, we are naturally drawn much more to the former than the latter. (We attach the modifier “social” to “group” to convey our special interest in groups of substantial size.) When the national government and the resources it commands become the objects of politics, then group attachments and oppositions based in particularistic features, like kin or local community, are subordinated to attachments rooted in broader groupings, such as race, gender, and religion (Posner 2004). This may seem straightforward, but from one perspective at least, it is quite puzzling. Broad social groups—blacks and whites, men and

11

12

SOCIAL GROUPS AND THE VOTE

women—are invisible. That is, such groups are, as Solomon Asch once put it, “too large to be perceived at once” (1952, p. 227). Yet we experience them as familiar, real, and whole, and the puzzle is why. Donald Campbell’s (1958) ingenious analysis of perceiving aggregates of individuals as social entities provides a solution. Conceding that social groups are not as solid as material objects and their boundaries are less clear, Campbell argues that we nevertheless see them in the same way we see material objects. Perceptual principles “are equally applicable to stones and social groups” (Campbell 1958, p. 18).6 Campbell offers four principles from research on human perception that govern when discrete elements (individuals) are perceived as parts of a whole organization (group), and Robert Abelson has more recently added a fifth:7 1. Similarity: Similar elements are more likely to be perceived as parts of the same organization. Similarity—grouping by common features—is a necessary condition for group perception. 2. Proximity: Elements close together are more likely to be perceived as parts of the same organization. 3. Common Fate: Elements that move together in the same direction and otherwise share a “common fate” are more likely to be perceived as parts of the same organization. Insofar as outcomes and opportunities are shared across many occasions, to that degree, individuals will tend to be seen as constituting a single group. 4. Good Figure (Pregnance): Elements forming a part of a spatial pattern tend to be perceived as part of the same unit. This principle is relevant to the specification of boundaries. Good figures “resist” intrusion; they are “opaque” to probing and their boundaries are relatively impermeable. 5. Coordinated Action: Abelson points out that often in political discussions, “ethnic groups and nations are treated as if they are active organisms with hopes, plans, intentions, grievances, moods, and the like: The Palestinians yearn for a homeland, the Serbs have a deep-seated animus against the Muslims, the Christian fundamentalists are expanding their power base in the Republican Party, and so forth” (Abelson et al. 1998, p. 248). Aggregates of individuals are more likely to be taken as a group when they are seen to carry out coordinated action to achieve common objectives.

SOCIAL GROUPS AND THE VOTE

In sum, we are prone to see aggregation of individuals as members of groups when they display the perceptual properties of similarity, proximity, common fate, good figure, and coordinated action.8 Notice that racial groups fulfill these criteria well. Members of racial groups share some conspicuous features: the “physical insignia” of skin color, hair texture, facial features, accent and cadence of speech, and so forth. Thanks in part to the stubborn persistence of segregation, they often find themselves isolated, in close proximity only to one another. They share a common fate: to some degree, they are treated alike, suffering injury and insult or opportunity and honor not because of who they are as individuals, but because of the racial group they happen to represent. They display boundary maintenance, as expressed in strong (if slowly declining) preferences for within-group marriage. And finally, members of racial groups also display signs of coordinated action; to this degree, they are seen not only as social entities but also as political ones, with collective aspirations and common interests.

Categorical Inequality In the United States, as in other advanced industrial societies, individuals vary tremendously in wealth, power, and status. Inequal ity is generated in part by individual differences in talent and enterprise; by luck, good and bad; and, most relevant to our purpose here, by recurrent social processes whereby different social groups are subject to systematically different treatment. Some form of inequality accompanies virtually all social interactions. Most of the time, such inequality is fleeting. Durable inequality generated by recurrent social processes is a different matter. In Charles Tilly’s (1998) analysis, differences in advantage that pivot on categorical opposites— black versus white, Muslim versus Jew, male versus female, citizen versus foreigner—are especially likely to endure. According to Tilly, “paired and unequal categories do crucial organizational work, producing marked, durable differences in access to valued resources. Durable inequality depends heavily on the institutionalization of categorical pairs” (1998, p. 8). In Tilly’s scheme, systems of enduring categorical inequality are established by two general processes. The first of these is exploitation, whereby

13

14

SOCIAL GROUPS AND THE VOTE

members of a categorically bounded network command resources from which they draw significantly increased returns, accomplished by coordinating the efforts of outsiders whom they exclude from the full value added by their effort. Over all of recorded history, “both mighty emperors and petty tyrants have organized exploitation around categorical distinctions” (1998, p. 88). Complementing exploitation is a second mechanism, opportunity hoarding, whereby members of a categorically bounded network gain control over a valued resource from which outsiders are excluded. Depending on time and place, hoarding might refer to territory, education, financial capital, instruments of coercion, employment, or any other valued resource. Once established, categorical inequality is generalized by a process of emulation, whereby existing inequalities are transplanted from one setting to another. This can take place when exterior categorical differences are matched to internal categorical differences—as in labor markets, when a firm assigns high-paying jobs that promise advancement to one group (say, whites) and low-paying, dead-end jobs to another group (say, blacks). Other firms follow suit. Eventually, the practice generates pools of workers with different experiences and capabilities defi ned along group lines. Firms hire and promote accordingly. The result is categorical inequality entrenched within an entire industry.9 According to Tilly, inequality is locked into place through adaptation, whereby daily routines are organized around categorical distinctions. One variety is the invention of norms governing day-to-day interaction between members of categorically unequal groups, as in the extensive and intricate system of deference that grew up in the Jim Crow South. Racial “etiquette” guided every detail of every encounter—forms of address, topics of conversation, appropriate demeanor, and more—thereby providing blacks and whites a regular reminder of the unbridgeable gulf that separated them.10 As categorical inequality spreads, members of advantaged groups begin to create what Elizabeth Anderson calls “stigmatizing stories.” Their purpose is to explain and rationalize inequality. Glaring differences between groups in wealth, power, and status reflect corresponding differences between groups in talent or virtue or culture. By and large, such stories do not cause inequality. Recurrent social processes do. Exploitation, opportunity hoarding, emulation, and adaptation

SOCIAL GROUPS AND THE VOTE

are the engines of inequality. Stories are important, however, because they justify and fortify inequality organized around categorical opposites.11

Race as Categorical Inequality A grotesque example of categorical inequality is supplied by slavery. Beginning in the early part of the seventeenth century, West Africans were taken forcibly from their homelands, put in chains, and shipped under nightmarish conditions across the ocean to the American South, there to provide cheap labor for the burgeoning plantation economy. By the time of the first U.S. Census in 1790, African Americans—nearly all slaves—made up roughly 20 percent of the national population and more than one-third of the population of the South. Slavery, imposed and maintained by violence, was at the center of the new American economic order.12 Today, of course, slavery is gone; the Jim Crow regime of racial oppression that followed emancipation has been dismantled, and discrimination on account of race is now illegal. All of this is true, and all of it is important. As far as race is concerned, the United States is a far more egalitarian society than it once was. Does this mean that race and disadvantage are no longer connected? No. Evidence to the contrary is overwhelming. African Americans have made significant inroads into the middle class over the last sixty years, sharing in the economic prosperity that came to all of American society following World War II. However, racial differences remain and they are imposing. One-third of the black children in the United States live in poverty, more than three times the rate of white children, and black children are much more likely to experience continuous and persistent poverty. Among adults, blacks are twice as likely to be unemployed; they are substantially overrepresented among “discouraged workers,” those who have given up looking for work and therefore do not appear in official unemployment figures; and when blacks are employed, they earn less. These differences are large, but they are nothing compared to racial differences in wealth. According to recent figures, the average white household commands more than ten times the financial assets of the average black household.13 Racial differences in fundamental aspects of health have proven stubborn as well. Infant mortality provides a disturbing case in point. While black women who bear children today are much less likely to lose an in-

15

16

SOCIAL GROUPS AND THE VOTE

fant than were their parents and grandparents before them, the mortality rate remains more than twice as high for blacks than for whites. Black children who survive their first year can expect poorer health, more illness— asthma, diabetes, heart disease, and cancer—and, on average, a shorter life.14 Discrimination by race has been illegal since the 1964 Civil Rights Act, and surely it is neither as flagrant nor as pervasive today as it once was. But it is not gone. Scores of careful studies make clear that African Americans still face discrimination in labor markets. Black people looking to purchase homes are still steered away from white neighborhoods and are still subject to racial bias in mortgage lending. African Americans continue to endure racist epithets on the streets; harassment by police officers in public spaces; rudeness, excessive surveillance, and higher prices while they shop; coolness from their teachers and bosses; and racist jokes from their coworkers. While whites tend to believe that discrimination is a problem of the past, many blacks see it as pervasive in society and a demoralizing presence in their own lives.15 What about inequality in politics? In the early years of the twentieth century, white-dominated legislatures and constitutional conventions throughout the South enacted an assortment of devices designed to banish blacks from political life. These included the poll tax, literacy and property tests, the understanding clause, the good character clause, and not least, the white primary. Blacks initiated legal action, held meetings, organized election campaigns, petitioned constitutional conventions bent on rescinding their suffrage, and where permitted, voted against the new suffrage restrictions. But their efforts were unavailing. Disfranchisement proceeded apace. Blacks disappeared from politics.16 All these formal obstacles to black participation are gone now, swept away by scores of local struggles, Supreme Court decisions, the Voting Rights Act of 1965, and the threat of federal intervention. Black participation in political life towers over what it was a generation or two ago. As a consequence, many blacks now hold positions of political authority.17 Progress on this front has been dramatic. In 1965, the year of the Voting Rights Act, just four of the 435 elected officials serving in the U.S. House of Representatives were black. Not a single African American served in the Senate; only three were mayors of American cities. In the entire coun-

SOCIAL GROUPS AND THE VOTE

try, fewer than three hundred blacks held elected office, most as members of school boards or city councils. Within a decade, the number of blacks holding elective office across the nation increased more than tenfold. The upward trend continued through the 1970s, but now is leveling off—and leveling off well below strict proportionality. While African Americans make up roughly 13 percent of the voting-age population in the United States, they comprise less than 2 percent of elected officials. Blacks have made impressive gains in politics—illustrated in a most dramatic way by Barack Obama’s historic victory in the 2008 presidential election—but they remain, as a general rule, substantially underrepresented in the corridors of power.18 Taken all around, the quality of life experienced by black Americans has improved notably since World War II. However, over the same period, racial differences in the quality of life have persisted. Some differences between blacks and whites have diminished, others have increased, and still others have changed not one iota. All in all, race in the United States continues to provide a compelling case of enduring inequality.

Voting and the Social Group Our purpose here is to offer a general framework for analyzing and understanding how social difference shapes politics—what might be called “A Theory of Voting with Social Groups in Mind.” (The language is extravagant, but it does convey what we are up to.) The theory is intended to apply to the case of race and Obama in 2008, of course, but it should prove useful in understanding social groups other than race and political moments other than 2008. First Premise: Politics Is a Sideshow In a series of influential essays written in the aftermath of World War I, Walter Lippmann, perhaps the most prominent public intellectual of the day, argued that the typical citizen—parochial in interest, modest in intellect, and most of all preoccupied with private affairs—lacks the wherewithal to grasp political matters in any deep way. People are busy; politics is complicated. To expect ordinary people to become absorbed in the

17

18

SOCIAL GROUPS AND THE VOTE

affairs of government would be to demand of them an appetite for political knowledge quite peculiar, if not actually pathological. Lippmann was right. Family, work, and health are central preoccupations. In the meantime, for nearly all of us almost all the time, the events of political life remain peripheral curiosities. In modern societies like the United States, “politics is a sideshow in the great circus of life.”19 If this is so, citizens may well wonder why they should take the trouble to become informed about public affairs. Indeed, many do not. On matters of politics, Americans are often astonishingly ignorant. This fact places a premium, from the voter’s point of view, on cues that are readily at hand and rich in information. Cues like these: Barack Obama is black, John Kennedy is Catholic, Hillary Clinton is a woman.20 Second Premise: Social Motivation On those occasions when voters do turn their attention to politics, they are motivated in part by social concerns. In “Rational Fools,” presented as a rebuke to his fellow economists, Amartya Sen argued that a person propelled entirely by calculations of self-interest would be “close to a social moron” (1977, p. 336). With Sen, we assume that people are not social morons, that they are motivated by more than sheer egoism. They act not only on their own behalf, but for and against social groups as well. Human nature includes a social aspect, and this is expressed in all domains of life, including the political.21 Social motivation is underpinned by ethnocentrism, a deep human predisposition to divide the social world into in-groups and out-groups. The term was introduced by William Graham Sumner, who proposed that members of human groups are sure that their way of doing things is superior to the way things are done elsewhere. As Sumner put it, “There is a right way to catch game, to win a wife, to make one’s self appear, to cure disease, to honor ghosts, to treat comrades or strangers, to behave when a child is born, on the warpath, in council, and so in all cases which can arise” (Sumner 1906, p. 28). Sumner was convinced that ethnocentrism was a universal feature of human society—and he was not far wrong. When referring to outsiders, human populations resort readily to terms of condescension, distrust, and contempt. When referring to insiders, an entirely different vocabu-

SOCIAL GROUPS AND THE VOTE

lary comes into play: respect, trust, and pride are now the terms of choice. Around the world, ethnocentrism prevails. Around the world, that is to say, there exists a human appetite for politics organized by social groupings.22 Third Premise: Bounded Rationality Our third premise goes to how voters decide. Here we take instruction from a general theory of human judgment, the cumulative and considerable achievement of the last half century of cognitive science, a development led most notably by Herbert Simon, Daniel Kahneman, and Amos Tversky. The general theory begins with Simon’s notion of bounded rationality, the assertion that “human thinking powers are very modest when compared with the complexities of the environments in which human beings live” (Simon 1979, p. 3). We presume that the decisions voters make in the polling booth, like the decisions they make in other domains of life, are governed by bounded rationality.23 From the extensive literature on bounded rationality, we take three important lessons for voter decision making. First, voters form judgments and make decisions based on only a small sample of what they know. They are limited in computational capacity, and they search very selectively. The “search is incomplete, often inadequate, based on uncertain information and partial ignorance, and usually terminated with the discovery of satisfactory, not optimal, courses of action” (Simon 1985, p. 295). Second, voters form judgments and make decisions intuitively. Under intuitive thinking, considerations come quickly and spontaneously to mind, without conscious search or computation. The operations of the intuitive system are fast, automatic, and emotionally charged. Intuitive thinking can be interrupted and superseded by effortful reflection, but this happens only under special circumstances: “People are not accustomed to thinking hard, and are often content to trust a plausible judgment that quickly comes to mind.” This is especially so for politics, for most of us, most of the time.24 Third, voters form judgments and make decisions in light of a particular way of looking at a problem—in light of a particular frame. Frames operate by altering the relative salience of different aspects of the decision, highlighting some features of the situation and masking others. Framing

19

20

SOCIAL GROUPS AND THE VOTE

is powerful because people generally passively accept the frame they are given—in politics, as in other domains of life.25 Social Groups as Factors in the Vote We propose that social groups can become a factor in the voter’s decision in one of two ways. The first is through identification with ingroups. The second is through attitude toward out-groups. These two paths reflect the dual role played by social groups in everyday life. Ingroups provide solidarity and opportunities for coordination. Out-groups supply points of comparison and targets of resentment. social groups as objects of identification In The People’s Choice, Paul Lazarsfeld, Bernard Berelson, and Helen Gaudet sought to discover “how and why people voted as they did” (1944, p. 1). In this particular instance, the people were residents of Erie County, Ohio, and the election in question was the 1940 contest between Franklin Roosevelt and Wendell Willkie. Lazarsfeld, Berelson, and Gaudet were sociologists by training, and they believed that the primary function of the presidential campaign was to activate and reinforce predispositions rooted in social groups. Consistent with this hunch, they found that middle-class Protestants living in the country tended to vote for Willkie in large numbers; meanwhile, city-bound, working-class Catholics supported Roosevelt overwhelmingly. The authors concluded “A person thinks, politically, as he is, socially. Social characteristics determine political preferences” (1944, p. 27). Lazarsfeld, Berelson, and Gaudet were on to something. However, their assertion of the primacy of social characteristics was more a manifesto than it was a theory. They had little to say about why religion or class should govern political choice. We say it is inequality that largely underlies the relationship between social characteristics and political preferences. Social groups—some social groups—are sites of durable inequality. This means that, to take one example, blacks and whites will naturally develop different and distinct interests. Their views on family, children, schooling, work, and more will differ. The values they take to be central to the orga ni zation of society and to their own lives will differ. Accordingly, what they want, need, and hope to receive from government will differ as well.

SOCIAL GROUPS AND THE VOTE

Further, we say that the political consequences of social group membership are carried disproportionately by those who belong to the group psychologically. Identification presumes that Americans are social creatures and that political opinions are “badges of social membership,” serving as “declarations, to others and to ourselves, of social identity” (Smith, Bruner, and White 1956). Identification implies a willingness to say “we”; it entails not just membership but also awareness and attachment— self-consciousness that one is a member of a group and value invested in the membership.26 Group identification comes in two main varieties: common fate and emotional interdependence. Common fate refers to the extent to which individuals believe that their life chances and outcomes are intertwined with the opportunities and experiences of their group—that what happens to their group will happen to them. Those who identify with their group on grounds of common fate will come to a political choice with their group’s interests prominently in mind.27 Group identification is grounded also in emotional interdependence, occurring when individuals feel close to their group, experiencing pride when other group members do well and anger when they are treated unfairly. Emotional interdependence reflects the expressive side of politics. To the degree that Americans derive their sense of self from their membership in social groups, political choice becomes an act of affirmation and solidarity.28 Group identification is categorical in that social groups are types or kinds. At the same time, group identification is dimensional in that psychological attachment to a group varies continuously. For some group members, attachment is effectively zero; for others, identification with a social group constitutes a central aspect of identity; and there exist all shades in between. Strength of identification is a sign of a person’s priorities. The stronger the identification, the more powerful the political consequences of group membership will be.29 social groups as objects of attitude Membership groups are important, but as reference group theory reminds us, people “frequently orient themselves to groups other than their own” (Merton and Rossi 1968, p. 35). Such orientations—what we call “attitudes”—are occasionally sympathetic, but they are more often hostile, a reflection, in part,

21

22

SOCIAL GROUPS AND THE VOTE

of ethnocentrism, the human predisposition of turning difference into animosity. By “attitude,” we refer to a form of orga nized readiness, a tendency to act in favor of or in opposition to some object. Attitudes have consequences: they affect “perception, judgment, memory, learning, and thought” and “determine for each individual what he will see and hear, what he will think, and what he will do” (Allport 1935, p. 806). Attitudes toward social groups are especially relevant here because they offer an appealing (humans are social creatures) and efficient (politics is a sideshow) way to simplify the complexities of electoral choice. In his famous essay on belief systems in mass publics, Converse (1964) demolished the idea that ordinary Americans might approach the political world equipped with broad ideological points of view. He also offered a proposal for how citizens might reason about politics if ideology was beyond their reach. Perhaps, Converse suggested, citizens orga nize their opinions on policy according to their attitudes toward the social groups that such policies seem to benefit or harm. To illustrate his argument, Converse invited the reader to imagine a set of policies formulated to benefit or harm a single social group—in this case, and in the vernacular of the times, “Negroes”: Negroes should be kept out of professional athletics. The government should see to it that Negroes get fair treatment in jobs and housing. Even though it may hurt the position of the Negro in the South, state governments should be able to decide who can vote and who cannot. Converse argued that in these hypothetical proposals, the key feature is not states’ rights or federal authority; it is the social group. Because abstractions “take on meaning only with a good deal of political information and understanding, the attitude items given would tend to boil down for many respondents to the same single question: ‘Are you sympathetic to Negroes as a group, are you indifferent to them, or do you dislike them?’ ” (p. 235). Converse was using race to make a general point about the potential of attitudes toward social groups to orga nize political opinions—and we now know that he was right. Scores of studies show that public opinion

SOCIAL GROUPS AND THE VOTE

on matters of policy is group-centric: shaped in powerful ways by the attitudes that citizens harbor toward the social groups they see as the principal beneficiaries or victims of the policy. For example, support for tightening welfare benefits derives from hostility toward the poor (Gilens 1999); opposition to government action against AIDS turns on contempt for homosexuals (Price and Hsu 1992); resistance to immigration reflects suspicions that the new immigrants are somehow un-American (Citrin, Reingold, and Green 1990). Attitude toward out-groups is not the only force driving opinion in these various policy disputes, but it is always present, and of all the forces that shape opinion, it is often the most powerful.30 We expect the same logic to apply to voting. When a presidential candidate is seen as standing for or against a certain social group, voters will be attracted or driven away, depending on their attitude toward the group in question. Social Groups as Short-Term Forces In either manifestation—as an object of identification or as an object of attitude—social groups operate as short-term electoral forces. Imagine the vote cast by a population partitioned into two basic components. First is the “normal vote,” the vote that would be expected from the population under normal conditions; the vote we would see in the absence of short-run disturbances. In American presidential elections, the normal vote is represented primarily by party identification. Most Americans think of themselves as Democrats or Republicans. This attachment to party is a standing commitment—a “persistent adherence,” as the authors of The American Voter put it—that profoundly influences how citizens see the world of politics. In the United States, voting for president is first and foremost an affirmation of partisanship.31 The second component of the vote reflects deviation from baseline—a departure from normal conditions—due to immediate circumstances. Such short-term forces push voters toward one side or the other. Such forces are important—indeed, in particular cases, they can be crucial—but their influence is temporary. They affect the vote in one election without necessarily carrying over to the next.32 Unpopular wars, sharp recessions, eruptions of domestic unrest, the emergence of new and divisive issues: All these can be electoral forces

23

24

SOCIAL GROUPS AND THE VOTE

operating in the short-term. Under the right circumstances, so too can social groups.33 Activation Americans belong to many social groups at once. They are simultaneously black, white, or brown; Catholic, Protestant, Jewish, Muslim, or atheist; male or female; bankers or carpenters; urbanites or suburbanites; Southerners or Yankees; and so on. This means that Americans have available, in principle at least, an extensive repertoire out of which to construct a social identity. In parallel fashion, Americans possess attitudes toward social groups—toward Jews, women, bankers, and many others— any of which could be relevant to their electoral choice. In the short run, which aspects of identity and attitude become important—which are activated—depend on political circumstances. Social groups will have more or less potency as short-term electoral forces depending on the election frame: on the prominence and clarity of cues signaling that the candidates differ substantially in the social groups they favor and oppose. Putting the point in extreme form, in the absence of such cues, social groups will disappear from the voter’s calculus.34 Cues signaling a candidate’s alignment with social groups can take various forms. Candidates can propose particular policies that visibly favor some groups at the expense of others: think of Al Gore’s fierce defense of Social Security. They can emphasize or neglect problems that are of special concern to a particular group: consider Bill Clinton’s pledge to “end welfare as we know it”. They can keep certain company, spending time in the public eye with iconic representatives of one group or another: think of Richard Nixon’s foray into the Deep South in 1968, accompanied by Senator Strom Thurmond of South Carolina, a hero of Southern resistance. Perhaps the most effective signal of all is membership itself. In 1960, John Kennedy claimed to be Catholic; others said he was, too—no one denied it. Kennedy was Catholic, and as we see in chapter 3, this simple and well-publicized fact played a prominent and consequential part in the story of the 1960 election. Even more effective is to embody membership, as Barack Obama embodied race in 2008. Obama did not need to claim to be a black American. Indeed, for the most part, he designed a campaign to evade the claim. But

SOCIAL GROUPS AND THE VOTE

Obama’s effort to neutralize race could not succeed completely, for it was continually subverted by his body—the color of his skin, the features of his face, and the texture of his hair. Whatever Obama said about society and government and about problems and policies, at the end of the day, every time American voters caught a glimpse of him, he was black.

Summary Politics is a struggle for authority. In democratic systems, that struggle is carried out importantly (though not exclusively) through elections. Social groups become relevant to politics insofar as they are sites of durable inequality—and in the American context, there is no better example than race. Voters are immersed in private life, motivated in part by concern for others, and boundedly rational in their decision making. They are partisans, first and foremost, but they can be influenced by social groups as well and in two ways: through identification with in-groups and through attitude toward out-groups. Which aspects of identification and attitude become important—which are activated—depend on political circumstances. With these conceptual distinctions made and the theoretical machinery built, we are ready to begin our empirical investigation. We start with Barack Obama’s remarkable victory over Hillary Clinton in the 2008 contest for the Democratic presidential nomination.

25

2 DIVIDED BY RACE—AND BY GENDER: THE 2008 DEMOCRATIC NOMINATION CONTEST

More people have voted for me than for anyone who has ever run for the Democratic nomination. —Hillary Clinton

In the summer of 2004, Barack Obama was an Illinois state legislator, a part-time professor of constitutional law at the University of Chicago, and the author of a well-regarded and modest-selling memoir (Dreams of My Father). Fresh from a surprising victory in the Illinois Democratic Party primary, he was spending his days and nights in places like Waukegan, Pekin, and Rockford, running hard for a seat in the U.S. Senate. On the national scene, that is to say, Obama had not yet arrived.1 As Obama was beating the bushes in Illinois, John Kerry was doing the same all over the country in pursuit of the Democratic Party’s presidential nomination. When Kerry’s schedule brought him to Chicago, he and Obama made several joint campaign appearances—at a vocational center, a bakery, a town hall meeting, and a fund-raiser at the Hyatt Hotel downtown. Kerry saw firsthand Obama’s ability to reach an audience. After securing the nomination, Kerry invited Obama to deliver the Tuesday night keynote address at the 2004 Democratic National Convention, to be held in August at the Fleet Center in Boston. Obama and his staff were pleased to be asked, but disappointed that the keynote address was not scheduled

Quote from Hendrik Hertzberg, “Memory Lapse,” New Yorker, June 2, 2008.

26

DIVIDED BY RACE—AND BY GENDER

to be covered live by the major networks. The campaign had been lobbying for a slot in “the window.” The keynote address would have to do.2 It did. Obama’s speech electrified the party faithful in the Fleet Center and millions more watching at home. He began that night with the story of his life and his improbable rise to prominence: His father was born and raised in a village in Kenya, attending school in a tin-roofed shack; his mother grew up in a small town in Kansas, a child of farmers. “I stand here,” Obama said, “knowing that my story is part of the larger American story, that I owe a debt to all of those who came before me, and that in no other country on Earth is my story even possible.” Obama went on to sound what would become a familiar theme in his presidential campaign—the deep connection that he believed joined all Americans together: “There is not a liberal America and a conservative America,” Obama declared, “there is the United States of America. There is not a black America and white America and Latino America and Asian America; there’s the United States of America. We are one people.” Obama concluded by calling upon Americans to overcome their cynicism and to participate instead in a politics of hope: “hope in the face of difficulty, hope in the face of uncertainty, the audacity of hope! In the end, that is God’s greatest gift to us, the bedrock of this nation, a belief in things not seen, a belief that there are better days ahead.” Inside the Fleet Center, Democrats nodded their heads, rose from their seats, hollered, and shrieked. Some wept. Even the world-weary press corps seemed impressed. A few went as far as to predict that Obama might become the country’s first black president—though no one was so rash as to imagine it could happen in 2008.3 The keynote speech marked Obama’s arrival on the national stage. Returning to the campaign trail in Illinois, he began to draw enormous and enthusiastic crowds. Obama’s memoir, Dreams of My Father, originally published in 1995 to favorable reviews and modest sales, was reprinted and quickly climbed onto best seller lists. He began to raise huge amounts of money and to show up at campaign events outside Illinois, speaking on behalf of Democratic candidates facing close contests. As Obama’s star was rising, the Republican Party in Illinois was imploding. Jack Ryan had handily won the Republican nomination for the Senate seat coveted by Obama. (The Republican incumbent, Peter Fitzgerald,

27

28

DIVIDED BY RACE—AND BY GENDER

had decided against running for reelection.) At first glance, Ryan seemed a formidable adversary. He was young, handsome, Ivy-educated, rich, and an articulate and forceful advocate of economic conservatism. He was, however, divorced—from a Hollywood actress. The papers from this legal proceeding were sealed, but when, in the heat of the campaign, Ryan was forced to release them, they proved devastating to his chances. The divorce was not exactly amicable. Allegations of coerced visits to sex clubs in Paris and New York, as well as references to whips, cages, and “other apparatus” gave the press a field day. Within the week, Ryan was gone. The Republican Party was unable to find a replacement for Ryan until August, turning finally—perhaps in desperation—to Alan Keyes. Keyes was an unlikely choice: extremely conservative, highly articulate and contentious, an unsuccessful two-time candidate for the presidency, African American, and, bizarrely, a resident of Maryland. Under these exceedingly congenial conditions, Obama won an overwhelming victory in November. He took 70 percent of the vote, the most one-sided Senate election in Illinois history. Obama arrived in Washington as his party’s rising star, and he arrived with a plan—“The Plan,” as the Obama inner circle called it.4 The purpose of the plan was to position Obama for a possible vice presidential bid in 2008. At the time, this seemed wildly ambitious—after all, just a few months earlier, Obama had been toiling in obscurity in the Illinois state legislature. As things turned out, of course, it was not ambitious enough. Obama and his advisers were actually aiming too low. Following the plan, Obama worked hard and kept out of the limelight, hoping to demonstrate to his colleagues that he could do more than give a rousing speech. As a member of the Foreign Affairs Committee, he traveled widely, visiting Russia, Eastern Europe, the Middle East, and Africa. The trip to Africa in August 2006 ended in Kenya, on a farm near Lake Victoria where Obama’s father had been born and raised. Obama was accompanied by scores of reporters—American, African, and European— and attracted huge and passionate crowds. David Mendell, a Chicago Tribune reporter who traveled with Obama and was nearly trampled for his troubles, called the six days spent in Kenya “utter madness.” Within weeks of his return from Africa, Obama began to think seriously about running for president. His press was intoxicating. Flattering

DIVIDED BY RACE—AND BY GENDER

profiles appeared in national magazines. His name was mentioned as presidential material by influential columnists. Obama’s second book, The Audacity of Hope, was published in October to good reviews and eye-popping sales, eventually dislodging John Grisham from the top of the New York Times’ best seller list.5 And while the news could hardly have been better for Obama, it could scarcely have been worse for Republicans. In the November 2006 midterm elections, the Republican Party suffered a staggering defeat. The Democrats regained control of both houses of Congress for the first time in a dozen years, gaining six seats in the Senate and thirty in the House. President George W. Bush’s policy in Iraq appeared to be failing, and his popularity was plummeting. Obama, just two years into his first term as a U.S. senator, began to think that the Democrats could take the presidency in 2008 and that he could be the one to do it. Obama announced his decision to enter the race in early February 2007 in Springfield, where he had served eight years in the Illinois state senate. The day was frigid, even for midwinter on the Illinois prairie, but Obama was determined to make his announcement outside, where he could stand in front of the Greek Revival Old State Capitol and speak, he hoped, to a hardy band of supporters. When Obama and his family walked out into the cold, an enormous cheer went up. In a sign of things to come, a crowd of more than fifteen thousand had gathered on the capitol grounds. People had come, many from hundreds of miles away, attracted by the possibility that history was in the making. Obama told the crowd that he was running for president to transform the country. Change was urgently needed. The “time is now,” he said, “to shake off our slumber and slough off our fears and make good on the debt we owe past and future generations.” Obama acknowledged the improbability of his quest, but he drew strength from history. In the face of overwhelming challenges, Americans, acting together, have often achieved the improbable: In the face of tyranny, a band of patriots brought an empire to its knees. In the face of secession, we unified a nation and set the captives free. In the face of Depression, we put people back to work and lifted millions out of poverty. We welcomed immigrants to our shores, we opened railroads to the West, we landed a man on

29

30

DIVIDED BY RACE—AND BY GENDER

the moon, and we heard King’s call to let justice roll down like water, and righteousness like a mighty stream. Each and every time, a new generation has risen up and done what’s needed to be done. Today we are called once more—and it is time for our generation to answer that call. For that is our unyielding faith—that in the face of impossible odds, people who love their country can change it. When Obama finished, a huge roar came from across the grounds. Cameras flashed and clicked. Parents hoisted up their children so that they might catch a glimpse of the man who had just declared his intention to become the first black president of the United States.6 Press coverage of the announcement acknowledged the historic significance of Obama’s aspirations, but was quick to point out the long odds he faced. In the next day’s New York Times, Adam Nagourney and Jeff Zeleny wrote: For all the excitement on display, Mr. Obama’s speech also marked the start of a tough new phase in what until now has been a charmed introduction to national politics. Democrats and Mr. Obama’s aides said they were girding for questions about his experience in national politics, his command of policy, a past that has gone largely unexamined by rivals and the news media, and a public persona defined more by his biography and charisma than by how he would seek to use the powers of the presidency. The formal entry to the race framed a challenge that would seem daunting to even the most talented politician: whether Mr. Obama, with all his strengths and limitations, can win in a field dominated by Senator Hillary Clinton, who brings years of experience in presidential politics, a command of policy and political history, and an extraordinarily battled-tested network of fundraisers and advisers.7 Senator Clinton was indeed the presumptive nominee. She enjoyed the backing of her party, an experienced and professional staff, endorsements from prominent African Americans, money to burn (or so she thought), and widespread public support. By the fall of 2007, with the first nomina-

DIVIDED BY RACE—AND BY GENDER

tion contests just coming into view, Clinton possessed what appeared to be a commanding lead over her rivals. David Plouffe, Obama’s principal political strategist, thought the odds so favored Clinton that he doubted Obama would run. Even looking back on the race a year later, Plouffe said, “If you ran the Democratic primaries a hundred times, Clinton probably wins ninety-five times.”8 But as we know, in a tight and fiercely contested race, it was Obama who eventually emerged victorious. The business of this chapter is to scrutinize Obama’s surprising nomination victory. In doing so, we pay particular attention to race. We show how race helped Obama among African Americans and hurt Obama among white voters. Throughout, race is our principal concern. But we also pay attention to gender. In the 2008 contest for the Democratic Party’s presidential nomination, Hillary Clinton made gender salient in the same way that Obama made race salient—by embodying it. Gender, like race, is a durable inequality. Indeed, it could be argued that gender is the most durable inequality of all, present at the beginning of human society and present still today. As we argue in chapter 1, social groups become relevant to politics insofar as they are sites for persistent inequality. Race fulfills this condition well, but so does gender. Senator Clinton’s prominence in the Democratic nomination contest affords us the opportunity to expand our analysis, to compare race with gender, treating both as potentially significant short-term electoral forces. To set the stage for this comparison, we begin by recounting the story of the nomination race, from the opening contests in January to Clinton’s concession in June. Then we discuss the place of race and gender in the campaign. Both candidates did their best to talk past race and gender, and for the most part—with the conspicuous exception of Pastor Jeremiah Wright’s sudden, theatrical, and discombobulating entrance onto the national scene in March—both candidates succeeded. In the heart of the chapter, we turn to voters and the role played by their beliefs and feelings about race and gender in the 2008 nomination contest.

The 2008 Democratic Nomination Campaign The current presidential nomination system owes much to the Democratic Party’s tremendously unsatisfying experience of 1968, which

31

32

DIVIDED BY RACE—AND BY GENDER

featured President Lyndon B. Johnson’s surprise withdrawal, a police riot at the Democratic National Convention, the nomination of a candidate— Hubert Humphrey—who had entered not a single primary contest, and the loss of the White House in November. Reform elements within the Democratic Party succeeded in changing the rules, and the Republicans quickly followed suit. These alterations—perhaps “the most extensive planned changes in the entire history of American parties” (Shafer 1983, p. 524)—ensured that nominees would be selected less by party leaders and more by the party’s rank and file, who would be offered the opportunity to cast their votes in a sequence of state-based primaries and caucuses. Under the new system, results from these separate contests determine the number of pledged delegates committed to vote for each candidate at the party’s national convention. Delegates are awarded to candidates proportional to their success in each state’s primary or caucus. The Democratic Party’s nomination process typically begins with an open caucus in Iowa in January, proceeds briskly to New Hampshire for the first primary, and then meanders through all the states and territories, not reaching completion until June. As the 2008 version of this process was about to kick off, Hillary Clinton was the clear front-runner on the Democratic side. In addition to Clinton and Obama, the field included Joe Biden, Chris Dodd, John Edwards, Mike Gravel, Dennis Kucinich, Bill Richardson, and Tom Vilsack. Of these, only Clinton, Obama, and Edwards were widely known, and only these three had managed to amass the kind of money required to mount a serious campaign.9 Among the three front-running candidates, Clinton enjoyed a comfortable lead in the national polls. In December 2007, before Iowa and New Hampshire, Clinton was the choice of 41.1 percent of Democrats nationwide, followed by Obama with 28.4 percent and Edwards with 20.1 percent. Others in the field trailed far behind.10 Clinton also led among “superdelegates.” Superdelegates are drawn from the party elite: members of the U.S. House of Representatives and Senate, state and territorial governors, members of the Democratic National Committee, and notable party leaders. Altogether, superdelegates comprised nearly 20 percent of the total delegate count in 2008. While officially uncommitted until the convention, superdelegates are permitted

DIVIDED BY RACE—AND BY GENDER

to endorse a candidate before then. Many chose to do so, generally lining up behind Senator Clinton. On the eve of the Iowa caucus, Clinton led Obama among committed superdelegates by a three-to-one margin.11 The picture was not entirely rosy for the Clinton campaign, however. Nationwide, Democrats gave Senator Clinton favorable ratings, but they liked Senator Obama nearly as much. In ideological terms, Democrats saw Clinton and Obama as essentially indistinguishable: no indication here of Obama as the radical outsider. And while Democrats were impressed by Clinton’s experience, they regarded Obama as the more trustworthy.12 Moreover, there were signs that Clinton was in trouble in Iowa, the first of the 2008 nomination contests. In a private campaign memorandum prepared in May 2007, Mike Henry, then Senator Clinton’s deputy campaign manager, recommended that the Clinton campaign pull out of Iowa completely. According to Henry, Iowa was Clinton’s “consistently weakest state,” and campaigning there was expensive. Clinton would be better served by spending time and money elsewhere, especially in the large Super Tuesday states that would vote on February 5. Henry’s memorandum was leaked to the New York Times. When asked for comment, the Clinton campaign said that it had considered and then rejected Mr. Henry’s advice. Howard Wolfson, the campaign’s communications director, claimed that Senator Clinton had not even seen the memorandum. Senator Clinton herself said that the Henry memorandum did not represent the campaign and that it was not her view. To prove her point, she then hurried off to spend the next few days campaigning in Iowa.13 Henry may not have had the right remedy, but he was surely right in identifying a problem. According to a CBS/New York Times poll carried out in the first week of November 2007, Clinton was leading in Iowa, but her margin was paper-thin. Twenty-five percent of Iowans said they intended to vote for Clinton; 22 percent said they were going to vote for Obama; and 23 percent declared their support for Edwards. Iowa was up for grabs.14 Clinton’s lead was even more precarious than these simple percentages suggest. When asked for their second choice (on the possibility that their first pick might not gather sufficient support in the caucus system), Iowa Democrats were more likely to name Obama (24 percent) than Clinton

33

34

DIVIDED BY RACE—AND BY GENDER

(16 percent). They also found Obama more likable than Clinton. And while they were much more likely to see Clinton as prepared for the job (80 percent versus 42 percent), they were just as one-sided in the opposite direction in thinking that Obama was more likely to say what he really believed (82 percent as against 47 percent). The closer one looked, the more fragile Clinton’s lead seemed. Mike Henry was not the only campaign insider to think Clinton could lose Iowa. David Plouffe, Obama’s principal political strategist and director of field operations, thought so, too. In previous presidential elections, Plouffe had worked for Senator Tom Harkin and for Representative Richard Gephardt, devising strategies to challenge front-runners, and he knew every inch of Iowa. Plouffe knew that none of Senator Clinton’s core advisers had any experience in the state. He knew that Bill Clinton had never campaigned seriously in Iowa, so there was no residual Clinton organization for Hillary to reactivate. He knew that Iowans generally opposed the Iraq War, and thought that Senator Clinton’s vote in favor of authorizing U.S. military intervention could hurt her. As Plouffe saw it, the way for Obama to win the nomination was “to destabilize Hillary early.” Absent an Obama victory in Iowa (or New Hampshire), Plouffe said, Clinton “would be off to the races.”15 The Obama team put all its chips on Iowa. The campaign spent some $9 million on television advertisements alone, and another $7 million on grassroots organizing. By the end of 2007, 160 paid field organizers were operating in the state, supported by a small army of volunteers. The campaign staged mock caucuses to train staff and prospective voters on how the caucus process works (it is not straightforward). As the actual caucus neared, the organization was making between 25,000 and 30,000 telephone calls to potential supporters every day. The effort here was aimed partly at expanding the electorate—David Axelrod called this the “Iowa Dream”: to mobilize the young, to find pockets of racial and ethnic minorities and bring them, perhaps for the first time, into the caucus proceedings. As for the candidate himself, Obama spent almost one-quarter of 2007—eighty-nine days in all—in Iowa and visited ninety of the state’s ninety-nine counties.16 All this effort paid handsome dividends. On January 3, Obama won 37.6 percent of the Iowa vote, as against 29.8 percent for Edwards and 29.5

DIVIDED BY RACE—AND BY GENDER

percent for Clinton. The margin of victory was narrow and only a handful of delegates were at stake, but Obama’s win was trumpeted as historic. According to the New York Times, he had “rolled to victory” in Iowa, “lifted by a record turnout of voters who embraced his promise of change.” Obama was showered with money and attention. Diane Sawyer and Brian Williams chased after him for prime-time interviews. Flattering photographs of the senator filled the covers of the national news magazines. In newspapers, on television, and no doubt in everyday conversations as well, Obama was now a serious contender, if not the actual front-runner. Almost as important as Obama winning was Clinton losing. The press wrote this part of the story just as melodramatically. Senator Clinton had suffered a “startling setback,” as the New York Times put it. The Times went on to say—as if there could be a question about this—that despite her devastating loss, Clinton “vowed to stay in the race.” Reporters sniffed panic among Clinton supporters. Rumors of a staff shake-up began to circulate. The Iowa victory did not guarantee Obama the nomination. Eulogies on the Clinton campaign composed in the immediate aftermath of Iowa were rescinded just a few days later, when Clinton won a narrow victory over Obama in New Hampshire (Clinton and Obama actually tied in the delegate count). Iowa did level the playing field, however. In losing, Clinton no longer seemed the inevitable nominee. Now there was a race.17 In short order, moreover, it was a two-person race. A week and a half after New Hampshire, while Clinton and Obama were dividing the delegates in the Nevada caucus, John Edwards’s support collapsed. Sensing that voters were coalescing around the two front-runners, Edwards suspended his campaign ten days later. For all intents and purposes, the Democratic contest was down to two. The race turned next to South Carolina, the first contest to feature large numbers of African American voters. Trailing badly in preelection polls, Clinton decided to spend her time in several Super Tuesday states, leaving her husband, the former president, to fly the Clinton flag in South Carolina. Bill Clinton did his wife no favors (not for the first time), complaining publicly about the Obama campaign in a series of statements that CBS News labeled “inflammatory and negative,” including the suggestion that, like Jesse Jackson’s victory in the South Carolina primary in 1988,

35

36

DIVIDED BY RACE—AND BY GENDER

Obama’s likely win in 2008 should be dismissed as merely symbolic. Symbolic or not, Obama defeated Senator Clinton in South Carolina decisively, winning 55 percent of the vote to her 27 percent. Two days later, Obama received the endorsement of Senator Edward Kennedy. At a rally at American University in Washington, D.C., Kennedy likened Obama’s candidacy to his brother’s successful run for the presidency in 1960, and he suggested, gently but pointedly, that the Clintons should stand down: There was another time, when another young candidate was running for president and challenging America to cross a New Frontier. He faced public criticism from the preceding Democratic president, who was widely respected in the party. Harry Truman said we needed “someone with greater experience. May I urge you to be patient?” And John Kennedy replied: “The world is changing. The old ways will not do. It’s time for a new generation of leadership.” So it is with Barack Obama. He has lit a spark of hope amid the fierce urgency of now.18 In December, Clinton had led Obama among Democrats nationwide by some 20 percentage points. At the end of January, after South Carolina and on the eve of Super Tuesday, the lead was gone.19 On February 5, contests were held in 23 states and territories, with nearly 1,700 delegates at stake. When the dust settled, both Obama and Clinton could claim victory. Obama won 13 states and territories, including Connecticut and Missouri. Clinton won 10, including the delegaterich prizes of California and Massachusetts. With all the votes counted, Obama had amassed 847 delegates to Clinton’s 834. In the race for the nomination, Obama now held a slight lead (see figure 2.1). The day after Super Tuesday, Senator Clinton announced that she had personally loaned her campaign $5 million in January. The news came as a surprise and set off another round of stories that featured troubled Clinton supporters concerned about the campaign. The news was particularly striking juxtaposed against Obama’s announcement that his campaign had raised an astonishing $32 million in January. The bad news for Clinton continued. Over the next two weeks, Obama racked up eleven consecutive victories in caucuses and primaries alike.

DIVIDED BY RACE—AND BY GENDER

2000 1800

Obama

1600 1400

Clinton

1200 Pledged 1000 Delegates 800 600 400 200 0

2008 Primary Election

Figure 2.1 Delegates Won: 2008 Democratic Nomination Primaries and Caucuses Source: Cable News Network: http://www.cnn.com/ELECTION/2008/primaries/ results/scorecard/.

Obama prevailed in Louisiana, Washington, Maine, Maryland, Virginia, Wisconsin, and more. The press interpreted these as “staggering” losses for Clinton and described her campaign as demoralized and in disarray. Facing oblivion, Clinton launched a direct attack on Obama’s qualifications. A Clinton television advertisement questioned whether Obama would be prepared to handle a crisis as commander in chief. Clinton scored a convincing win in Ohio and a narrow victory in the Texas primary. She framed her wins that night as a comeback, but the delegate count continued to favor Obama. Obama actually won the total delegate race in Texas (the Texas contest was part primary and part caucus), stayed close to Clinton in Ohio, and won contests in Wyoming and Mississippi the following week. In comparison to Clinton’s deficit, the number of pledged delegates still to be won was small, and with delegates allocated by proportional representation, overwhelming victories in the delegate count were virtually impossible. At this stage of the contest, Senator Clinton had only two moves left. One was to insist that the Michigan and Florida delegations be seated at the convention. In August 2006, the Democratic National Committee (DNC) had adopted a proposal that only four states—Iowa, New Hampshire, Nevada, and South Carolina—would be permitted to hold presidential

37

38

DIVIDED BY RACE—AND BY GENDER

primaries or caucuses before February 5, 2008. Arguing that this unfairly assigned too much influence to small and unrepresentative states, both the Florida and Michigan legislatures voted to move their primaries up to take place before the permitted date. In response, the DNC stripped Florida and Michigan of their delegates. Both primaries went ahead as scheduled— that is to say, early—though, by mutual agreement, neither Clinton nor Obama spent any time or money in either state. (Obama’s name did not even appear on the Michigan ballot.) As her prospects dimmed, Clinton began to argue for reinstating the Michigan and Florida results (she had “won” both). When that argument failed, she suggested that the two states should be permitted to vote again. This recommendation, too, went nowhere.20 Clinton’s final option was to win over the superdelegates. She tried to do so by raising doubts about Obama’s electability in the fall and by arguing that she would make the stronger candidate against John McCain and the Republicans. This maneuver also failed. Party leaders, including House Speaker Nancy Pelosi, implied that the party would court disaster in November should superdelegates overturn the judgment made by millions of rank-and-file Democrats. Senator Clinton kept fighting. She scored a convincing win in Pennsylvania, eked out a victory in Indiana, and took West Virginia and Kentucky by large margins. But she lost to Obama in North Carolina and again in Oregon. On June 3, the day of the final primary contests, the Obama campaign rolled out commitments from nearly sixty superdelegates. Senator Clinton suspended her campaign, and on June 7, in a gracious concession speech, endorsed Obama for president.

Race and Gender in the 2008 Nomination Contest We have so far said little about the place of race and gender in the 2008 nomination contest. In some ways, this is an accurate reflection of what, in fact, transpired. For the most part, the campaign studiously avoided both subjects. Clinton and Obama faced similar dilemmas. Senator Clinton could hope for increased support among women, but she could not risk being seen as a “women’s” candidate. Clinton’s campaign was organized around

DIVIDED BY RACE—AND BY GENDER

presenting her as an experienced, strong leader, ready to be commander in chief. Mark Penn, Clinton’s chief strategist, looked to Margaret Thatcher, the former British prime minister, as offering the right model. In making this recommendation, Penn argued that Thatcher had succeeded in national politics not because she came across as warm but because she came across as tough. To win the nomination and win again in the fall, Clinton had to appeal to men as well as women, and perhaps especially to men. Likewise, although Obama could count on the lion’s share of the black vote, to be thought of as the black candidate would sink his chances for the nomination and ruin his prospects in November. Obama had to win support from whites as well as blacks. Because the campaign could not avoid race—Obama embodied blackness—it attempted to neutralize it. By and large, Obama did not talk about race, did not go out of his way to seek the endorsement of prominent black leaders, spent most of his time in front of white audiences, and avoided the rhetoric of racial grievance and compensation. The “we” in “Yes we can” referred to the American people, united by common human hopes and love of country. “We are,” as Obama said time and again, “one people.”21 And so, throughout the campaign, race “was the thing always present, the thing so rarely mentioned” (Remnick 2008, p. 83). Much the same could be said about gender. There were exceptions, of course—moments when race and gender erupted into public view, when race and gender became explicit topics of conversation. One such moment occurred in New Hampshire. In a debate held just before the primary, John Edwards appeared to join sides with Obama in his criticism of Clinton. This was read in some circles as bullying. A few days later, Clinton grew teary-eyed as she talked about how hard it was to keep going in the face of criticism and defeat. These episodes drew intense scrutiny, and they were framed in terms of gender. As for race, around the time of the South Carolina primary, Hillary Clinton appeared to slight Martin Luther King’s contribution to the Voting Rights Act of 1965, saying that it took a president to make Dr. King’s dream real. Meanwhile, her husband, enraged over what he took to be Obama’s free ride with the press, ridiculed Obama’s opposition to the Iraq War as a “fairy tale” and dismissed Obama’s victory in South Carolina as mere racial symbolism. These comments kicked up a fuss. They were read,

39

40

DIVIDED BY RACE—AND BY GENDER

in some quarters, as racially inflammatory. But like their gender counterparts, the Clintons’ comments amounted to minor and momentary disturbances. Race and gender were quickly hustled offstage. The campaign moved on.22 This leaves one disturbance unaccounted for, which was neither minor nor momentary. At one utterly unmanaged moment, race came crashing onto center stage. Reverend Jeremiah Wright’s sudden appearance in the campaign presented the Obama candidacy with a potentially fatal threat. Obama first met Wright in the late 1980s, while he was working as a community organizer in Chicago after college and before law school. Wright was then senior pastor of Chicago’s Trinity United Church of Christ. He officiated at the wedding ceremony of Barack and Michelle Obama, he presided over their children’s baptisms, and he played an important role in Obama’s journey to Christianity. The title of Obama’s 2006 book, The Audacity of Hope, was inspired by one of Wright’s sermons.23 In the early spring of 2008, shortly after the Ohio and Texas primaries, with Obama and Senator Clinton still locked in a close fight for the Democratic nomination, excerpts of several of Wright’s sermons began circulating on YouTube. ABC News picked up the story. In a sermon delivered shortly after 9/11, Wright was seen arguing that the terrorist attacks were payback for American atrocities: genocide against Native Americans, the bombing of Hiroshima and Nagasaki, and support for apartheid in South Africa. In another sermon, which also received widespread play, Wright denounced the government for its treatment of African Americans: When it came to treating her citizens of African descent fairly, America failed. She put them in chains, the government put them on slave quarters, put them on auction blocks, put them in cotton fields, put them in inferior schools, put them in substandard housing, put them in scientific experiments, put them in the lowest paying jobs, put them outside the equal protection of the law, kept them out of their racist bastions of higher education and locked them into positions of hopelessness and helplessness. The government gives them the drugs, builds bigger prisons, passes a three-strike law and then wants us to sing “God Bless America.”

DIVIDED BY RACE—AND BY GENDER

No, no, no, not God Bless America. God damn America—that’s in the Bible—for killing innocent people. God damn America, for treating our citizens as less than human. Many white Americans found Wright’s sermons infuriating. They were delivered with strut, swagger, and great theatricality, and they seemed to repudiate a central theme of the Obama campaign: that as president, Obama would rise above partisan rancor and bring the country together. Obama first tried to distance himself from his pastor and former confidant. He referred to Wright’s statements as “inflammatory and appalling.” This did little to quiet the frenzy. Obama recognized that his relationship with Pastor Wright was threatening to take over his campaign and consume his candidacy. To confront the issue directly, Obama decided to deliver a speech to the nation. In David Axelrod’s estimation, Obama’s address, delivered from the National Constitution Center in Philadelphia, was “probably the most important moment of the whole campaign.”24 Obama began by pointing out the stark contradiction between the ideals of justice and equality expressed in the Constitution, on the one hand, and the national disgrace of slavery, on the other. He framed his campaign for president as the continuation of the long struggle to narrow the gap between American ideals and American reality. And he offered his campaign’s success as a sign of Americans’ hunger for national harmony. He then turned to Jeremiah Wright, referring to him as “my former pastor.” Obama noted that there was more to Wright than what was showing up in snippets of his sermons. Wright had served his country as a U.S. marine and for more than thirty years had led a church locally famous for its good works: “housing the homeless, ministering to the needy, providing day-care ser vices and scholarships and prison ministries, and reaching out to those suffering from HIV/AIDS.” There was goodness in Wright, Obama said, but such goodness could not excuse his “incendiary language.” His sermons expressed “a profoundly distorted view of this country—a view that sees white racism as endemic and that elevates what is wrong about America above all that is right about America.” In Obama’s judgment, Wright’s words served to widen the racial divide, denigrate the greatness and the goodness of the nation, and offend white and black alike.

41

42

DIVIDED BY RACE—AND BY GENDER

Still, Obama was not prepared to abandon Wright: I can no more disown him than I can disown the black community. I can no more disown him than I can my white grandmother—a woman who helped raise me, a woman who sacrificed again and again for me, a woman who loves me as much as she loves anything in this world, but a woman who once confessed her fear of black men who passed by her on the street, and who on more than one occasion has uttered racial or ethnic stereotypes that made me cringe. These people are a part of me. And they are a part of America, this country that I love. Obama said that to dismiss Wright as a mere “crank or a demagogue” would be a mistake, for there was something real and important in the anger Wright displayed, even if he displayed it crudely. Obama reminded his audience that Wright had come of age at a time when segregation was the law of the land and discrimination was blatant: For the men and women of Reverend Wright’s generation, the memories of humiliation and doubt and fear have not gone away, nor has the anger and bitterness of those years. That anger may not get expressed in public, in front of white co-workers or white friends. But it does find voice in the barbershop or the beauty shop or around the kitchen table. At times, that anger is exploited by politicians, to gin up votes along racial lines or to make up for a politician’s own failings. And occasionally it finds voice in the church on Sunday morning, in the pulpit and in the pews. The fact that so many people are surprised to hear that anger in some of Reverend Wright’s sermons simply reminds us of the old truism that the most segregated hour of American life occurs on Sunday morning. Obama went on to acknowledge that there was real and legitimate anger in the white community as well. When whites are ordered to bus their children to a school across town, when they hear that an African American has landed a good job or a spot in an elite college, and when they’re

DIVIDED BY RACE—AND BY GENDER

told that their fear about crime in city neighborhoods is nothing more than prejudice, anger pools and deepens. Such anger may often prove counterproductive, yet “to wish away the resentments of white Americans, to label them as misguided or even racist, without recognizing they are grounded in legitimate concerns—this too widens the racial divide and blocks the path to understanding.” Obama called upon African Americans to bind their particular grievances to the larger aspirations of all Americans and to take responsibility for their own lives.25 He called upon whites to acknowledge the legacy and persistence of discrimination as a real presence in American life. Working together, progress on race was possible. As Obama saw it, contrary to Reverend Wright’s sermons, the nation need not be “irrevocably bound to a tragic past.” Wright’s dramatic entrance into the campaign and Obama’s speech took place during an otherwise quiet interlude: a week and a half after the Ohio and Texas contests (March 4), and a month before the Pennsylvania primary (April 22). This is convenient for determining whether the episode damaged Obama’s candidacy. Remarkably enough, it seemed not to. Figure 2.2 presents results from a sequence of national surveys of Democrats conducted by Gallup. The surveys begin early in March, shortly after the Ohio and Texas primaries, and run through the middle of April, ending a week before the Pennsylvania contest. As the figure shows, after Ohio and Texas and before Reverend Wright’s sudden appearance onto the national scene, Obama held a small lead over Clinton of about 3 percentage points. At the peak of the furor over Wright and just before the nationwide speech on race, Obama abruptly fell behind. This turned out to be a momentary setback, however. Within a day or two after his speech and well before the Pennsylvania primary, Obama had regained the advantage. The fierce competition for delegates between Obama and Clinton resumed almost as if nothing had happened.26 This is surprising. This is really surprising. At the time, many analysts said that Pastor Wright’s sudden celebrity could destroy Obama’s candidacy. Why didn’t it? The answer, we think, lies in the nature of the relationship between ordinary citizens and political elites. As we say in chapter 1, citizens have

43

44

DIVIDED BY RACE—AND BY GENDER

55 Obama 50 45 40 Percent Support

35 30

Clinton Ohio, Texas primaries

Pastor Obama's Wright speech story breaks

25 20 15

2008 Primary Election

Figure 2.2 Clinton versus Obama before and after Reverend Wright Source: Gallup via http://www.pollster.com/polls/us/08-us-dem-pres-primary.php.

limited interest in politics and they reach judgments about public life quickly and rather casually. Politics—the “mystery off there,” as Walter Lippmann (1925, p. 24) once referred to it—is complicated, ridden with uncertainty, and ambiguous in meaning. Insofar as ordinary citizens keep up, they do so by depending on others; more precisely, by relying on the interpretations offered by political elites. This claim is at the center of John Zaller’s (1992) influential theory of opinion formation and change. In Zaller’s account, the engine that drives public opinion is “elite discourse.” Public opinion moves in response to what elites say: the endorsements they offer, the arguments they spin, and the stereotypes and frames they advance. In Zaller’s telling, such cues “enable citizens to form conceptions of and, more importantly, opinions about events that are beyond their full understanding” (p. 14).27 Zaller’s theory directs our attention not to Obama’s speech itself, but rather to commentary on the speech by elites. Millions of Americans watched as Obama delivered his remarks on race at the National Constitution Center in Philadelphia. More important, from Zaller’s point of view, immediately afterward and over the next several days, millions more took notice of what elites had to say about the speech. For several days, it was

DIVIDED BY RACE—AND BY GENDER

the news story. Reporters, editors, anchors, pundits, and others wielding large megaphones all had their say. When citizens paused from their daily routines to glance up at this discussion, they saw something remarkable. Opinion leaders of every persuasion—Democratic and Republican, liberal and conservative— applauded Obama’s speech. It was endorsed in glowing terms by the New York Times and the New York Review of Books.28 No surprise there. But Obama’s speech received accolades from unexpected quarters as well: John McCain, Secretary of State Condoleezza Rice, Newsweek, New York City mayor Michael Bloomberg, General Colin Powell, and many others. Writing in the Wall Street Journal, Peggy Noonan called the speech “strong, thoughtful and important.” Charles Murray, coauthor of The Bell Curve, wrote that the speech was “just plain flat out brilliant—rhetorically, but also in capturing a lot of nuance about race in America.” In an address before the American Enterprise Institute, former Republican Speaker of the House Newt Gingrich declared that Obama “gave us a very courageous speech.”29 Across the political spectrum, from predictable and unpredictable sources alike, Obama’s speech was greeted with bouquets of praise. This is important since we know that the reactions of ordinary people to dramatic political events are influenced, often decisively, by elite interpretation. Because elite reaction to Obama’s address was so lopsidedly positive, what could have been a catastrophe for Obama turned out to be little more than a momentary setback. Obama delivered quite a speech. Perhaps it saved his candidacy.30 The campaign moved on and Obama eventually prevailed, thereby becoming the Democratic Party’s candidate for president in 2008 and the first African American to win a major party’s presidential nomination. This was a remarkable thing—astonishing, from certain points of view—and let’s keep in mind, a very close thing. During the nomination contest, some 36 million Democrats voted either for Clinton or Obama. If we include the Florida and Michigan votes, and if the Michigan votes marked “uncommitted” are allocated entirely to Obama (Obama had removed his name from the ballot after Michigan violated party rules by jumping the queue), then Obama wins the popular vote by a hair: 50.09 percent to 49.91 percent. If, instead, we set the Michigan uncommitted voters aside, then Clinton takes the popular vote, also by a tiny margin: 50.25 percent to 49.74 percent.31

45

46

DIVIDED BY RACE—AND BY GENDER

The score that really counted, of course, was measured in delegates.32 This race was close, too, and it was close from beginning to end (see figure 2.1). Altogether, Obama needed the support of 2,117 delegates to win the nomination; he received 2,307. Clinton fell just short of the required mark, with 1,972. In an exceedingly close race, Obama became the nominee.

Race, Gender, and the 2008 Democratic Nomination Vote Presidential primaries and caucuses are intrasquad contests; they take place within parties. This means that partisanship—whether voters identify as Democrats or Republicans—is largely nullified as a basis for choice.33 With party removed, other predispositions come more prominently into play. Among such predispositions operating as short-term forces in 2008, we suspect, are race and gender. At first glance, both certainly seemed important. Obama and Clinton split the primary and caucus vote almost precisely in half. But Obama did much better in 2008 among black voters than he did among white voters, while Hillary Clinton fared better among women than among men. Obama took 76.6 percent of the black vote and 47.9 percent of the white vote. Clinton won 52 percent of the vote among women and 40 percent of the vote among men.34 In-Group Solidarity As we argue in chapter 1, one way that social cleavages like race and gender become important in politics is through the activation of in-group solidarity. In the case of 2008, we are interested in determining the extent to which feelings of racial solidarity lay behind blacks’ support for Barack Obama, the extent to which feelings of gender solidarity lay behind women’s support for Hillary Clinton, and whether race played a more important role in this respect for Obama than gender did for Clinton. Based on previous research, we might expect race to be the more important of the two. Racial group identity tends to play a larger role in politics than gender group identity. But there were reasons to think the results might go the other way in 2008.35

DIVIDED BY RACE—AND BY GENDER

Obama’s rise to prominence revitalized a long-standing debate within the African American community. Was Obama black enough? Was he black at all? Would he—could he—remain loyal to his race? Such questions fi rst surfaced in Obama’s ill-fated 2000 campaign for Congress. Restless, pushing forty, and worried about his family’s future, Obama decided to challenge Bobby Rush for a seat in the U.S. House of Representatives. Rush was elected to Congress in 1992 and had represented the South Side of Chicago ever since. As a participant in the civil rights movement, a former member of the Black Panther Party, and an alderman under Mayor Harold Washington, Rush would seem to have been safely ensconced in his largely African American district. But to Obama, Rush looked vulnerable. According to David Mendell, who covered Obama during his years in Chicago, Obama “saw Rush as an aging politician ready to be replaced by a younger man with a fresh vision and new enthusiasm for tackling the ills of the black community” (2007, p. 128). In his willingness to take Rush on, Obama struck many as brash and ungrateful. Many black Americans “wondered what Obama was doing trying to unseat this black elder statesman; and some African Americans privately began to question not only Obama’s motives but his black credentials” (Mendell 2007, p. 129). Obama had strong ties to powerful white institutions—Harvard and Columbia, where he had gone to school; the University of Chicago, where he was teaching; and the Chicago Tribune, which endorsed him, calling Obama “a rising star in the Democratic Party.” None of this did Obama much good on the South Side. Allegations began to spread that Obama was the tool of North Shore liberals or a Hyde Park “mafia.” Donne Trotter, a black state senator, said that many in the black community viewed Obama as “the white man in blackface.” Rush dismissed Obama as “a person who read about the civil rights protests and thinks he knows all about it.”36 Obama was wrong about Rush being ready for replacement. Or maybe Obama was wrong about himself being ready to do the replacing. Either way, Rush delivered a beating. Obama lost by 31 percentage points—a “drubbing,” Obama called it. Mortified and chastened, he returned to Springfield and the Illinois senate to fight another day. Obama’s subsequent forays into electoral politics have turned out rather more successfully, as we know, but his racial authenticity has continued to

47

48

DIVIDED BY RACE—AND BY GENDER

be questioned. One reason for this is ancestry. Obama’s mother was white—“white as milk,” as Obama wrote in Dreams from My Father—and this was enough, in some eyes, to render him inauthentic. His father was black, but a Kenyan national who came to the United States voluntarily— indeed, eagerly. Without a deep, personal connection to slavery, so goes the argument, Obama could not fully appreciate the role played by race in American life.37 In addition, Obama was raised in Hawaii and Indonesia and attended elite schools in the United States. He experienced segregation and discrimination, but he did not come from the ghetto. He did not live the life. And finally, throughout his political career (starting, one might say, at Harvard Law School), whites have always been among Obama’s strongest supporters and closest advisers. For all these reasons, then, perhaps not all black Americans would see Obama as a full-fledged member of their racial group. If this were so, then racial group identification among African Americans might turn out to be less important to Obama than gender group identification among women would be to Clinton.38 To find out, we analyzed survey data supplied by the Cooperative Campaign Analysis Project (CCAP), a national panel study of registered voters, conducted online in six waves, commencing in December 2007 and concluding in November 2008, immediately after the presidential election.39 The questions we placed on the CCAP included standard instrumentation to assess group identification. Women were asked how often they felt pride over the accomplishments of women, and in a separate question, how often they felt angry about the way women were treated. African Americans were asked about the accomplishments and treatment of blacks. The exact questions are presented in table 2.1. As shown there, black Americans were much more likely to strongly identify with their race than women were with their gender: 46.9 percent of African Americans reported that they felt pride in the accomplishments of blacks “a lot,” compared to just 23.4 percent of women saying the same about the accomplishments of women. The difference is nearly as great on anger over mistreatment.40 We created two scales of group identification, one for blacks and the other for women. We build composite scales based on multiple indicators

DIVIDED BY RACE—AND BY GENDER

Table 2.1 Group Solidarity in 2008 Women (percent)

African Americans (percent)

How often do you fi nd yourself feeling a sense of pride in the accomplishments of women/blacks? A lot 23.4 46.9 Fairly often 39.4 28.2 Once in a while 28.2 19.5 Hardly ever 9.1 5.4 N 549 107 Women (percent)

African Americans (percent)

How often do you fi nd yourself feeling angry about the way women/blacks are treated in society? A lot 15.4 37.2 Fairly often 16.7 30.2 Once in a while 47.4 31.2 Hardly ever 20.5 1.5 N 543 107 Source: 2007–2008 Cooperative Campaign Analysis Project.

whenever we can. Representing complex concepts like group identification with single indicators is hazardous. Measurement problems do not disappear when answers to multiple questions are combined into scales, but they diminish. Scales provide greater precision, finer discrimination, and better estimates. In this case in particular, we simply averaged responses across the two questions and scored the result to range from 0 (low identification) to 1 (high identification). The empirical question before us, then, was the extent to which the primary vote was driven by feelings of group solidarity. Did Obama do especially well among African Americans with close ties to their racial group? Did Clinton do especially well among women who identify closely with their gender?41 To determine the effect of group identification on vote in the Democratic primary, it is not enough simply to show that scores on the identification

49

50

DIVIDED BY RACE—AND BY GENDER

scales are correlated with the vote. To estimate the effect of group identification, our analysis must take into account the effects due to other factors. We are interested in the independent effect of group identification, holding other relevant considerations constant. For this reason, our analysis included a measure of strength of partisanship, on the idea that Democrats might see Clinton as the party regular and Obama as the outsider. We also included measures of social background: education, age, religion, and region. Such characteristics share two features: they identify social cleavages that often give rise to political differences, and they are (or might plausibly be) correlated with group identification. For example, well-educated African Americans may be more likely to identify with their racial group than African Americans who have less schooling; or young women may be more apt to identify with their gender than their older counterparts. In short, we included strength of partisanship, education, age, religion, and region in our analysis in order to obtain an unbiased estimate of the effect due to group identification (in more technical language, to avoid omitted variable bias). We estimated the effect of group identification in two separate analyses, one for African Americans (racial solidarity) and the other for women (gender solidarity).42 To maximize sample size, we estimated the relationship between group identification and ratings of Obama and Clinton among Democrats, voters and nonvoters alike.43 In both the December and March interviews, CCAP respondents were asked to evaluate Barack Obama and Hillary Clinton on a simple scale, running from very unfavorable to very favorable. For convenience in interpretation, we coded all variables on the 0–1 interval, and we estimated effects by least squares regression.44 The results, presented in table 2.2, reveal a sharp contrast between race and gender. Group solidarity among women has little or nothing to do with support for Hillary Clinton. The effect is positive in December and positive in March, but on both occasions it is small and on neither occasion can we be sure that the effect is not actually zero. In contrast, group solidarity among African Americans plays a major role in generating support for Barack Obama. The effect is positive, significant, and substantial in December; it is positive, significant, and larger still in March.

DIVIDED BY RACE—AND BY GENDER

Table 2.2 The Effect of Group Solidarity on Ratings of Clinton and Obama in 2008 Least Squares Regression (Standard Errors in Parentheses) Ratings of Clinton by Women

Group Solidarity Partisan Strength Race (Black)

December

March

0.08 (0.08) 0.24** (0.08) 0.09** (0.04)

0.10 (0.11) 0.22** (0.10) −0.08 (0.06)

Sex (Female) R2 N

0.09 237

0.04 182

Ratings of Obama by African Americans December

March

0.25* (0.13) −0.08 (0.13)

0.35** (0.13) −0.08 (0.14)

−0.00 (0.06) 0.08 73

−0.06 (0.06) 0.17 60

*p < 0.10 **p < 0.05 ***p > Catholics Protestants > Catholics Protestants = Catholics

6.4 25.2 69.9

Source: 1956 National Election Study.

Coding of Variables for Key Equations Chapter 2 Evaluation of Obama = β0 + β1Racial Group Solidarity + β2Strength of Party Identification + β3Education + β4 Age + β5Sex + ε [2.1] where Evaluation of Obama is a five-point favorability scale, coded to run from 0 (Very Unfavorable) to 1 (Very Favorable). Racial Group Solidarity is coded so that respondents had to answer both questions to be included (see above for more detail). Strength of Party Identification is a four-point scale based on the Democratic side of the seven-point party identification scale, recoded to run from 0 (Independent) to 1 (Strong Democrat). Education is a five-point scale. The variable runs from 0 (No High School) through 1 (Postgraduate). The categories “Some College” and “College 2 Years” were coded into the same category at 0.5. Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Sex is a dummy variable with Male coded as 0 and Female coded as 1.

APPENDIX

Evaluation of Clinton = β0 + β1Gender Group Solidarity + β2Strength of Party Identification + β3Education + β4Catholic + β5South + β6Age +β7Race + ε [2.2] where Evaluation of Clinton is a five-point favorability scale, coded to run from 0 (Very Unfavorable) to 1 (Very Favorable). Gender Group Solidarity is coded so that respondents had to answer both questions to be included (see above for more detail). Race is a dummy variable for blacks and whites, with whites as the omitted reference group. Catholic is a dummy variable for Catholics, with all other religions as the omitted reference group. South is a dummy variable for South, with all other regions as the omitted reference group. Other variables are coded as above. Primary Vote = β0 + β1Racial Resentment + β2Strength of Party Identification + β3Education + β4West + β5South + β6Northeast + β7Age +β8Sex + ε [2.3] where Primary Vote is coded 0 if Clinton and 1 if Obama (votes for other candidates are excluded). Racial Resentment is coded so that respondents had to answer all questions to be included (see above for more detail). Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region, and 0 if not. North Central is the omitted reference group. Other variables are coded as above. Primary Vote = β0 + β1Women’s Place + β2Strength of Party Identification + β3Education + β4South + β5Race + β6Age + ε [2.4] where Primary Vote is coded 0 if Clinton and 1 if Obama (votes for other candidates are excluded). Women’s Place is coded so that respondents had to answer all questions to be included (see above for more detail). Other variables are coded as above.

211

212

APPENDIX

Chapter 3 Democratic Vote = B0 + B1Party Identification + B2Race + B3Education + B4Region + B5Catholic + B6Jewish + B7Other Religion + B8[Catholic × 1956] + B9[Catholic × 1960] + . . . B19[Catholic × 2008] + B20[1956] + B21[1960] + . . . B39[2008] + e [3.1] where Democratic Vote is coded 1 if Democratic, 0 if Republican (other than two-party votes are excluded). Party Identification is the standard seven-point party identification scale, scored to run from 0 (Strong Republican) to 1 (Strong Democrat). Religion is a series of dummy variables (Catholic, Jewish, Religion/Other) that are coded 1 if a respondent is a member of that religion, and 0 if not. Protestant is the omitted reference group. Year is a series of dummy variables that are coded 1 if a respondent took the survey in that year and 0 if not. The omitted reference year is 1952. Race is a dummy variable for blacks and whites, with whites as the omitted reference group. Education is a series of dummy variables (Less than High School, High School, Some College, and College) that are coded 1 if a respondent has that education level and 0 if not. Less than High School is the omitted reference group. Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region and 0 if not. North Central is the omitted reference group. Democratic Vote = B0 + B1Party Identification + B2Religious Group Solidarity + B3Education + B4Region + e [3.2] where Democratic Vote is coded 1 if Democratic, 0 if Republican (other than two-party votes are excluded). Party Identification is an average of the seven-point party identification value given by a panel respondent in 1956 and 1958. The combined variable is scored to run from 0 (Strong Republican) to 1 (Strong Democrat). Religious Group Solidarity is an average of the respondent’s closeness to Catholics in 1956 and 1958 (see above for more detail). Education is a series of dummy variables (Less than High School, High School, Some College, and College) that are coded 1 if a respondent has that

APPENDIX

education level and 0 if not. Less than High School is the omitted reference group. Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region and 0 if not. North Central is the omitted reference group. Democratic Vote = B0 + B1Party Identification + B2 Anti- Catholic Sentiment + B4Race +B5Education + B6Region + e [3.3] where Anti-Catholic Sentiment is based on 1956 responses (see above for more detail). Race is a dummy variable for blacks and whites, with whites as the omitted reference group. Other variables are coded as immediately above. Democratic Vote = B0 + B1Race + B2[Race × 2008] + B32008 + B4Party Identification + B5Religion + B6Education + B7Region + B8Age + e [3.4] where Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Other variables are coded as above. Democratic Vote = B0 + B1Race + B2[Race × 2008] + B3[Race × (1952, 1956, 1960)] + B4[Race × (1964, 1968)] + B5Party Identification + B6Religion + B7Education + B8Region + B9Age +B10(2008) + B11[(1952, 1956, 1960)] + B12[(1964, 1968)] + e [3.5] where All variables are coded as above. Democratic Vote = B0 + B1Party Identification + B2Racial Group Solidarity + B3Education + B4Region + B5Sex + e [3.6] where Democratic Vote is coded 1 if Democratic, 0 if Republican (other than two-party votes are excluded). Racial Group Solidarity is coded so that respondents had to answer both questions to be included (see above for more detail). Party Identification is the standard seven-point scale, recoded to run from 0 (Strong Republican) to 1 (Strong Democrat). Other variables are coded as above.

213

214

APPENDIX

Democratic Vote = B0 + B1Party Identification + B2Racial Resentment + B3Education + B4Region + B5Sex + B6Bush’s Per formance + B7Egalitarianism+ B8Limited Government + e [3.7] where Democratic Vote is coded 1 if Democratic, 0 if Republican (other than two-party votes are excluded). Party Identification is the standard seven-point scale, recoded to run from 0 (Strong Republican) to 1 (Strong Democrat). Racial Resentment is coded so that respondents had to answer all questions to be included (see above for more detail). Bush Per formance is a five-point scale coded to run from 0 (Strongly Disapprove) to 1 (Strongly Approve). Limited Government is an index composed of responses to three questions. As with other indices, the overall scale score is simply the average score across the items answered, with each answered item weighted equally. The variable is coded so that 0 is the response in favor of more government and 1 is the response in favor of less government. The items are: For each question, choose the response that comes closest to your own views: The less government, the better OR There are more things that government should be doing. We need a strong government to handle today’s complex economic problems OR The free market can handle these problems without government being involved. The main reason government has become bigger over the years is because it has gotten involved in things that people should do for themselves OR Government has become bigger because the problems we face have become bigger.

Mean 0.55

Standard Deviation

Range

Reliability

0.42

0–1.0

0.81

APPENDIX

Egalitarianism is an index composed of responses to two questions. As with other indices, the overall scale score is simply the average score across the items answered, with each answered item weighted equally. The variable is coded so that 0 is the most egalitarian response and 1 is the most inegalitarian response. The items are: After each statement, please mark the appropriate box to indicate how strongly you agree or disagree: Our society should do whatever is necessary to make sure that everyone has an equal opportunity to succeed. It is not really that big a problem if some people have more of a chance in life than others.

Mean 0.33

Standard Deviation

Range

Reliability

0.26

0–1.0

0.69

Other variables are coded as above.

Chapter 4 Turnoutt = β0 + β1Racet + β2Educationt + β3Incomet + β4 Aget + β5[Aget > 65] + β6Sext + ε [4.1] where Turnout is coded 1 if the respondent reported voting, 0 if not. Income is a five-category scale that ranges from 0 (0–16 percentile of family income) to 1 (96–100 percentile). Age > 65 is a dummy variable that is coded 1 if the respondent is greater than or equal to 65 years old and is 0 otherwise. Sex is a dummy variable, with Male coded as 0 and Female coded as 1. Education is a four-category scale, where 0 is Grade School or Less and 1 is College or Advanced Degree. Other variables are coded as above.

215

216

APPENDIX

explaining presidential elections

Year

Et

Vt

IMt

Wt

1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008

0.023 0.010 0.035 0.009 0.053 0.035 0.032 0.030 −0.002 0.060 0.032 0.018 0.018 0.037 0.025 −0.003

0.523695 0.445481 0.577517 0.499174 0.613447 0.495941 0.617861 0.489477 0.446947 0.591696 0.539016 0.465451 0.547353 0.50268 0.51244 0.463116

−0.14 0.29 0.64 0.43 1.44 −0.41 1.12 0.47 0.50 0.12 0.65 0.08 −0.65 −0.37 0.11 0.64

0 −0.075 0 0 0 −0.187 0.371 0 0 0 0 0 0 0 0.051 −0.172

Note: Vt (the dependent variable) is the incumbent party’s share of the two-party vote. Et is the percentage annual change in the economy over the preceding year leading up to electiont (October to October), measured in real disposable income per capita. Wt is the balance of support in the general public for U.S. military involvement in the presence of war (1952, 1968, 1972, and 2004) and 0 otherwise; k = 1 if the incumbent party is the initiator of war (1952, 1968, 2004) and -1 if not (1972). IMt = ⏐Challenger’s Ideology t⏐-⏐Incumbent’s Ideology t⏐.

Chapter 5 Evaluation of Jackson = B0 + B1Racial Group Solidarity + B2Strength of Party Identification + B3Education + B4Region + B5Sex + B6Age + e [5.1] where Evaluation of Jackson is based on the one hundred–point thermometer scale, coded here to run from 0 (Very Cold) to 1 (Very Warm). Racial Group Solidarity is a difference score between Black and White thermometer ratings (see above for more detail). Strength of Party Identification is a four-point scale based on the Democratic side of the seven-point party identification scale, recoded to run from 0 (Independent) to 1 (Strong Democrat). Education is a seven-point scale. The variable runs from 0 (Less than High School) through 1 (Advanced Degree).

APPENDIX

Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Sex is a dummy variable, with Male coded as 0 and Female coded as 1. Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region and 0 if not. North Central is the omitted reference group. Evaluation of Obama = B0 + B1Racial Group Solidarity + B2Strength of Party Identification + B3Education + B4Region + B5Sex + B6Age + e [5.2] where Evaluation of Obama is based on the one hundred–point thermometer scale, coded here to run from 0 (Very Cold) to 1 (Very Warm). Racial Group Solidarity is a difference score between Black and White thermometer ratings (see above for more detail). Strength of Party Identification is a four-point scale based on the Democratic side of the seven-point party identification scale, recoded to run from 0 (Independent) to 1 (Strong Democrat). Education is a four-point scale. The variable runs from 0 (Less than High School) through 1 (College/Advanced Degree). Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Sex is a dummy variable, with Male coded as 0 and Female coded as 1. Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region, and 0 if not. North Central is the omitted reference group. Primary Vote 1988 = B0 + B1Racial Resentment + B2Strength of Party Identification + B3Education + B4Region + B5Sex + B6Age + e [5.3] where Primary Vote is coded 0 if Dukakis and 1 if Jackson (votes for other candidates are excluded). Racial Resentment is coded so that respondents had to answer three of the four component questions to be included (see above for more detail). Other variables are coded as above. Primary Vote 2008 = B0 + B1Racial Resentment+ B2Strength of Party Identification + B3Education + B4Region + B5Sex + B6Age + e [5.4] where Primary Vote is coded 0 if Clinton and 1 if Obama (votes for other candidates are excluded).

217

218

APPENDIX

Racial Resentment is coded so that respondents had to answer three of the four component questions to be included (see above for more detail). Other variables are coded as above. Evaluation of Powell = B0 + B1Racial Group Solidarity + B2Party Identification + B3Education + B4Sex + B5Age + e [5.5] where Evaluation of Powell is based on the one hundred–point thermometer scale, coded here to run from 0 (Very Cold) to 1 (Very Warm). Racial Group Solidarity is coded so that respondents had to answer three of the four component questions to be included (see above for more detail). Party Identification is a five-point scale, coded to run from 0 (Strong Republican) to 1 (Strong Democrat). Education is a nine-point scale. The variable runs from 0 (Grade School) through 1 (Doctorate/Law Degree). Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Sex is a dummy variable, with Male coded as 0 and Female coded as 1. Evaluation of Obama = B0 + B1Racial Group Solidarity + B2Party Identification + B3Education + B4Sex + B5Age + e [5.6] where Evaluation of Obama is based on a ten-point thermometer scale, coded here to run from 0 (Very Cold) to 1 (Very Warm). Racial Group Solidarity is coded so that respondents had to answer two of the three component questions to be included (see above for more detail). Party Identification is the standard seven-point scale, recoded to run from 0 (Strong Republican) to 1 (Strong Democrat). Education is a six-point scale. The variable runs from 0 (No High School) through 1 (Postgrad). Sex is a dummy variable, with Male coded as 0 and Female coded as 1. Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Evaluation of Powell = B0 + B1Racial Resentment + B2Party Identification + B3Education + B4Region + B5Religion + B6Sex + B7 Age + e [5.7] where Evaluation of Powell is measured in 1996 and based on the one hundred– point thermometer scale, coded here to run from 0 (Very Cold) to 1 (Very Warm).

APPENDIX

Racial Resentment is measured in 1992 and coded so that panel respondents had to answer three of the four component questions to be included (see above for more detail). Party Identification is the standard seven-point scale, recoded to run from 0 (Strong Republican) to 1 (Strong Democrat). Education is a seven-point scale. The variable runs from 0 (Less than High School) through 1 (Advanced Degree). Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region and 0 if not. North Central is the omitted reference group. Religion is a series of dummy variables (Catholic, Jewish, Religion/Other) that are coded 1 if a respondent is a member of that religion and 0 if not. Protestant is the omitted reference group. Sex is a dummy variable, with Male coded as 0 and Female coded as 1. Age is the respondent’s age in years at the time of the survey. It is coded to range from 0 to 1. Evaluation of Obama = B0 + B1Racial Resentment + B2Party Identification + B3Education + B4Region + B5Religion + B6Sex + B7 Age + e [5.8] where Racial Resentment is coded so that respondents had to answer all of the component questions to be included (see above for more detail). Region is a series of dummy variables (South, West, North Central, and Northeast) that are coded 1 if a respondent lives in that region, and 0 if not. North Central is the omitted reference group. Religion is a series of dummy variables (Catholic, Jewish, Muslim, Religion/None, Religion/Other) that are coded 1 if a respondent is a member of that religion, and 0 if not. Protestant is the omitted reference group. Other variables are coded as above.

Chapter 6 Evaluation of Obama = B0 + B1Racial Resentment + B2Party Identification + B3ObamaVote + B4 McCainVote + B5LimitedGovt + e [6.1] where Evaluation of Obama is based on a seven-point thermometer scale, coded here to run from 0 (Very Cold) to 1 (Very Warm). Racial Resentment is coded so that respondents had to answer all questions to be included (see above for more detail). Party Identification is the standard seven-point scale, recoded to run from 0 (Strong Republican) to 1 (Strong Democrat).

219

220

APPENDIX

ObamaVote and McCainVote are dummy variables that equal 1 if the respondent voted for that candidate; and 0 if the respondent voted for the other candidate, for a third-party candidate, or did not vote at all. Limited Government is a seven-category variable coded such that 1 is a position in favor of limiting government and 0 is a position in favor of expanding government.

APPENDIX

Auxiliary Results Chapter 2 Table 2.2 Complete Results: The Effect of Group Solidarity on Ratings of Clinton and Obama in 2008 Least Squares Regression (Standard Errors in Parentheses) Ratings of Clinton by Women

Group Solidarity Partisan Strength Race (Black) Sex (Female) Age Education South Catholic Intercept R2 N

December

March

0.08 (0.08) 0.24** (0.08) 0.09** (0.04) –

0.10 (0.11) 0.22** (0.10) −0.08 (0.06) –

−0.08 (0.10) 0.03 (0.08) −0.01 (0.04) −0.05 (0.05) 0.56*** (0.10) 0.09 237

−0.12 (0.13) 0.02 (0.10) −0.02 (0.05) −0.06 (0.07) 0.57*** (0.12) 0.04 182

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 2007–2008 Cooperative Campaign Analysis Project.

Ratings of Obama by African Americans December

March

0.25* (0.13) −0.08 (0.13) –

0.35** (0.13) −0.08 (0.14) –

−0.00 (0.06) −0.19 (0.15) 0.12 (0.14) –

−0.06 (0.06) −0.27* (0.14) −0.06 (0.15) –

–

–

0.77*** (0.21) 0.08 73

0.89*** (0.21) 0.17 60

221

Table 2.6 Complete Results: The Effect of Men’s Beliefs about Women’s Place on their Vote in the 2008 Democratic Primary Logistic Regression (Standard Errors in Parentheses)

Women’s Place Partisan Strength Race (Black) Age Education South Intercept N

Intended Vote (December)

Reported Vote (March)

Recollected Vote (November)

5.26* (3.05) −2.78 (1.73) 0.59 (0.96) −5.48** (1.77) 1.52 (1.46) 0.15 (0.79) 1.32 (1.91) 45

2.31 (2.26) −0.45 (1.15) 2.32** (1.13) −4.34** (1.81) 4.30** (1.45) −0.80 (0.65) −1.06 (1.41) 69

−0.70 (0.85) 0.09 (0.99) 3.22*** (0.78) 0.21 (1.26) 2.20 (2.35) −1.17** (0.56) −1.67 (2.37) 130

*p < 0.10 **p < 0.05 ***p < 0.001. Note: Each column represents an equation. Source: 2007–2008 Cooperative Campaign Analysis Project (columns 1 and 2) and 2008 National Election Study (column 3).

APPENDIX

Chapter 3 The Religious Factor in Presidential Voting, 1952– 2008 Logit Coefficients (Standard Errors in Parentheses)

Catholic Catholic × 1960 1960 Partisanship Race (Black) Jewish Other religion High School Some College College Northeast South West Catholic × 1956

1

2

0.07 (0.06) 1.75*** (0.32) −0.26** (0.12) 5.19*** (0.08) 2.18*** (0.13) 1.32*** (0.16) 0.80*** (0.08) 0.01 (0.08) −0.01 (0.09) 0.25** (0.09) −0.01 (0.07) −0.60*** (0.06) −0.00 (0.07) –

0.10 (0.18) 1.73*** (0.38) 0.35** (0.16) 5.49*** (0.09) 2.13*** (0.13) 1.39*** (0.17) 0.65*** (0.09) −0.17** (0.08) −0.29** (0.09) −0.09 (0.09) 0.03 (0.07) −0.66*** (0.07) −0.00 (0.07) −0.32 (0.25) 0.48 (0.31) 0.60** (0.29) −0.33 (0.25) −0.31 (0.27)

Catholic × 1964

–

Catholic × 1968

–

Catholic × 1972

–

Catholic × 1976

–

(continued)

223

(Continued) 1

224

Catholic × 1980

–

Catholic × 1984

–

Catholic × 1988

–

Catholic × 1992

–

Catholic × 1996

–

Catholic × 2000

–

Catholic × 2004

–

Catholic × 2008

–

1956

–

1964

–

1968

–

1972

–

1976

–

1980

–

1984

–

1988

–

1992

–

1996

–

2000

–

2004

–

2 −0.04 (0.30) 0.08 (0.25) −0.01 (0.27) −0.29 (0.25) −0.36 (0.28) −0.19 (0.26) 0.45 (0.30) 0.03 (0.31) 0.14 (0.13) 1.70*** (0.16) 0.19 (0.15) −0.13 (0.13) 0.98*** (0.14) 0.23 (0.16) 0.24* (0.14) 0.77*** (0.15) 1.46*** (0.14) 1.66*** (0.15) 1.16*** (0.14) 0.83*** (0.17)

(Continued)

2008

1

2

−3.18*** (0.10) 16,223

1.14*** (0.16) −3.80*** (0.13) 16,223

–

Intercept N

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 1952–2008 National Election Study.

Table 3.1 Complete Results: Religious Identification and the Catholic Vote in 1960 Logistic Regression (Standard Errors in Parentheses) JFK Vote among Catholics Religious Identification Partisanship High School Some College College Northeast South West Intercept N

2.53** (1.25) 4.73*** (1.06) −0.20 (0.65) 0.10 (0.62) −1.20* (0.68) −0.96 (0.63) −0.58 (1.40) −1.43* (0.83) −1.39 (1.02) 198

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 1956–1960 National Election Panel Study.

225

Table 3.2 Complete Results: Distrust of Catholics and the Protestant Vote in 1960 Logistic Regression (Standard Errors in Parentheses) JFK Vote among Protestants Attitude toward Catholics Partisanship Race (Black) High School Some College College Northeast South West Intercept N *p < 0.10 **p < 0.05 ***p < 0.001. Source: 1956–1960 National Election Panel Study.

226

−0.96** (0.38) 5.10*** (0.46) 1.44** (0.52) −0.16 (0.32) −0.78** (0.32) −0.26 (0.41) −0.13 (0.40) −0.38 (0.30) 0.66* (0.39) −3.32*** (0.39) 593

Table 3.3 Complete Results: The Race Factor in Presidential Voting, 1952– 2008 Logistic Regression (Standard Errors in Parentheses) Vote for the Democratic Candidate 1

2

Race × [1952, 1956, 1960]

2.07*** (0.13) 2.26*** (0.67) –

Race × [1964, 1968, 1972]

–

2.09*** (0.16) 2.21** (0.68) −0.93** (0.33) 0.58* (0.35) 5.27*** (0.09) 1.33*** (0.16) 0.67*** (0.08) 0.14** (0.06) −0.14* (0.08) −0.26** (0.09) −0.02 (0.09) 0.05 (0.07) −0.63*** (0.06) 0.00 (0.07) 0.20* (0.11) −0.53*** (0.07) −0.37*** (0.06) −2.87*** (0.10) 16,223

Race (Black) Race × [2008]

Party Identification

1952, 1956, 1960

5.20*** (0.08) 1.31*** (0.16) 0.76*** (0.08) 0.18** (0.06) −0.01 (0.08) −0.05 (0.09) 0.21** (0.09) 0.02 (0.07) −0.61*** (0.06) 0.00 (0.07) 0.36*** (0.11) –

1964, 1968, 1972

–

Jewish Religion/Other Catholic High School Some College College Northeast South West 2008

Intercept N

−3.21*** (0.10) 16,223

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 1952–2008 National Election Studies.

227

Table 3.5 Complete Results: Racial Resentment and the White Vote in 2008 Logistic Regression (Standard Errors in Parentheses) Vote for Obama

Racial Resentment Party Identification Approval of Bush’s Performance Limited Government Egalitarianism Age Education Sex West South Northeast Intercept N

1

2

−3.45*** (0.74) 2.96*** (0.49) −4.07*** (0.55) –

−3.31*** (0.79) 2.78*** (0.50) −3.97*** (0.56) −0.72* (0.39) 0.39 (0.61) −1.50* (0.76) 0.19 (0.62) −0.54* (0.31) −0.38 (0.41) −0.53 (0.34) −0.08 (0.48) 2.95** (0.94) 650

– −1.52** (0.74) 0.22 (0.62) −0.42 (0.28) −0.37 (0.40) −0.55* (0.33) −0.02 (0.46) 2.61** (0.88) 650

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 2007–2008 Cooperative Campaign Analysis Project.

228

Table 3.6 Complete Results: Racial Attitude and the White Vote in 2008 Logistic Regression (Standard Errors in Parentheses) Racial Attitude Represented by:

Racial Attitude Party Identification Education Age Sex Northeast South West Constant N

Racial Resentment

Racial Stereotype

Subtle Racism

−3.73*** (0.62) 6.12*** (0.50) −2.87** (1.12) −1.50** (0.52) 0.03 (0.26) −0.05 (0.43) −0.39 (0.33) 0.32 (0.34) 2.21** (1.09) 755

−3.83** (1.49) 6.50*** (0.53) −1.90 (1.19) −1.34** (0.53) −0.06 (0.26) 0.24 (0.41) −0.66** (0.33) 0.30 (0.33) 1.12 (1.44) 728

−1.89*** (0.54) 6.21*** (0.50) −1.78 (1.16) −1.81*** (0.52) −0.06 (0.26) 0.15 (0.44) −0.66** (0.33) 0.36 (0.33) 0.14 (1.04) 741

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 2008 National Election Study.

229

230

APPENDIX

Chapter 4 The Effect of Race on Turnout, 1952– 2008 Logistic Regression (Standard Errors in Parentheses)

Race Income Age Age > 65 Education Sex Intercept N

1952

1956

1960

1964

1968

1972

1976

−1.48*** (0.20) 1.44*** (0.25) 3.48*** (0.52) −0.92** (0.26) 2.14*** (0.35) −0.52*** (0.13) −0.59** (0.24) 1,555

−1.22*** (0.20) 1.47*** (0.26) 3.29*** (0.49) −0.47* (0.26) 1.92*** (0.31) −0.49*** (0.12) −0.78** (0.23) 1,671

−1.08*** (0.30) 1.10** (0.43) 2.74** (0.84) −0.43 (0.37) 1.95** (0.56) −0.62** (0.22) −0.14 (0.44) 1,092

−0.30 (0.20) 1.52*** (0.30) 3.15*** (0.55) −0.52* (0.29) 1.24*** (0.31) −0.09 (0.14) −0.70** (0.27) 1,381

−0.01 (0.22) 1.98*** (0.30) 3.10*** (0.53) −0.54** (0.27) 1.42*** (0.30) 0.02 (0.14) −1.16*** (0.28) 1,333

0.19 (0.23) 1.39*** (0.26) 2.27*** (0.49) −0.35 (0.26) 2.28*** (0.32) −0.38** (0.14) −0.19 (0.25) 1,837

0.20 (0.19) 1.55*** (0.26) 3.06*** (0.42) −0.48** (0.23) 1.88*** (0.27) −0.25** (0.12) −1.32*** (0.23) 1,671

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 1952–2008 National Election Studies.

APPENDIX

231

1980

1984

1988

1992

1996

2000

2004

2008

0.15 (0.22) 1.43*** (0.29) 3.68*** (0.50) −0.46 (0.28) 2.34*** (0.30) −0.03 (0.14) −1.75*** (0.27) 1,194

0.10 (0.19) 1.76*** (0.25) 3.65*** (0.47) −0.34 (0.27) 2.41*** (0.29) 0.15 (0.13) −1.95*** (0.24) 1,610

0.06 (0.18) 1.89*** (0.27) 3.29*** (0.48) −0.16 (0.27) 2.81*** (0.28) 0.19 (0.14) −2.49*** (0.26) 1,428

0.09 (0.16) 2.11*** (0.26) 3.12*** (0.46) −0.17 (0.26) 2.74*** (0.29) 0.12 (0.13) −2.07*** (0.25) 1,810

0.13 (0.22) 1.93*** (0.30) 2.73*** (0.58) 0.30 (0.31) 2.42*** (0.31) 0.01 (0.16) −1.99*** (0.31) 1,232

0.45* (0.25) 1.42*** (0.32) 3.18*** (0.58) −0.19 (0.32) 2.41*** (0.31) −0.11 (0.15) −1.93*** (0.32) 1,132

−0.07 (0.24) 1.21** (0.37) 1.64** (0.63) −0.05 (0.36) 2.08*** (0.38) 0.41** (0.19) −1.06** (0.33) 820

0.95*** (0.25) 0.98* (0.39) 2.35*** (0.64) .16 (0.42) 2.27*** (0.43) 0.52* (0.21) −1.70*** (0.37) 939

Table 4.1 Complete Results: The Impact of Racial Resentment on White Americans’ Vote for President, 1988– 2008 Logistic Regression (Standard Errors in Parentheses) Vote for the Democratic Candidate

Racial Resentment Party Identification High School Some College College Northeast South West Jewish Catholic Religion/ Other Intercept N

1988

1992

1996

2000

2004

2008

−1.70*** (0.45) 5.52*** (0.35) 0.40 (0.40) −0.09 (0.42) 0.29 (0.41) −0.35 (0.28) −0.89** (0.27) 0.10 (0.27) 1.24* (0.66) 0.19 (0.22) 0.55 (0.36) −2.28*** (0.54) 927

−2.35*** (0.50) 6.39*** (0.39) 0.50 (0.45) 0.63 (0.48) 0.04 (0.48) 0.34 (0.28) −0.36 (0.27) 0.23 (0.31) 1.69** (0.67) −0.18 (0.26) 0.90** (0.33) −2.13*** (0.64) 1,005

−2.10** (0.96) 6.88*** (0.74) −0.22 (1.01) 0.02 (1.03) −0.37 (1.01) 1.25* (0.65) 0.19 (0.52) 1.23* (0.66) –

−1.26** (0.51) 7.08*** (0.47) 0.69 (0.68) 0.82 (0.68) 1.20* (0.69) −0.20 (0.35) −0.65** (0.28) 0.40 (0.33) 2.99** (1.20) 0.07 (0.27) 0.59* (0.33) −3.87*** (0.86) 863

−2.04** (0.75) 6.70*** (0.52) −0.29 (0.71) −0.43 (0.71) −0.07 (0.70) −0.26 (0.40) −0.86** (0.39) −0.29 (0.37) 0.85 (1.08) 0.40 (0.35) 0.21 (0.35) −1.89** (0.96) 586

−3.73*** (0.65) 6.21*** (0.50) −1.90** (0.78) −1.96** (0.79) −2.44** (0.78) 0.04 (0.42) −0.37 (0.32) 0.33 (0.32) 3.01 (1.99) −0.11 (0.34) 0.49* (0.28) 1.09 (0.85) 798

−0.78 (0.57) 0.28 (0.66) −1.78 (1.20) 288

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 1988, 1992, 1992–1996 Panel, 2000, 2004, and 2008 National Election Studies.

APPENDIX

Chapter 5 Table 5.2 Complete Results: The Effect of Racial Group Identification on Black Americans’ Evaluation of Jesse Jackson and Barack Obama Least Squares Regression (Standard Errors in Parentheses)

Racial Group Identification Partisan Strength Education Age Sex West Northeast South Intercept N

Jackson (1988)

Obama (2008)

0.46** (0.17) 0.08 (0.07) 0.06 (0.07) 0.10 (0.10) 0.03 (0.04) 0.01 (0.07) −0.08 (0.06) −0.00 (0.04) 0.42** (0.13) 160

0.23*** (0.07) 0.13*** (0.03) 0.04 (0.03) −0.04 (0.04) 0.02* (0.02) 0.03 (0.03) −0.05* (0.03) −0.00 (0.02) 0.62*** (0.05) 410

*p < 0.10 **p < 0.05 ***p < 0.001. Source: 1988 and 2008 National Election Studies.

233

Table 5.3 Complete Results: The Effect of Racial Resentment on the White Democratic Primary Vote, 1988 versus 2008 Logistic Regression (Standard Errors in Parentheses) Jackson versus Dukakis (1988)

Obama versus Clinton (2008)

−3.75** (1.14) 0.35 (0.83) 0.05 (0.73) −2.46** (1.13) 0.62 (0.69) 1.31** (0.64) 0.73 (0.65) 0.66 (0.44) 0.22 (1.19) 161

−2.23** (0.85) −0.36 (0.65) 0.98 (0.76) 0.52 (0.93) −1.68** (0.70) 0.03 (0.52) −0.55 (0.51) 0.16 (0.38) 0.51 (1.16) 172

Racial Resentment Partisan Strength Education Age Northeast West South Sex Intercept N *p