ch4

THE ANALYSIS PRESENTED in chapter 3 suggests that much of the Republican Party's electoral success in the postwar era—and thus, much of the escalation in economic inequality associated with Republican administrations and policies—is a by-product of partisan biases in economic accountability. From the standpoint of democratic theory, that is a peculiarly unsatisfying conclusion. Good democrats like to think that government policy stems, directly or indirectly, from the will of the people. If it stems instead from such irrelevant quirks of voter psychology as myopia, misperceptions, and responsiveness to campaign spending, the warm glow seems distinctly diminished.

Democratic optimists, however, will note that elections are by no means the only possible avenue by which the will of the people may shape public policy. Political science provides many examples of analyses suggesting that public opinion may have significant direct effects on public policy regardless of whether Republicans or Democrats happen to be in charge.¹ Thus, my accounts of policy-making in chapters 5 through 9 will involve careful weighing of the apparent influence of public sentiment, on one hand, and partisan elite ideology, on the other.

Assessing and interpreting the role of public opinion in the policy-making process requires, first, some clear understanding of what that public opinion is. Citizens may have crystallized views regarding specific matters of public policy, or they may not.² Insofar as citizens do have meaningful policy preferences, those preferences may be internally consistent and sensibly related to their broader political values, or they may not.3 Public opinion touching on issues related to inequality is likely to be especially complex given the deep and multifaceted resonance of the value of equality in American political culture.

Much popular commentary suggests that ordinary Americans find inequality both natural and unobjectionable—that the ideal of equality, despite its prominence in our official ideology, has little real resonance in American political culture. According to sociologist Nathan Glazer, for example, “Americans, unlike the citizens of other prosperous democracies, not to mention those of poor countries, do not seem to care much about inequality. … [E]ven after the Enron and other scandals, most Americans remain apathetic about inequality: What we have today is outrage against those who do not play fair—not outrage over inequality as such.” In a similar vein, business writer Robert Samuelson argued that, “on the whole, Americans care less about inequality—the precise gap between the rich and the poor—than about opportunity and achievement: are people getting ahead?”⁴

Analysts of American ideology have often emphasized the potential for conflict and contradiction between the core values of economic opportunity and political equality. For example, Jennifer Hochschild found that people have quite distinct normative intuitions about economic, social, and political inequality. In the economic domain, her rich and poor respondents tended to “agree on a principle of differentiation, not of equality. … They define political freedom as strict equality, but economic freedom as an equal chance to become unequal.” Similarly, Herbert McClosky and John Zaller wrote that vast differences in wealth and life chances result “partly from the play of economic interests and the desire of those who have prospered to retain their advantages, and partly from the widespread acceptance of a powerful set of values associated with the private enterprise system that conflicts with egalitarianism.” Stanley Feldman provided evidence that people use distinct, often conflicting principles to assess “micro-justice” (individual rewards) and “macro-justice” (the social distribution of rewards).⁵

In this chapter I explore three important facets of Americans' views about equality. First, I examine public support for broad egalitarian values and the political consequences of that support. Second, I examine public attitudes toward salient economic groups, including rich people, poor people, big business, and labor unions, among others. Third, I examine public perceptions of inequality and opportunity, including perceptions of growing economic inequality, normative assessments of that trend, and explanations for disparities in economic status—and how these views are shaped by the interaction of political information and political ideology.

The first half of the chapter provides strong evidence of the importance of egalitarian values in American political culture. The vast majority of Americans express support for equality in the abstract, and perhaps more surprisingly, those egalitarian values are translated into support for a variety of concrete social welfare policies, including government services, employment programs, health insurance, and aid to African-Americans. At a more visceral level, survey respondents say that they feel much warmer toward “working-class people” and “poor people” than toward “business people” and “rich people”—and those class feelings, too, bolster support for egalitarian policies. Most people also believe that the rich pay less than they should in taxes.

All of this sounds like fertile ground for a populist backlash against the escalating inequality of the New Gilded Age. Perhaps it is. However, the second half of this chapter provides some significant grounds for doubt. Three-fourths of Americans said that “the difference in incomes between rich people and poor people in the United States today” was larger than it was 20 years ago, but those responses appear to reflect cynical folk wisdom more than close attention to actual economic trends. The proportion of the public agreeing with the sardonic adage that “the rich get richer and the poor get poorer” increased markedly in the 1960s and early 1970s—a period in which the poor became rather less poor, though poverty became more salient. However, the substantial increase in economic inequality over the following three decades had virtually no effect on perceptions of inequality. Meanwhile, among people who did recognize that income differences had increased, more than one-third said that they had not thought about whether that was a good thing or a bad thing.

Public ignorance and inattention have important implications for the politics of inequality—implications I shall explore in more detail in chapters 5 and 6. However, my consideration of public opinion in the realm of inequality also highlights a very different limitation on the translation of egalitarian values into populist politics. Whereas uninformed people may often fail to grasp the political relevance of their beliefs and values, well-informed people may often deny or distort inconvenient facts, preserving a false consistency between their beliefs about the causes, extent, and meaning of inequality in contemporary American society, on one hand, and their ideological or partisan predilections, on the other. In this way, too, genuine allegiance to the ideal of equality may comfortably coexist with fervent support for policies that exacerbate inequality.

Sure, we hold this truth to be self-evident, that all men are created equal. But it is far from self-evident what that means. Is it inconsistent with American ideals for corporate CEOs to earn 300 times as much as their workers? Are ordinary Americans offended by the fact that the top 1% of income-earners have hauled in more than half of the nation's total gains in real income over the past two decades? Is it wrong for millions of children to grow up without access to decent schools and routine health care? If not, then escalating economic inequality may, after all, be a perfectly understandable and unobjectionable outcome of (small “d”) democratic politics.

One useful starting point in examining public thinking about economic inequality in the New Gilded Age is to assess people's allegiance to equality as a general value, sidestepping for the moment the complexities that may arise in applying that value to specific cases. While it would be foolhardy to confuse pronouncements of support for equality in the abstract with support for specific policies promoting specific kinds of equality in specific circumstances, it would be equally foolhardy to deny that people's allegiance to abstract values may have important practical political consequences.

The first question in table 4.1 has consistently elicited the most enthusiastically egalitarian opinion. In periodic surveys going back to the early 1980s, more than 85% of Americans have agreed that “our society should do whatever is necessary to make sure that everyone has an equal opportunity to succeed,” and more than 60% have agreed strongly with that statement. Interpreted literally, these responses imply an astounding level of public support for what would have to be a very radical program of social transformation. Inherited wealth would have to be outlawed; local funding of public schools would have to cease; every child would have to be assigned competent and loving parents; and prejudices based on race, ethnicity, religion, and sex would have to be eradicated, along with social and economic advantages derived from intelligence, physical attractiveness, and freakish athletic skills. Of course, this is not really what people mean when they agree that we “should do whatever is necessary to make sure that everyone has an equal opportunity to succeed.” Nevertheless, their willingness to endorse such a statement, even as a matter of verbal ritual, provides a striking testament to the force of “equal opportunity” in American culture.

More generally, the questions in table 4.1 referring to “an equal chance” or equal treatment generated favorable responses from half to two-thirds of the respondents, while those referring to “how equal people are” and “pushing equal rights” were rather less popular.⁸ These differences suggest a consequential distinction in public thinking between equality of opportunity and equality of results. Nevertheless, more detailed analysis of individual response patterns suggests that there is a fair degree of consistency in people's responses to all six questions, regardless of their precise wording. Thus, it is convenient to combine responses to all six questions in a summary scale measuring general support for egalitarian values.⁹ The summary scale values range from −1 to +1, with an average value of +.21. Twenty percent of the ANES respondents had very egalitarian scale scores (in excess of +.5), while fewer than 3% had very inegalitarian scores (below −.5).

The analyses reported in the top panel of the table show how positions on each of these issues varied with attachment to egalitarian values. For government jobs and aid to blacks, the most egalitarian people were, on average, more than three points to the left of the most inegalitarian people on the seven-point policy scale—a difference between moderately liberal positions, on one hand, and solidly conservative positions, on the other. For views about government spending and services and health insurance, the average differences were almost as large, about 2.8 points.

Of course, people's views about equality are strongly correlated with their broader political ideology and partisanship. Thus, one may wonder whether the strong relationships documented in the top panel of table 4.2 are simply reflections of these more concrete political allegiances. However, that does not turn out to be the case. Not surprisingly, liberals and Democrats were significantly more supportive of government action in each of these issue areas than conservatives and Republicans were.¹⁴ Nevertheless, even after taking account of differences in policy preferences attributable to ideology and partisanship, the analyses presented in the bottom panel of table 4.2 suggest that egalitarian values had a substantial independent impact. For two of the four issues (government jobs and aid to blacks), the estimated effects of egalitarian values on policy preferences exceeded the combined effects of ideology and partisanship.

The fact that social welfare policy preferences seem to be strongly shaped by broader views about equality is especially important in light of the overall level of support for egalitarian values in the American public. If opponents of equality were just as numerous as proponents of equality, even the strong effects of egalitarian values on policy preferences reported in table 4.2 would not alter the average levels of public support for government action in these policy areas. However, the combination of net support for egalitarian values in table 4.1 and strong effects of egalitarian values on policy preferences in table 4.2 translates into significant additional net public support for liberal policies. For the issues of government jobs and aid to blacks, the overall distributions of public opinion are a good deal less conservative than they would otherwise be; for the issues of government services and health insurance, distributions of opinion that would otherwise be precisely balanced between liberal and conservative extremes instead lean in a liberal direction.¹⁵ In this respect, at least, Americans' support for egalitarian values is much more than just lip service—it has significant political consequences. As Sidney Verba and Gary Orren put it, “values are instrumental in shaping the public policies that give practical effect to political belief.”¹⁶

Ordinary people's responses to such abstract questions as whether we should “make sure that everyone has an equal opportunity to succeed” or whether we have “gone too far in pushing equal rights in this country” turn out to be strongly related to their views about important issues of public policy. However, it seems unlikely that most public thinking about inequality occurs at such a high level of abstraction. Indeed, political scientists have amassed a good deal of evidence suggesting that ordinary citizens engage in rather little abstract reasoning in most realms of politics, relying instead on positive or negative attitudes toward salient social groups to shape their reactions to specific public policies, political candidates, and social conditions.¹⁷

In the realm of economic inequality, such salient groups might include rich people, poor people, big business, and the working class, among others. ANES surveys over the years have measured public attitudes toward these and a variety of other social groups using a “feeling thermometer.” Survey respondents are invited to rate each group on a scale ranging from 0 to 100, with 50 meaning “you don't feel particularly warm or cold” toward the group, ratings above 50 representing “favorable and warm” feelings, and ratings below 50 representing “unfavorable and cool” feelings. Table 4.3 reports the average thermometer ratings for various groups in the 2004–2012 ANES surveys. The groups are listed in order of popularity: at one extreme, “working-class people” received an average rating of 83 out of 100; at the other extreme, the Republican Party received an average rating of 50—just below Muslims and gay men and lesbians.

The results presented in table 4.4 suggest that class sympathies do have significant effects on policy preferences, over and above the effects of egalitarian values, political ideology, and party identification. Most people's class sympathies did not push them strongly in either direction (because they reported fairly warm feelings about all of the groups included in the summary measure). Nonetheless, those with moderately egalitarian class sympathies (one standard deviation above the mean) took positions that were about half a point more liberal than those with moderately inegalitarian class sympathies (one standard deviation below the mean) on three of the four seven-point scales. For these three policies—government jobs, government services, and health insurance—the estimated effects of class sympathies rival in importance the estimated effects of egalitarian values and ideology. (For aid to blacks, a policy domain in which class sympathies are presumably less relevant, the estimated effect was much smaller, about one-fifth of a point.)

The 2002 and 2004 ANES surveys included another set of questions tapping attitudes toward rich people and poor people in the specific context of a policy issue where class sympathies are presumably even more relevant. These questions asked whether rich people, poor people, and the respondents themselves are asked to pay too much, too little, or about the right amount in federal income taxes. The distributions of responses to these questions are presented in table 4.5.¹⁸

Ordinary citizens' values and sympathies are likely to be politically significant only when they are connected to consequential political actors or issues. But the political activation of values and sympathies is far from automatic. For example, egalitarian impulses are likely to gain real political traction only when citizens perceive contradictions between egalitarian ideals and social realities.²⁰ Given the enormous complexity of politics and public affairs, connections of that sort may require a good deal of political information and attention. As the acute political observer Walter Lippmann noted almost a century ago, the fact that much everyday political thinking occurs in a “pseudo-environment” only loosely connected to social reality often postpones or prevents “the murder of a Beautiful Theory by a Gang of Brutal Facts.”²¹

In the present context, it seems well worth exploring how much ordinary Americans know about the Brutal Facts of escalating inequality sketched in chapter 1. Recent ANES surveys have included a variety of questions probing respondents' perceptions of economic inequality, its causes, and its consequences. Table 4.6 presents the distribution of responses to a question in the 2002–2012 surveys assessing people's recognition of growing income inequality. The question asked whether “the difference in incomes between rich people and poor people in the United States today is larger, smaller, or about the same as it was 20 years ago.” Those who said “larger” or “smaller” were asked whether the difference was much larger or smaller or only somewhat larger or smaller.

If we ask where there is potential political support for egalitarian redistribution in these responses, the obvious place to begin is with the 32% of the public who said that the difference in incomes between rich and poor was much larger than it had been and that that was a bad thing. Those people were at the core of a somewhat larger group—44% of the public—who both recognized and regretted the fact that economic inequality had increased. In contrast, outright supporters of economic inequality—those who applauded the fact that inequality had increased or believed with regret that inequality had declined—constituted less than 6% of the public.

In addition to asking respondents about their perceptions and evaluations of economic inequality, the 2002 ANES survey included a battery of questions inviting respondents to explain why, “in America today, some people have better jobs and higher incomes than others do.”²⁴ The survey offered a variety of potential explanations ranging from “some people just don't work as hard” to “discrimination holds some people back” to “God made people different from one another.” Respondents were asked to indicate whether each potential explanation was “very important,” “somewhat important,” or “not important.” Their answers are summarized in table 4.8, which lists the seven potential explanations in order of popularity.

The quintessential American belief that economic success is a matter of hard work fares well in table 4.8, with 45% of the public saying that unequal effort is a “very important” cause of economic inequality. However, there is even more support (about 55%) for the notion that unequal access to a good education is very important. Two other social factors, discrimination and government policies, also loom fairly large as explanations for economic inequality. And while one-third of the respondents said that differences in “inborn ability to learn” are very important, almost half rejected the idea that income differences exist because “God made people different from one another.” These responses certainly do not suggest that most Americans view economic inequality as a merely natural phenomenon, even if they do think it is attributable in part to differences in character and ability.

It should not be surprising that conservatism provides moral justification for inequality. However, it should be surprising that ideological commitments also had substantial effects on responses to the factual questions included in the table. Extreme conservatives were twice as likely as extreme liberals to say that hard work is a very important explanation for income differences (60% versus 28%). They were also more than three times as likely as extreme liberals to insist that poor people have the same access to justice that rich people do (36% versus 11%), though even most extreme conservatives disagreed with this proposition. Given the complexity of these questions and the paucity of hard evidence available to adjudicate them, liberals and conservatives managed to construct very different, ideologically congenial pictures of the social reality of inequality.

Even in the case of the most straightforwardly factual question included in the table—whether the income gap between rich and poor had increased—the responses of conservatives and liberals were strongly colored by their ideological perspectives. Almost 90% of extreme liberals, but only 73% of extreme conservatives, thought that income differences had increased. Put the other way, conservatives were more than twice as likely as liberals to deny that the difference in incomes between rich people and poor people had grown larger. In light of the magnitude of the economic trends described in chapter 1, it is striking that more than one in four extreme conservatives persisted in believing that income differences between rich people and poor people have not, in fact, increased over the past two decades.

The statistical results presented in table 4.9 are derived from the responses of a representative national sample of the adult population. Such a sample includes many people who pay little attention to news and have only the vaguest grasp of public affairs. If these inattentive citizens relied on ideological surmise because they lacked solid information about actual trends in economic inequality, the ideological disparities in perceptions evident in the table might simply reflect the (potentially remediable) limits of civic education in this domain.

In order to address that possibility, the statistical analyses presented in table 4.9 must be elaborated to allow for the (potentially distinct) effects of political information within each ideological group. I do that using general measures of political information in the ANES surveys. The 2004 ANES survey included questions asking: (1) which party had more members in the House of Representatives prior to the election; (2) which party had more members in the Senate prior to the election; (3) which party was more conservative at the national level; (4) what job or political office Dennis Hastert held; (5) what job or office Dick Cheney held; (6) what job or office Tony Blair held; and (7) what job or office William Rehnquist held. I constructed a political information scale based on correct answers to these seven questions. (Two of the 1,066 respondents answered all seven questions correctly; 22% gave five or more correct answers, while 42% gave two or fewer correct answers.) The 2002 ANES survey did not include an analogous battery of factual questions about politics; instead, I employed a subjective rating of respondents’ “general level of information about politics and public affairs” (on a five-point scale ranging from “very low” to “very high”) provided by the interviewer at the end of each interview.³⁰

Most of the differences between better-informed and less-informed people in table 4.10 parallel those between liberals and conservatives in table 4.9. Better-informed people were more likely to say that income inequality had increased and more likely to say that was a bad thing. They were also more likely to say that poor people were disadvantaged in the legal system and (slightly) less likely to say that differences in hard work were a very important cause of inequality. In each of these respects, political information shifted people's views in the same direction as liberal political ideology.

Although the primary effects of political information and liberal ideology were generally reinforcing, table 4.10 also presents strong evidence of interactions between information and ideology: increasing political awareness seems to have had very different effects on liberals and conservatives. Indeed, for each of the four questions included in the table, greater political awareness shifted the views of liberals and conservatives in opposite directions.

FIGURE 4.2 Perceptions of Increasing Income Differences by Ideology and Information Level

The pattern of ideological polarization in figure 4.2 will appear familiar to readers of John Zaller's influential book The Nature and Origins of Mass Opinion. Zaller's model of opinion formation encompasses situations in which a “mainstream message” produces uniform shifts in opinion among politically aware conservatives and liberals. However, in situations where political elites are ideologically polarized, Zaller's model implies a very different pattern, with politically aware conservatives and liberals pulled in opposite directions by the contrasting arguments of their respective elites. The result is a characteristic diverging pattern, with undifferentiated (and generally moderate) policy preferences among less-aware conservatives and liberals, but increasing ideological disparity among those who are more attentive to elite rhetoric.35

The significance of political awareness in Zaller's model stems from the fact that logical (or merely conventional) connections between broad ideologies and specific policies are not likely to be automatically and identically evident to all conservatives and liberals. Rather, their apprehension requires careful thought—or, more likely, secondhand exposure to careful thought—about specific facts and values in the light of broad ideological commitments. People who are unusually attentive to political discourse and sophisticated in their political thinking are especially likely to grasp those connections.

The difference here is that political awareness produced sharp ideological polarization not just in policy preferences but also in perceptions of a seemingly straightforward, objective fact. Rather than contributing to accurate apprehension of that fact by conservative and liberal observers alike, political awareness seems mostly to have taught people how the political elites who share their ideological commitments would like them to see the world. In particular, what is most significant in figure 4.2 is that the conservatives who were most politically aware were most likely to deny that income differences had increased. In this instance, political awareness did more to facilitate ideological consistency than it did to promote an accurate perception of real social conditions.³⁶

Figure 4.3 shows how ideology and political information combined to influence the probability that respondents in the ANES surveys would both recognize and regret the growth of economic inequality. Among extreme liberals (represented by the solid line in the figure), that probability increased dramatically with increases in general political awareness. At the bottom of the distribution of information, only about two-thirds of extreme liberals recognized that inequality had increased, and fewer than half of those who recognized an increase said that was a bad thing. (Most of the rest said that they had not thought about whether increasing inequality was good or bad.) However, among the best-informed liberals, more than 95% recognized that inequality had increased, and more than 95% of that 95% said that was a bad thing. In contrast, among extreme conservatives (represented by the dotted line in figure 4.3) the effect of political information was to depress both recognition of growing inequality and opposition to growing inequality. Thus, while the least-informed conservatives were about as likely as the least-informed liberals to recognize and regret growing inequality, only about one in six of the best-informed conservatives did so.

FIGURE 4.3 Probability of Recognizing and Regretting Increasing Income Differences

FIGURE 4.4 Perceptions That “Some People Just Don't Work as Hard” by Ideology and Information Level

In examining the role of public opinion in the policy-making process, it will be crucial to consider the ways in which citizens' policy preferences reflect political values and sympathies of the sort examined in this chapter. However, it will be equally important to bear in mind the extent to which many ordinary citizens fail to translate their broad values and sympathies into consistent views about specific policy issues. As we saw in tables 4.2 and 4.4, egalitarian values and sympathies do seem to undergird liberal positions on a variety of important policy issues. But the analyses presented in chapters 5 and 6 will suggest that the political significance of economic inequality is mostly lost on many Americans. Although they may express genuine allegiance to egalitarian values, they are not sufficiently attuned to the political debate to apply their values sensibly and consistently in formulating responses to some of the most important issues in contemporary American politics.

¹ Perhaps the most ambitious work along these lines is by Erikson, MacKuen, and Stimson (2002), though, as I note in chapter 11, their statistical analyses seem to imply smaller direct effects of public opinion on policy and larger electoral effects than one might suppose from their prose.

² Writing in response to the first wave of enthusiasm for plebiscitary democracy at the turn of the 20th century, A. Lawrence Lowell (1913, 45–46) noted that “no people, however civilized, are capable of forming real opinions on all subjects,” and that “no real public opinion is possible” unless “the essential facts are matters of common knowledge from everyday experience, or where they have been so much discussed that familiarity with them has been generally diffused.”

³ Philip Converse's (1964) essay on “The Nature of Belief Systems in Mass Publics” is the single most influential assessment of the coherence and consistency of political opinions. Achen and Bartels (2016, especially chap. 2) summarized the relevant scholarly literature and its implications for democratic responsiveness.

⁴ Glazer (2003, 111); Robert J. Samuelson, “Indifferent to Inequality?” Newsweek, May 7, 2001, 45.

⁵ Hochschild (1981, 111, 278); McClosky and Zaller (1984, 63); Feldman (2003).

⁶ Hochschild (1981, 280); McClosky and Zaller (1984, 177–178).

⁷ Feldman (1988) provided a detailed discussion of the theoretical pedigree and measurement properties of these items.

⁸ The three items in table 4.1 inviting respondents to agree with egalitarian sentiments all elicit more egalitarian responses than the three inviting them to disagree with egalitarian sentiments. Are people simply prone to agree with such statements regardless of their content? A comparison of responses to the third and fourth questions, which seem closest to being mirror images of each other, is fairly reassuring on this score: the two questions generated similar distributions of responses, with pluralities of 52–34 and 49–32, respectively, for the egalitarian position.

⁹ A factor analysis of responses to all six questions produces a strong first factor reflecting support for egalitarian values and a much weaker second factor reflecting a response tendency to agree with all six items regardless of whether they measure agreement or disagreement with the egalitarian position. Since all six items load fairly similarly on the first (substantive) factor (and since the items are balanced between agreement and disagreement), I combine them into a summary scale by simply averaging each respondent's answers to the six questions.

¹⁰ “Some people feel the government in Washington should see to it that every person has a job and a good standard of living. Suppose these people are at one end of a scale, at point 1. Others think the government should just let each person get ahead on their own. Suppose these people are at the other end, at point 7. And, of course, some other people have opinions somewhere in between, at points 2, 3, 4, 5, or 6. Where would you place yourself on this scale, or haven't you thought much about this?”

¹¹ “Some people think the government should provide fewer services even in areas such as health and education in order to reduce spending. … Other people feel it is important for the government to provide many more services even if it means an increase in spending. … ”

¹² “Some people feel that the government in Washington should make every effort to improve the social and economic position of blacks. … Others feel that the government should not make any special effort to help blacks because they should help themselves… .”

¹³ “There is much concern about the rapid rise in medical and hospital costs. Some people feel there should be a government insurance plan which would cover all medical and hospital expenses for everyone. … Others feel that all medical expenses should be paid by individuals through private insurance plans like Blue Cross or other company paid plans… .”

¹⁴ Partisanship is measured using the standard ANES seven-point party identification scale. Ideology is measured by self-placement on a seven-point liberal-conservative scale, with respondents who said that they didn't know or “haven't thought much about” their position (about 28% of the sample) reclassified as moderates. For comparability with the egalitarian values scale, both measures are recoded to range from −1 to +1.

¹⁵ The average values on the 1 (conservative) to 7 (liberal) policy scales are 3.51 for aid to blacks, 3.66 for government jobs, 4.15 for government services, and 4.17 for health insurance. The corresponding averages setting the effects of egalitarian values in the bottom panel of table 4.2 to zero would be 3.22 for aid to blacks, 3.40 for government jobs, 3.96 for government services, and 4.02 for health insurance.

¹⁶ Verba and Orren (1985, 2).

¹⁷ The classic statement of this point is by Converse (1964). Nelson and Kinder (1996) examined how the “group-centrism” of issue preferences can be promoted or inhibited by variations in how issues are framed.

¹⁸ The tax burden questions were each asked twice in the 2002 ANES survey, before and after the election. The pre-election responses are reported in table 4.5, since they seem less likely to be affected by sensitization stemming from the extensive battery of questions focusing on inequality and tax policy in the 2002 ANES survey.

¹⁹ “The Rich, the Poor and the Growing Gap Between Them,” The Economist, June 17, 2006, 28.

²⁰ Kluegel and Smith (1986); McCall (2013).

²¹ Lippmann (1922, 10).

²² McCall (2005, 8).

²³ Hochschild (1981, 279, 278, 283).

²⁴ Respondents were also asked an open-ended version of the same question. Half were asked the fixed-choice questions in the pre-election survey and the open-ended version in the post-election survey; the other half got the open-ended version first and the fixed-choice questions in the post-election survey. The responses to the fixed-choice questions from the two random half-samples were generally similar, except that those who responded in the post-election survey attached somewhat less importance to “government policies”—despite the fact that the intervening survey content called attention to a variety of relevant government tax and spending policies.

²⁵ Osberg and Smeeding (2003, tables 2.1 and 4.2–1).

²⁶ For example, in a 1998 Gallup poll focusing on “Perceptions of Fairness and Opportunity,” 52% of the respondents with household incomes below $15,000 and 51% of those who described themselves as “have-nots” nevertheless said that “the economic situation in the United States is basically fair.” Jost et al. (2003, 55–56, 58, 60). See also Lerner (1980); Jost and Banaji (1994).

²⁷ Jost et al. (2003, 63). On personality and political ideology more generally, see McClosky (1958) and Stenner (2005).

²⁸ These probabilities represent response tendencies for extreme conservatives and extreme liberals with median incomes, and they split the difference in responses between people interviewed in 2002 and 2004.

²⁹ Jost et al. (2003, 63).

³⁰ For evidence regarding the validity of the interviewers’ assessments of respondents’ political information, see Zaller (1985). For respondents interviewed both before and after the 2002 election, I averaged the ratings provided (almost always by different interviewers) at the end of the pre-election and post-election interviews. For respondents who were not reinterviewed after the election (11% of the sample in 2002 and 12% in 2004), I used the pre-election interviewers’ ratings only. In every case, I calibrated the resulting information scale to reflect each respondent's position in the overall distribution of political information, from zero (for the least informed) to one (for the most informed). The actual range is from .014 to .999.

³¹ These percentage differences are calculated for ideological moderates with average incomes. Since the statistical results presented in table 4.10 provide strong evidence of interactions between political information and ideology, the effects of information among liberals and conservatives are quite different from those reported here. These differences are highlighted in figures 4.2, 4.3, and 4.4.

³² Better-informed people also provided systematically different explanations for economic inequality, stressing social causes (inequality in educational opportunities, discrimination, and government policies) more heavily than less-informed people did. They were also somewhat less likely to say that rich people are asked to pay too much in taxes—but no more or less likely to say that poor people are asked to pay too much or that they are asked to pay too much. Nor were they more or less likely to think that corporate accounting scandals are widespread. As for perceptions of the partisan politics of inequality, they were much more likely to recognize the differences in positions of the Democratic and Republican Parties on specific tax policies and much more likely to say that the Republicans are “generally better for rich people” and that the Democrats are “generally better for poor people.”

³³ Among ideological moderates, recognition of the growing income gap increased less dramatically than among liberals, but still significantly: from 74% at the lowest information level to 84% at the highest information level.

³⁴ For extreme conservatives, the information effect implied by the probit analysis is −.606 (with a standard error of .192); for extreme liberals, the effect is +1.336 (with a standard error of .235).

³⁵ Zaller (1992, chaps. 6 and 9).

³⁶ Christopher Achen and I have developed a mathematical model of political inference in which political awareness increases the weight attached to both partisan predispositions and reality in the construction of political judgments (Achen and Bartels 2006). Insofar as the former effect dominates the latter effect, this model implies not only that partisan disparities in political judgment will be concentrated among people who are high in political awareness, as in Zaller's model, but also that people who are generally better informed about politics may be most inaccurate in their political perceptions. Danielle Shani (2006) has presented a good deal of empirical evidence suggesting that partisan biases in perceptions of political conditions are exacerbated by political information. On partisan biases in perceptions more generally, see Bartels (2002a).

³⁷ For extreme conservatives, the information effect implied by the probit analysis is −.566 (with a standard error of .245); for extreme liberals, the effect is +2.022 (with a standard error of .306).

³⁸ The implied information effects in the probit analysis are +1.088 (with a standard error of .421) for extreme liberals and −.149 (with a standard error of .344) for extreme conservatives.

³⁹ The implied information effects in the probit analysis are +.693 (with a standard error of .309) for extreme conservatives and −.993 (with a standard error of .347) for extreme liberals.

⁴⁰ Hochschild (2001); Hochschild and Einstein (2015).

⁴¹ The one instance in table 4.10 in which the main effect of ideology (that is, the effect of ideology among people at the bottom of the distribution of political information) is large enough to be reliably distinguishable from zero is for the factual question about increasing differences in incomes—and that effect suggests, quite implausibly, that uninformed conservatives were more likely than uninformed liberals to recognize that incomes had become more unequal. I suspect that this anomalous result reflects some non-linearity in the effect of political information that is not adequately captured by the simple statistical specification reported in the table.

TABLE 4.1
Public Support for Egalitarian Values
	Agree strongly	Agree somewhat	Neither; DK	Disagree somewhat	Disagree strongly
Our society should do whatever is necessary to make sure that everyone has an equal opportunity to succeed	60.9%	26.8%	5.4%	4.8%	2.0%
If people were treated more equally in this country, we would have many fewer problems	32.4%	33.7%	14.2%	14.2%	5.4%
One of the big problems in this country is that we don't give everyone an equal chance	22.6%	29.9%	13.6%	23.0%	10.8%
It is not really that big a problem if some people have more of a chance in life than others	7.7%	23.9%	19.2%	28.2%	20.9%
We have gone too far in pushing equal rights in this country	15.9%	28.1%	16.0%	20.6%	19.3%
The country would be better off if we worried less about how equal people are	18.6%	28.8%	16.0%	20.7%	15.9%
Source: 1984–2012 ANES surveys. N = 19,357–20,652.

TABLE 4.2:
Egalitarian Values and Policy Preferences
Ordinary least squares regression parameter estimates (with standard errors in parentheses). Policy preferences range from 1 (for the most conservative position) to 7 (for the most liberal position). Observations clustered by year; year-specific intercepts not shown.
	Government jobs	Government services	Aid to blacks	Health insurance
Egalitarian values only
Egalitarian values (−1 to +1)	1.69 (.10)	1.40 (.06)	1.68 (.03)	1.39 (.12)
Standard error of regression	1.69	1.49	1.60	1.86
Adjusted R²	.14	.14	.16	.09
N	14,625	14,883	16,456	11,674
Egalitarian values, ideology, and party identification
Egalitarian values (−1 to +1)	1.23 (.10)	.89 (.06)	1.35 (.05)	.72 (.08)
Liberal ideology (−1 to +1)	.49 (.05)	.53 (.06)	.41 (.02)	.75 (.08)
Democratic party identification (−1 to +1)	.49 (0.3)	.50 (.04)	.31 (.05)	.61 (.05)
Standard error of regression	1.63	1.42	1.57	1.77
Adjusted R²	.20	.22	.19	.18
N	14,183	14,472	16,009	11,220
Source: 1984–2012 ANES surveys.

TABLE 4.3
Average “Feeling Thermometer” Ratings for Social Groups
Ratings range from zero (least favorable) to 100 (most favorable).
	2004	2008	2012	Average
Working-class people	82.6	82.7	83.9	83.1
The military	79.8	79.6	83.2	80.9
Middle-class people	76.8	76.4	77.3	76.8
Poor people	73.4	72.0	71.0	72.1
Business people	69.3	NA	NA	69.3
Catholics	69.0	67.3	66.3	67.5
Conservatives	60.9	60.3	57.6	59.6
The Democratic Party	58.7	57.2	55.3	57.1
Rich people	59.5	57.3	54.2	57.0
Christian fundamentalists	58.8	56.3	53.4	56.2
Feminists	56.2	56.6	54.2	55.7
Labor unions	58.2	55.6	52.5	55.4
Big business	56.1	53.2	53.4	54.2
People on welfare	55.9	54.4	52.3	54.2
Liberals	54.9	54.7	52.1	53.9
Muslims	53.6	50.3	47.9	50.6
Gay men and lesbians	47.5	49.4	53.7	50.2
The Republican Party	53.9	47.7	47.2	49.6
N	1,003–1,178	1,886–2,259	1,777–2,012	—
Source: 2004–2012 ANES surveys.

TABLE 4.4
Class Sympathies and Policy Preferences
Ordinary least squares regression parameter estimates (with standard errors in parentheses). Policy preferences range from 1 (for the most conservative position) to 7 (for the most liberal position). Observations clustered by year; year-specific intercepts not shown.
	Government jobs	Government services	Aid to blacks	Health insurance
Class sympathies (−1 to +1)	1.69 (.14)	1.39 (.34)	.56 (.26)	1.51 (.19)
Egalitarian values (−1 to +1)	.92 (.04)	.66 (.10)	1.07 (.05)	.56 (.04)
Liberal ideology (−1 to +1)	.56 (.05)	.56 (.07)	.34 (.03)	.93 (.06)
Democratic Party identification (−1 to +1)	.55 (.04)	.59 (.10)	.54 (.03)	.57 (.17)
Standard error of regression	1.55	1.35	1.57	1.67
Adjusted R²	.28	.32	.22	.26
N	3,426	3,241	4,076	3,487
Source: 2004–2012 ANES surveys.

TABLE 4.5
Perceived Tax Burdens
“Do you feel you are asked to pay more than you should in federal income taxes, about the right amount, or less than you should? What about rich people? What about poor people?”
	2002	2004	Average
Own tax burden
Pay more than I should	48.3%	39.6%	44.0%
Pay about the right amount	46.2%	51.2%	48.7%
Pay less than I should	3.8%	4.8%	4.3%
Rich people
Pay more than they should	15.3%	10.3%	12.8%
Pay about the right amount	29.9%	24.9%	27.4%
Pay less than they should	53.2%	60.4%	56.8%
Poor people
Pay more than they should	45.8%	46.3%	46.1%
Pay about the right amount	43.7%	42.1%	42.9%
Pay less than they should	8.7%	7.0%	7.8%
N	1,495–1,504	1,210–1,211	—
Source: 2002 and 2004 ANES surveys.

TABLE 4.6
Perceptions of Economic Inequality
“Do you think the difference in incomes between rich people and poor people in the United States today is larger, smaller, or about the same as it was 20 years ago?”
	2002	2004	2008	2012	Average
Much larger	42.2%	47.9%	54.9%	52.5%	49.4%
Somewhat larger	32.3%	31.1%	23.7%	26.8%	28.5%
About the same	16.3%	15.9%	13.9%	10.4%	14.1%
Somewhat smaller	6.0%	2.5%	4.2%	3.6%	4.1%
Much smaller	1.8%	0.3%	1.3%	1.7%	1.3%
Don't know	1.5%	2.2%	2.0%	5.1%	2.7%
N	1,508	1,065	2,099	2,051	—
Source: 2002–2012 ANES surveys.

TABLE 4.7
Assessments of Economic Inequality
“Do you think the difference in incomes between rich people and poor people in the United States today is larger, smaller, or about the same as it was 20 years ago? [If larger or smaller,] do you think this is a good thing, a bad thing, or haven't you thought about it?”
	A good thing	A bad thing	Haven't thought about it; DK	Total
Much larger	2.4%	31.6%	10.6%	44.6%
Somewhat larger	2.1%	12.4%	17.3%	31.9%
About the same	—	—	—	16.1%
Somewhat smaller	1.5%	0.5%	2.5%	4.6%
Much smaller	0.3%	0.6%	0.2%	1.1%
Don't know	—	—	—	1.7%
Total	6.3%	45.2%	30.6%	100%
Source: 2002 and 2004 ANES surveys. N = 2,573.

TABLE 4.8
Explanations for Economic Inequality
“Next, we'd like to know why you think it is, that in America today, some people have better [worse] jobs and higher [lower] incomes than others do. I'm going to read you some possible explanations, and I want you to tell me how important you think each is—very important, somewhat important, or not important at all.”
	Very important	Somewhat important	Not important
Some people don't get a chance to get a good education	54.5%	35.1%	9.1%
Some people just don't work as hard	44.7%	41.4%	12.9%
Some people have more inborn ability to learn	33.8%	42.1%	23.4%
Discrimination holds some people back	26.4%	50.2%	22.5%
Government policies have helped high-income workers more	26.0%	38.2%	33.6%
Some people just choose low-paying jobs	19.4%	38.8%	39.6%
God made people different from one another	22.2%	26.5%	48.5%
Source: 2002 ANES survey. N = 1,427.

TABLE 4.9
Political Ideology and Perceptions of Inequality
Ordered probit parameter estimates (with standard errors in parentheses). Additional response thresholds not shown.
	The income gap has increased	The larger gap is a bad thing^a	The poor don't get a fair trial	Hard work is very important
Conservative ideology (−1 to +1)	−.298 (.049)	−.560 (.062)	−.418 (.077)	.408 (.066)
Family income (0 to 1)	.091 (.081)	.509 (.101)	.240 (.128)	−.201 (.110)
2004 survey	.229 (.044)	.121 (.054)	—	—
Intercept	.639 (.053)	−.066 (.063)	.670 (.072)	−.058 (.064)
Log likelihood	−3,040.4	−1,644.5	−1,010.6	−1,378.8
Pseudo-R²	.01	.03	.02	.01
N	2,525	1,976	1,499	1,414
Source: 2002 and 2004 ANES surveys. ^a Includes only those respondents who said that the difference in incomes between rich people and poor people has increased.

TABLE 4.10
Political Information, Ideology, and Perceptions of Inequality
Ordered probit parameter estimates (with standard errors in parentheses). Additional response thresholds not shown.
	The income gap has increased	The larger gap is a bad thing^a	The poor don't get a fair trial	Hard work is very important
Political information (0 to 1)	.365 (.093)	.728 (.117)	.470 (.165)	−.150 (.138)
Conservative ideology (−1 to +1)	.210 (.113)	.094 (.141)	−.111 (.191)	−.037 (.170)
Political information × ideology	−.971 (.193)	−1.294 (.251)	−.618 (.348)	.843 (.298)
Family income (0 to 1)	.015 (.086)	.347 (.106)	.144 (.135)	−.194 (.116)
2004 survey	.240 (.045)	.142 (.055)	—	—
Intercept	.502 (.062)	−.352 (.077)	.491 (.094)	.014 (.083)
Log likelihood	−3,024.0	−1,618.0	−1,005.8	−1,374.5
Pseudo-R²	.02	.05	.02	.02
N	2,525	1,976	1,499	1,414
Source: 2002 and 2004 ANES surveys. ^a Includes only those respondents who said that the difference in incomes between rich people and poor people has increased.