Chapter 6 • Public Opinion and Ideology

REPRESENTATIVE DEMOCRACY REQUIRES an understanding of public opinion—the collective attitudes of the public, or segments of the public, toward the issues of the day—and how well these preferences are reflected in what government does. Public opinion polls are an ever-present feature of American public life and contemporary journalism. You can scarcely flip on the TV or read a push notification from a media outlet on your phone without being inundated with the latest poll telling you what the American people think about issues ranging from whether people believe the 2020 election was stolen to whether Washington, DC, should become a state to how movements like Black Lives Matter (BLM) and policy proposals like the Green New Deal can be framed to maximize or minimize public support.1 Understanding the nature, measurement, and content of public opinion is crucial to explaining American political behavior. In this chapter, our learning objectives are:

· Understanding how public opinion is measured and interpreted

· Learning about the array of views Americans have on economic, racial, social, defense, and international issues

· Considering the nature of political ideology and whether Americans have a political ideology representing a coherent set of fundamental beliefs or principles about politics that serves as a guide to current political issues, much as partisanship does

· Exploring state- and national-level relationships between leadership and public opinion


The commercial opinion-polling organizations have spent more than eight decades asking Americans about their views on matters of public policy. Most of this investigation has taken one of two forms: (1) asking individuals whether they “approve or disapprove of” or “agree or disagree with” a statement of policy or (2) asking individuals to pick their preference among two or more alternative statements of policy. This form of questioning seriously exaggerates the number of people who hold views on political issues. People can easily say “agree” or “disapprove” in response to a question, even if they know nothing at all about the topic. If given the opportunity, many people will volunteer the information that they hold no views on specific items of public policy. For example, during the 2011 showdown between President Barack Obama and congressional Republicans over raising the debt ceiling, about 35 percent of those polled by Gallup said they did not know enough about the issue to have an opinion.2 In contrast, in the past several decades, less than 5 percent of all adults had no opinion on issues such as abortion or the death penalty. Other issues, like the North American Free Trade Agreement, have gone from 16 percent of the public having no opinion in 1997 to 6 percent having no opinion in 2017—shortly after Donald Trump raised the salience of trade issues on the campaign trail.3 More typically in recent years, approximately 10 percent of the electorate has had no opinion on major issues of public policy. Philip E. Converse has shown, in addition, that a number of those individuals who appear to have an opinion may be regarded as responding to policy questions at random.4 One reason for this is that many people form opinions during a telephone survey off the top of their heads, usually based on the most recent or salient information about the issue they can immediately recall.5 Indeed, people are more likely to be uninformed, rather than misinformed, about major issues in American politics.6

The lack of opinion and information on topics of public policy can be explained in several ways. In general, the factors that explain nonvoting also account for the absence of opinions. Individuals who are not very interested in or concerned about politics are the least likely to have opinions on matters of public policy. Beyond this basic relationship, those with lower socioeconomic status are associated with a lack of opinion on issues. Low income and little education create social circumstances in which individuals are less likely to be exposed to and learn the information they need to develop clear views on public policies.

Of course, not all issues are equal. Some issues of public policy, such as abortion or the death penalty, are relatively easy to understand; others may be much more difficult, requiring individuals to face complex considerations. Edward G. Carmines and James A. Stimson have argued that different segments of the public respond to “hard” issues that involve calculation of policy benefits and “easy” issues that call for symbolic, “gut responses.” Relatively unsophisticated, uninterested members of the electorate respond to easy issues; the more sophisticated, most interested citizens are more likely to take positions on hard issues and are just as likely to take positions on easy ones.7

Partisan labels also affect people’s opinions. Paul Sniderman and John Bullock found that adding words like “Democrats say” or “Republicans say” to an issue position increased individual ideological constraint.8 That is, liberals were more likely to agree with a position when it carried a Democratic Party label, and conservatives were more likely to favor preferences with a Republican Party endorsement.

It is no simple matter to describe the distribution of opinions in the American electorate because no obvious, widely accepted method has been established to measure these opinions. Asking different questions in public opinion polls will elicit different answers. Even on the issue of abortion, on which most people have views, the distribution of opinions can be substantially altered by asking respondents whether they approve of “killing unborn children” as opposed to “letting women have control over their own bodies.” Furthermore, unlike reports of voting behavior, no direct means exist to validate measures of opinions. As a consequence, descriptions of public opinion must be taken as more uncertain and more tentative than those drawn from the discussion of partisanship.

The mode of public opinion gathering can also affect the distribution of opinions. The results from surveys asking Americans for whom they planned to vote in the presidential election varied substantially, based on whether the survey was a phone survey using random-digit dialing of landlines, included cell phone numbers, was conducted on the internet, or was administered face-to-face. Now that more than half of American homes do not have landlines,9 survey researchers who rely on phone calls mix an increasing percentage of cell phone numbers in the samples.10 This is especially critical for capturing younger people in surveys. The American National Election Studies (ANES), which provides most of the data used in this book, used a mix of video, phone, and internet sampling in 2020.11

Beyond the mode of the survey, nonresponse to public opinion polling can skew results. The Elections Research Center (ERC) at the University of Wisconsin–Madison conducted a multiwave panel survey in 2020 in several swing states. In each wave, the ERC tried to reinterview the same people interviewed in the first wave, but of course, some people refused to participate again. This is called attrition. Working with the survey firm YouGov, the ERC replaced those who did not participate in multiple waves of the survey with new, carefully sampled respondents. Interestingly, the first survey the ERC conducted, back in February 2020,12 was the most accurate prediction of the election results in the swing states. One reason for this could be that as the election got closer, and President Trump looked like he would lose, his supporters became less willing to participate in surveys, making it look like he was performing worse than he actually was. Another potential explanation is that survey firms were systematically undersampling a key group of Trump supporters so that their intent to vote for the president was not captured in surveys. The Pew Research Center has studied whether polls that significantly overestimated Joe Biden’s vote in a state are a sign that polls about issues are similarly skewed, so far concluding that they are not.13

After public opinion is measured, it needs to be interpreted. When reading results from a public opinion survey, it is important to look for the nature of the sample (e.g., adults, registered voters, or likely voters) and the size of the sample. A good rule of thumb is that a national U.S. sample of one thousand adults will have a ±4 percent margin of error.14 This means that if the survey firm conducted the poll one hundred times, the results would be within 4 percentage points of the survey result ninety-five out of one hundred times. So, if a survey reported that 53 percent of people favored a policy with a margin of error of 3 percent, we could be very confident that the true level of support for the whole population would be between 50 and 56 percent. With this knowledge in mind about how opinions are measured and how polls are interpreted, let us move on to describing some contours of American political opinion.


Domestic Economic Issues

The liberal–conservative ideological divide related to the government’s role in managing the economy has dominated partisan politics since 1932. A little more than a month after he took office, major economic problems caused by the COVID-19 pandemic led President Biden to advocate for and sign a $1.9 trillion relief package that received no Republican votes, highlighting this ideological divide. The collapse of the financial sector in the fall of 2008 and the actions of both the Bush administration and the incoming Obama administration to stimulate the economy and to bail out and then reregulate banks and insurance companies brought cries of “socialism” from conservative politicians and commentators. Highly partisan debate over President Obama’s health care plan reminded Democrats of the failure of the Clinton administration’s health care plan in 1993. This failure helped usher in the loss of the House of Representatives to the Republicans in 1994 and promises of a “conservative revolution” that would reduce government involvement in the economy and cut support for various social programs. In similar fashion, Ronald Reagan’s victory in 1980 and his administration’s subsequent cutting of taxes and social programs was portrayed as a reversal of fifty years of economic liberalism. Elections have consequences, and in policy terms, there have certainly been consequences of Republican or Democratic victories. In terms of public opinion, however, broad and continuing public support remains for many governmental initiatives, regardless of election outcomes. As George W. Bush discovered when he proposed privatizing Social Security after winning reelection in 2004, long-standing programs that appear to benefit “deserving” segments of the population are difficult to reform because they enjoy widespread support.

The responses to public opinion questions, and public opinion itself, can be affected by political rhetoric and election slogans. For example, the General Social Survey asks a long series of questions on whether spending on various programs is “too much, not enough, or about right.” Over the years, sizable proportions of the public have said that too much is being spent on “welfare.” At the same time, even larger proportions have said not enough is being spent on “assistance to the poor.”15 Clearly, years of racist anecdotes about welfare queens and promises to end welfare as we know it have had their effect on the way particular programs are perceived, if not on the public’s general willingness to use government as an instrument for social purposes.

Figure 6-1 shows the distribution of attitudes toward spending for different governmental purposes, using the 2020 ANES. It is clear that there are major partisan differences across spending preferences for almost every issue. Majorities of Democrats favor increasing spending on social security, public schools, welfare programs, highway maintenance, aiding the poor, and protecting the environment. Majorities of Republicans prefer more spending on border security, public schools, dealing with crime, and highway maintenance.

On some issues, the disagreements between the parties are about whether spending should be increased or decreased. For example, Democrats want more spending on welfare programs while Republicans want less. However, most differences are between a spending change and the status quo. Republicans prefer more spending on dealing with crime and on border security while a plurality of Democrats prefer spending to stay as it is on those issues. More than 60 percent of Democrats want more social security spending whereas Republicans were evenly divided between preferring more spending and preferring the same level of spending.

Overall, Figure 6-1 shows fairly widespread willingness to support government spending on domestic social programs, something that has been true for the past several decades. Importantly, many government benefits that come from spending are hidden from public view. Suzanne Mettler’s The Submerged State shows that most Americans are unaware of the many government benefits they receive and are even less aware that the affluent are the most likely to benefit from these programs.16

Eight horizontal stacked graphs show the attitude of Democrats, Independents, and Republicans toward domestic spending in 2020.Description

Figure 6-1 Attitudes Toward Domestic Spending, 2020

Source: American National Election Studies, available at

Perhaps not surprisingly, examining how different social groups think about major political issues reveals some fairly dramatic differences in what people want from the government. Figure 6-2 is based on a question in the 2020 ANES that offers respondents a choice between “cutting spending and decreasing services” and “increasing services and increasing spending.” We look at responses to this question for several social categories, using race and ethnicity, education, and religion as variables. The pattern in the figure is not difficult to discern. The least economically secure—Blacks, Hispanics, and whites with no religious preferences and some or no college—support government services most strongly. The groups most distinctively in favor of cutting spending and services are white Protestants and Catholics with no college or some college.

Two bar graphs of attitudes toward spending and services by race, ethnicity, religion, and education in 2020.Description

Figure 6-2 Attitudes Toward Cutting Spending Versus Increasing Government Services, by Race, Ethnicity, Religion, and Education, 2020

Source: 2020 American National Election Studies, available at

Favoring services over spending cuts represents the type of choice in governmental policy that characterized the New Deal. Thus, it would be reasonable to expect a dramatic difference between Democrats and Republicans on such an issue. Economic issues have divided Democrats and Republicans since the 1930s, whereas other issues have been of only temporary significance for the parties. As a consequence, the relationship in Table 6-1 showing that Democrats disproportionately favor increased government services and Republicans prefer cuts is no surprise. Strong Democrats favor increased services over a reduction in spending by a margin of 75 percent to 3 percent; strong Republicans are the opposite, favoring a reduction in spending over increased services by a margin of 48 percent to 15 percent. This basic pattern has existed for decades. While becoming steadily more polarized, Republican views became less polarized in 2020, perhaps due to the COVID-19 pandemic.

Rising health care costs have been a concern for years. Health care was the dominant issue in the first years of the Obama administration. Reforming the health care system was also attempted early in the Clinton administration, when President Bill Clinton appointed First Lady Hillary Rodham Clinton to head a group to develop a proposal for a national health care program. The managed competition program eventually proposed was not the national health care program favored by the most liberal advocates, but the ensuing debate was cast in terms of governmentally mandated and regulated programs versus private insurance companies with individual choice in health care providers. Ultimately, the Clinton administration lost the battle in Congress and in the arena of public opinion. Initially in 1992, the public favored a governmental insurance plan, 44 percent to 24 percent, over private insurance plans.17 After health insurance and pharmaceutical companies and the American Medical Association launched an extensive advertising campaign featuring a middle American couple, “Harry and Louise,” worrying about the government taking away their choice of doctor, the Clinton program went down in defeat. By 1996, public sentiment had reversed, with 40 percent of the ANES sample saying they thought medical expenses should be paid by private insurance plans and 34 percent opting for a governmental plan. In the 2008 campaign, health care was again an issue, and the public was again tilting in favor of a governmental program, with 48 percent in favor of a governmental plan and 34 percent preferring private health insurance.18

As the debate in Congress geared up, Harry and Louise returned to the airwaves, sixteen years older and this time supporting a health care reform package. After the successful passage of the Affordable Care Act (ACA), known as Obamacare by political opponents and the news media, health care was again front and center in the 2012, 2016, and 2020 presidential elections. After President Trump was elected, House Republicans attempted to repeal the ACA more than seventy times. In the Senate, Democrats closed ranks, preventing the bill from being repealed. Health care reform remained a major issue in the 2020 campaign with Trump continuing to promise to eliminate the ACA if reelected, even though he had not been able to do so in his first term.

While the law contains many popular provisions, opinion of the entirety of the act was more negative than positive until the possibility of repeal became more serious. In 2021, the ACA is slightly more popular than not.19 Table 6-2 shows the durability of partisan divisions on health care issues as Democrats continued to favor government-sponsored health insurance and Republicans strongly favored private health insurance in 2020. Interestingly, after Trump’s victory, when the possibility of the ACA being overturned was suddenly real, the legislation became much more popular. Strong Republican support of government health insurance doubled from 2016 to 2020.

Table 6-1

Source: 2020 American National Election Studies, available at

Table 6-2

Source: 2020 American National Election Studies, available at

Racial Justice Issues

Race and attitudes associated with race hold a prominent place in American political history. For many years after Reconstruction, little overt public attention was paid to racial issues. The South was allowed to impose its system of segregation on its Black population by law, while informal, de facto methods created much the same system of separate neighborhoods and segregated schools in the North. After integration became a major national and international focal point of attention in the 1940s and 1950s, a number of significant developments in the political attitudes of the public occurred. First, Southern Blacks have become an organized, involved, politically motivated group. Removing the legal barriers to voting in the South enabled Black voters in Southern states to command the attention of politicians at the ballot box and in state legislatures and governors’ mansions.

Second, large numbers of Southern whites have adjusted their opinions to accept the realities of the new legal and political position of Black Americans. The public, as a whole, has come to support the general principle of racial equality. The recent near unanimity on this point means that the vast majority of Americans no longer express support for policies and practices that explicitly discriminate against racial groups, and making political appeals based on blatant racism is no longer consistent with the dominant political culture.

At the same time, the public has not moved significantly closer to supporting governmental programs designed to improve the economic and social position of racial groups. Northern as well as Southern whites have consistently opposed busing for the purposes of integration. More than 80 percent of whites oppose affirmative action on behalf of racial minorities, and less than half support the federal government enforcing fair employment practices. The proportion of the public supporting various forms of governmental action to aid Blacks is shown in Table 6-3. In contrast to the near unanimity of support for “letting [B]lack and white children go to school together,” less than half of the public supports positive actions on the part of government to improve the social and economic position of Blacks.

The lack of connection between broad principle and policy implementation has been the focus of both political debate and scholarly disagreement. One side argues that opposition to programs to aid Blacks is based, in general, on opposition to governmental activities and, in particular, on programs that benefit a subgroup of society.20 This position, often referred to as “racial conservatism,” is seen as stemming from a general philosophical commitment to limited government and a belief in individualism. The attitudes, it is argued, are based on principles, not racism.

The other view argues that opposition to policy proposals to use governmental programs to aid the social and economic circumstances of Blacks and other underrepresented groups stems from racial hostility, even though racial conservatives may have learned to cloak their racism in acceptable philosophical language. To complicate the matter further, scholars take different views of how racial hostility expresses itself in political attitudes. Scholars have used three dimensions of racial hostility to explain white support for or opposition to governmental policies regarding race:

1. Racial resentment (the feeling that Blacks are getting more than they deserve) or racial disapproval (the feeling that Blacks do not live up to certain value expectations, such as working hard).21 These contentions are often referred to as symbolic racism.

2. Group conflict (zero-sum conflicts over scarce resources).22

3. Social dominance (protection of the status quo by a dominant group).23

These dimensions of racial hostility can be interrelated and may reinforce one another. It is difficult to separate them or to be confident in measuring them or evaluating which dimension contributes the most to racial attitudes. Measuring racial attitudes is especially tricky because social norms in twenty-first-century America dictate against making explicitly racist statements and people want to give answers to survey interviewers that are socially desirable. Interpreting racial attitudes can also be tricky. Although these dimensions are often labeled as if they were positive or negative in content, they have both pro-Black and anti-Black extremes. In other words, if a Black person believes strongly that Black people are not getting what they deserve, racial resentment may be involved just as much as when a white person believes that Blacks are getting more than they deserve.

To connect these dimensions with attitudes about policies designed to provide governmental aid to underrepresented groups, it seems reasonable to assume that people must view potential beneficiaries of governmental aid as deserving. The extent to which Blacks and other people of color are viewed as deserving by white Americans may depend on whether Blacks are seen as individually responsible for their position or whether they are seen as victims of social and economic forces beyond their control. Presumably, whites who believe that Blacks can improve their situation through their own efforts will not view them as deserving of special governmental programs on their behalf. This basis of opposition would fit the symbolic racism perspective. Whites who see social structures and conditions imposing special hardships on Blacks regardless of their individual efforts will view Blacks as deserving of special assistance.

Even if Blacks are viewed by whites as deserving of government assistance, special programs may be opposed if whites see these programs as coming at the expense of whites. Another similar basis for opposition to programs for Blacks is the expectation that the status quo, which favors whites, will be disrupted, which would be undesirable from the point of view of whites. These objections are examples of the group conflict and social dominance perspectives.

A racial conservative might make the argument that there was once a time when all the relevant democratic principles were on the side of helping Blacks but that more recently such principles work both ways. Blacks should have an equal chance to get an education, find a job, and so forth, but they should not be given advantages over other deserving people. Indeed, this was the argument that the majority of the Supreme Court used in its 2013 decision overturning key provisions of the Voting Rights Act that had required federal approval when states with a history of racial discrimination wanted to change voting procedures. However, great differences are found in the perceptions of Blacks and whites about whether or not Blacks have an equal chance in American society.

Racial attitudes have had a profound effect on the American political landscape. In their book Issue Evolution, Carmines and Stimson argue that an evolution of the racial issue since the early 1960s has led increasingly to the Democratic Party being perceived as the liberal party on civil rights issues and the Republican Party being perceived as the conservative party.24 They see this distinction as the dominant perception of the parties in the eyes of the public, representing a fundamental redefinition of the issue alignment that has characterized the parties since the New Deal.

Before the 1960s, Republicans (the party of Abraham Lincoln) were seen as more progressive on civil rights than Democrats, particularly in light of the strongly segregationist cast to the Southern wing of the Democratic Party. Carmines and Stimson show that a change occurred during the 1960s and 1970s, when the elites of the two parties—members of Congress and presidential candidates—as well as party activists became distinctive in their racial views. The Democratic Party became dominated by northern liberals advocating stronger governmental action to ensure equal rights. At the same time, the leadership of the Republican Party became racially conservative—that is, opposed to government intervention to ensure equal rights for people of color. As the elites and activists sorted themselves into distinct groups on the basis of their attitudes toward racial issues, the perceptions that the mass public held of the parties followed suit. Increasingly through the late 1960s and 1970s, Carmines and Stimson argue, the partisan choices of individual citizens fell in line with their attitudes on racial questions.

The role of race and racial issues in American politics is not always easy to trace, however. Because certain issues that are not explicitly stated in terms of race are nevertheless symbols of race in the minds of some people, candidates can make appeals based on racist attitudes without using racial language. For example, “law and order” may mean “keeping Blacks in their place” to some, and “welfare” may carry racial overtones.25 Furthermore, the lack of support among whites for policies that target assistance to Blacks gives both parties an incentive to avoid embracing such policies, according to Donald R. Kinder and Lynn M. Sanders.26 Republican leaders can oppose these policies and win support from their overwhelmingly white constituency, particularly Southern whites. But Democratic leaders also have an incentive to avoid endorsing policies that would help Blacks, so as not to alienate white support.

If race is not an issue to be openly discussed in political campaigns, then uncovering the political significance of race in people’s attitudes and perceptions of the political parties becomes difficult. On the whole, straightforward efforts to capture distinctive party images along racial lines do not succeed. Although more than half of Americans believe there are differences in what the parties stand for, typically only a small percentage characterize the differences in racial terms. Overwhelmingly, when people articulate differences between the parties, it is in terms of symbols and issues associated with the New Deal realignment. This in all likelihood reflects the lack of overt discussion of racial issues by the political leadership of either party. In fact, Paul Kellstedt’s book The Mass Media and the Dynamics of Racial Attitudes has shown that shifts in mass opinion on racial attitudes have begun to mirror shifts in economic attitudes, suggesting that ideological opinions about race have “fused” onto economic preferences.27

In 2008, with the first Black presidential candidate nominated by a major party, race was an issue whether anyone talked about it or not. From Bill Clinton’s remarks downplaying a Black candidate’s win in the South Carolina primary, to the Reverend Jeremiah Wright’s videotaped sermons, to Obama’s March 18 speech on race, to the increased Black turnout and overwhelming support for Obama on Election Day in 2008 and 2012, race was an often unspoken but constant presence throughout the campaign. In retrospect, it needs to be remembered that Obama did not have the unanimous support of Black Democratic activists in the primaries. Former president Clinton was highly popular among Blacks during and after his presidency, and Hillary Clinton benefited from that association. And early on, Obama was seen in some circles as “not Black enough,” given his mixed racial heritage and his upbringing by his white mother and grandparents. Once nominated, any Democratic presidential candidate can count on around 90 percent of the Black vote in the November election. After Obama’s nomination, he received not only virtually unanimous support from Black voters but high enthusiasm and turnout as well in both 2008 and 2012. In their book Steadfast Democrats: How Social Forces Shape Black Political Behavior, Ismail White and Chryl Laird highlight how social pressure, built through strong social bonds of survival and resistance, is a critical element of Black Americans’ electoral support for Democratic Party candidates.28

While explicit examples of race negatively affecting Obama’s candidacy were few and far between, research examining people’s implicit, or “automatic,” responses to race showed that those with an implicit preference for whites over Blacks were more likely to support John McCain in 2008 and Mitt Romney in 2012, highlighting how complicated it is to untangle how racial attitudes affect public preferences and voting behavior. However, in the 2016 presidential elections, attitudes about race were a significant predictor of vote choice. An analysis of public opinion data led by Brian Schaffner showed that those individuals who thought racist incidents in the United States were few and far between were far more likely to vote for Trump than were those who thought racial incidents were a regular problem in the country.29

In 2020, the murder of George Floyd by a Minneapolis police officer sparked massive protests around the country, giving new energy to the BLM movement. As the protests grew, news coverage focused on what mass communication scholar Doug McLeod and colleagues have called the protest paradigm30—focusing on political elites as compared to ordinary citizens, framing coverage episodically instead of digging for the root causes of systemic racism, framing protesters as engaged in destructive action while police are framed as engaging in restoring order, and characterizing the movement based on the actions of a few.

Public opinion consequences of this kind of coverage can be inferred by looking at Table 6-3, which compares support for the BLM movement in June, shortly after Floyd was killed, and September, closer to Election Day. While support for the movement from Black people was steady, it dropped among every other group, though white Democrats continued to support the movement at a slightly higher level than Blacks.

Table 6-3

Source: Deja Thomas and Juliana Menasce Horowitz, “Support for Black Lives Matter Has Decreased Since June but Remains Strong Among Black Americans,” Pew Research Center, September 16, 2020,

Social Issues

One of the more emotional aspects of the polarization of the two political parties over the past two decades has been conflict over so-called traditional values. Issues such as abortion, gay rights, transgender rights, pornography, and sex education and prayer in the public schools have risen in prominence over the past few decades. Common wisdom positions Democrats on the liberal side and Republicans on the conservative side of these “culture wars,” as they are often called, but significant numbers of leaders and followers in both parties are not so easily placed. Like the process of sorting that occurred over racial issues in the 1960s and 1970s, a similar sorting over traditional values has been occurring more recently. The Republican Party, especially, has had a difficult time portraying itself as a “big tent” that welcomes a wide range of people with varying beliefs, but people who are pro-life or opposed to same-sex marriage have also come to feel uncomfortable in the Democratic Party.

Certainly, one of the most potent of these social issues is abortion. The 1973 Supreme Court decision in Roe v. Wade immediately generated a polarized response, turning election races in some areas into one-issue campaigns. Forty years later, intact dilation and extraction, far more commonly known as “partial birth” abortion, is a hot-button issue, and both sides in the abortion debate gird for battle over any anticipated retirement from the U.S. Supreme Court.

One of the difficulties in examining public opinion on abortion lies in the responses that different question wording elicits. Although responses to the same question are similar over time, different phrasing of questions produces differing proportions of “pro-choice” and “pro-life” answers. In the following analysis, we use data from the ANES, which has used the same question in each survey since 1980.

The public’s views on abortion are associated with several personal characteristics, most notably age, education, and religion. No matter what combination of characteristics is examined in the general public, invariably more than half of the people in the ANES surveys support the right to abortion under at least some circumstances. The two response choices at the pro-life end of the continuum are that abortion should never be permitted and that abortion should be permitted only in cases of rape or incest or to save the life of the mother. The first is a more extreme position than many pro-life advocates would take; the second includes circumstances that have been explicitly rejected by pro-life advocates in and out of Congress. The most extreme pro-choice alternative offered is that by law a woman should always be able to obtain an abortion as a matter of personal choice. We have used the most extreme category at either end of the continuum to indicate pro-life and pro-choice positions.

In simple terms, older people are generally more likely to be pro-life than younger people, and the less educated are less supportive of legal abortion than are the better educated. Given the frequent labeling of abortion as a women’s issue, it is worth noting that there is relatively little difference in the views of women and men, although women are more likely than men to take the extreme positions on both the pro-choice and pro-life sides.

Perhaps surprisingly, given the Roman Catholic Church’s clear position in opposition to abortion, little difference is found between Catholics and Protestants in their positions on abortion. This is true even when the frequency of church attendance is taken into account. Figure 6-3 shows the percentages taking the most extreme pro-life position and the most extreme pro-choice position for Catholics and Protestants with different frequencies of church attendance. Among both Catholics and Protestants, opposition to abortion declines as church attendance declines, but the percentages expressing pro-choice and pro-life sentiments are quite similar for the Catholic and Protestant groups. Regular churchgoers among Catholics are about as pro-life as Protestant regular churchgoers. Catholic occasional attenders are more likely to be pro-choice than Protestants who attend church infrequently. Underlying this change is another trend—a smaller proportion of all Catholics claim to be regular churchgoers. Regardless of their religious denomination, the figure shows that those who are faithful churchgoers are more pro-life.

Two line graphs of the attitudes of non-Catholics and Catholics toward abortion by the frequency of church attendance in 2020.Description

Figure 6-3 Attitudes Toward Abortion Among Catholics and Protestants, by Frequency of Church Attendance, 2020

Source: 2020 American National Election Studies, available at

Beyond the policy preferences individuals have on abortion is the level of relative importance they place on the issue. If abortion was an issue where people had clear opinions, but the opinions were not very important to them, it would not be a very consequential issue in American politics. Table 6-4 shows that nearly 62 percent of Democrats say abortion is very or extremely important while more than 56 percent of Republicans feel the same way.

Table 6-4

Source: 2020 American National Election Studies, available at

Perhaps not surprisingly, then, the public does a pretty good job of assessing candidate positions on the abortion issue. Table 6-5 shows that strong majorities of Democrats, independents, and Republicans placed Biden’s position solidly on the pro-choice end of the spectrum. Table 6-6 reveals more of a partisan difference with the placement of Trump’s position on abortion, with Democrats placing the president at the most extreme pro-life end of the spectrum, perhaps on the basis of the believed opinions about abortion that are held by Trump’s picks for the Supreme Court. Republicans place Trump on the pro-life end of the spectrum, but less than a quarter of Republicans thought that Trump was a pro-life evangelist.

Table 6-5

Source: 2020 American National Election Studies, available at

Table 6-6

Source: 2020 American National Election Studies, available at

Their position on abortion has become a litmus test for presidential hopefuls in both parties ever since the 1970s. Not surprisingly, then, a fairly strong relationship exists between partisanship and views on abortion, as can be seen in Table 6-7. Perhaps more unexpected is that the relationship is not stronger. Sizable minorities of Republicans and some Democrats take positions contrary to the stand of their party. This is one of several ways in which the polarization of the political activists and elites of the parties is not reflected in the rank and file.

Table 6-7

Source: 2020 American National Election Studies, available at

Note: The full text of choices is as follows:

1. By law, abortion should never be permitted.

2. The law should permit abortion only in case of rape, [in case of] incest, or when the woman’s life is in danger.

3. The law should permit abortion for reasons other than rape, incest, or danger to the woman’s life, but only after the need for the abortion has been clearly established.

4. By law, a woman should always be able to obtain an abortion as a matter of personal choice.

The issue of same-sex marriage was very contentious in the early 2000s, with some states banning same-sex marriage through legislative actions or referenda and others moving toward legalizing it. In 2004, the issue was suddenly injected into the presidential campaign when the Supreme Court of Massachusetts ruled that denying gay and lesbian people the right to marry was unconstitutional in that state, and when the mayor of San Francisco began issuing marriage licenses to same-sex couples. This was a no-win situation for the Democrats and their presidential candidate, given that the American public opposed same-sex marriage by a two-to-one margin at the time. Because proposed bans on same-sex marriage were on the ballot in several states that year, the issue also served to energize conservative voters in safe Republican states, thus raising President Bush’s popular vote margin. President Obama famously “evolved” to a position of support for same-sex marriage, following growing public sentiment favoring such a position. In 2013, the Supreme Court ruled that the federal government could not define marriage as being between one man and one woman, striking down the Defense of Marriage Act signed by former president Clinton.

Over time, the American public has become more tolerant of the idea of same-sex marriage or, at least, civil unions for same-sex couples. This is another social issue that has a fairly strong relationship with partisanship, as can be seen in Table 6-8. Democratic partisans are more accepting of the idea than Republicans, although a substantial minority of Democrats oppose it. Since 2008, strong Republicans’ support for same-sex marriage has grown from 9 percent to 40 percent, and strong Democrats’ support is up to 85 percent, from 46 percent in 2008. Weak Republicans’ support was at 63 percent in 2020, and weak Democrats’ support of same-sex marriage jumped to 84 percent, 17 points higher than in 2016.

Table 6-8

Source: 2020 American National Election Studies, available at

Homeland Security, Terrorism, and International Affairs

After September 11, 2001, terrorism dominated the front page for years.31 Not only have the events of that day been seared into people’s memories, but the fear of future terrorist attacks remains pervasive also—although concern is ebbing a little as time passes without another attack. In the fall of 2011, a CBS News poll found that 83 percent of the public believed that the threat of terrorism will always exist, though more than half also believed that an attack was not likely in the next few months.32

A wide-ranging survey in 2004 on Americans’ perceptions of the threat of terrorism showed that concerns about a terrorist attack focus on chemical and biological weapons as the most worrisome.33 Terrorism is seen as multifaceted, taking many possible forms with a wide array of possible targets. Even without another attack, many years will pass before substantial numbers of the public have no personal memory of September 11, so the feelings and issues surrounding terrorism will be around for a long time.

Terrorism is a valence issue—that is, one in which virtually every citizen of the United States agrees that it is bad—as opposed to a position issue, which some support and others oppose. At one level, terrorism is also an easy issue, in the sense that Carmines and Stimson use the term. It is easy, on a gut level, to understand what happened on September 11 and to find it abhorrent. Beyond this, however, the terrorist threat becomes more complicated—a hard issue—as one tries to imagine who might be terrorists, what motivates them, and what array of possible weapons they might use. Even more difficult is assessing the steps proposed and taken to thwart terrorists. Do they work? Are they cost-effective? How would one know?

More than most other international issues, the threat of terrorism involves domestic policies aimed at preventing terrorist attacks, such as security searches at airports and on mass transit systems. Most initial survey research in this area has focused on the public’s reaction to the USA PATRIOT Act. Among the roughly 80 percent of the public who had heard of the act, opinion was fairly evenly divided between those who saw it as a necessary tool for finding terrorists and those who thought it went too far and threatened civil liberties.34 These responses were strongly influenced by partisanship. Only 15 percent of Republicans thought it went too far, while 53 percent of Democrats held that view when George W. Bush was president.

Up to some unknown point, Americans are willing to see their civil liberties compromised if they believe doing so will help prevent terrorism, even though three-fourths of the public has little confidence that the government will use personal information, gained through antiterror measures, appropriately.35 Respondents are more willing to sacrifice the civil liberties of others than their own. For example, a CBS News poll in the spring of 2005 found that 56 percent of the public was willing to have governmental agencies monitor the telephone calls and emails of Americans the government finds suspicious; only 29 percent favored allowing this monitoring of “ordinary Americans.”36 These views are driven by partisanship as well. In 2013, with Democrat Barack Obama serving as commander in chief, 49 percent of Republicans supported the federal government’s monitoring of Americans’ phone calls, while 58 percent of Democrats did.

Typically, important valence issues, such as terrorism, have components that can be treated as position issues when segments of the public and their leaders hold differing views. So while everyone is opposed to terrorism, dissimilar views can be held about how well President Bush handled the war on terrorism or whether the Iraq war made the United States more or less safe from terrorism. As might be expected, the positions people take on these matters are related to partisanship.

Surprisingly, then, in 2020, worries that a terrorist attack might occur in the near future were surprisingly consistent across one’s location on the party identification scale. Figure 6-4 shows that independents were the least worried group with almost 49 percent slightly or not at all worried. Strong Republicans were the most worried group, but their numbers were almost identical to those of strong Democrats.

A clustered horizontal bar graph of the concern of Americans about a terrorist attach in the near future by political affiliation.Description

Figure 6-4 How Worried Are Americans About a Terrorist Attack in the Near Future?

Source: 2020 American National Election Studies, available at

During the Cold War years (around 1947–1991), foreign policy—and the study of public opinion about foreign policy—focused on relations between the United States and the Soviet Union. Questions centered on the relative strength of the two countries, the likelihood of nuclear conflict between them, and the preference for negotiation or military strength as a strategy for keeping the peace. With the breakup of the Soviet Union, the focus of foreign policy has shifted away from superpower military relations toward involvement, or noninvolvement, in trouble spots around the world, such as the Persian Gulf in 1990 and 1991; Afghanistan and Iraq since 2001 and 2003, respectively; Libya in 2011; the ongoing violence in the Sudan; and the political disarray in Egypt and Syria.

Issues of foreign affairs vary greatly in salience, particularly in response to involvement of the nation in a military conflict. In analyzing the public’s attitudes toward international events, a distinction needs to be made between brief conflicts and longer-lasting wars. A truism now in American politics says the American public will support—usually enthusiastically—brief military involvement in a foreign conflict that seems to be successful (and they will quickly forget if it is not successful). If, however, combat drags on, support typically diminishes, especially if there are significant U.S. casualties. In the case of Vietnam, opposition to the war developed at a sluggish pace over a considerable period of time, as did recognition of the seriousness of U.S. involvement. In the early years, opposition to the war (as measured by support for a prompt withdrawal) was low—less than 10 percent. By 1968, it had grown to 20 percent; by 1970, it was more than 30 percent; and by 1972, it had reached more than 40 percent. Clearly, as the war dragged on, support for military involvement declined dramatically. The Korean War more than seventy years ago offers similar evidence of the failure of public support for prolonged conflicts.

In contrast, military episodes that develop rapidly command great public attention and almost always garner public support. In the Persian Gulf War, deployment of American military forces to Saudi Arabia began in the late summer of 1990, in response to Iraq’s invasion of Kuwait. In the late fall, as deployment of U.S. troops continued, the public was somewhat divided in its support of this policy. Fifty-nine percent of the public polled said they believed it was the correct policy to pursue, and 39 percent said they believed it was not. These views shifted fairly dramatically after the brief and well-televised war with Iraq in early 1991. When the same people were interviewed again in June 1991 following the war, 81 percent felt the war had been the right thing to do, and only 18 percent thought the United States should have stayed out.

In the beginning, support for the second war in Iraq was high among the American public. In March 2003, immediately after the invasion, 69 percent of the public thought the United States had done the right thing in taking military action against Iraq. Only a quarter of the public thought the United States should have stayed out.37 A little over a year later, on the anniversary of the declared end of formal hostilities, the public was evenly divided on whether or not the United States had done the right thing—47 percent thought so; 46 percent did not.38 Small bursts of support appeared after the patriotic displays at both parties’ national nominating conventions in the summer of 2004 and right before the presidential election in November 2004, but generally, the trend of support was downward. By the fall of 2008, 70 percent of the public disapproved of President Bush’s handling of the war in Iraq.39 Political analysts have long noted a rally-around-the-flag phenomenon that occurs at times of international crisis.40 Presidents invariably get a boost in popularity ratings in the polls in the midst of an international incident, even when the actions of the administration are not particularly successful. John F. Kennedy got such a boost after the disastrous Bay of Pigs invasion of Cuba in 1961, as did Jimmy Carter—temporarily—after a militant student group seized the American embassy in Iran in 1979. And public approval of President George H. W. Bush’s handling of the Persian Gulf War was extremely high during and immediately following the conflict, representing a substantial increase over his earlier ratings. Similarly, George W. Bush saw an immediate gain in presidential job approval after the attacks on the World Trade Center and Pentagon on September 11, 2001. ABC News and the Washington Post completed a national survey the weekend before the attacks in which Bush’s job approval was at 55 percent, about average for modern presidents in the first year of their first term. Immediately after the disaster, his job approval climbed to 86 percent. The change in “strong” approval was from 26 to 63 percent.41 This enthusiasm did not survive the passage of time as President Bush left office with approval ratings that rivaled the worst evaluations of presidential performance in modern polling history.

As with domestic issues, partisan differences are apparent, with partisans of the president in office more supportive of whatever actions are taken than are partisans of the party out of power. This was seen most clearly during the war in Vietnam. Republicans were more likely to think getting involved in Vietnam was a mistake before 1969, when Democratic president Lyndon B. Johnson was in charge; thereafter, with Republican Richard M. Nixon in the White House, Democrats were more likely to view the war as a mistake.

Similarly, in 2004, 61 percent of Republicans said that Iraq was an immediate threat that required military action, whereas only 14 percent of Democrats and 26 percent of independents held that view. The partisan differences extend to many attitudes toward the Iraq war. Republicans, for example, are much more likely than Democrats or independents to believe that the war in Iraq was a major part of the war on terrorism. Republican support for the war in Iraq remained strong, whereas the support of Democrats and independents fell faster. In March 2003, 87 percent of Republicans said that attacking Iraq was the right thing to do. A year later, their support was still at 80 percent. Democratic support, meanwhile, dropped from 50 percent to 24 percent over the same time period, and independent support declined a comparable amount, from 70 percent to 45 percent.42 As can be seen in Table 6-9, what question was asked about Iraq did not matter much. Republicans, Democrats, and independents responded consistently. Republicans continued to support the Republican president, whereas Democrats did not. Independents were somewhere in between but became increasingly negative as time went on. Once Democrat Barack Obama took office, the percentage of Democrats who believed the war was going well jumped nearly 20 percentage points from 27 percent in September 2008 to 46 percent in June 2009, whereas Republican support fell from 88 percent to 76 percent during the same period.43

The public makes a distinction between the war in Iraq and the war in Afghanistan, although they share the characteristic of dragging on inconclusively. By the fall of 2008, 56 percent of the public thought sending military forces to Iraq had been a mistake, but only 28 percent thought it had been a mistake to send troops to Afghanistan.44 Similarly, a little over 60 percent of the public thought the United States should keep troops in Afghanistan, while only 45 percent thought troops should be kept in Iraq. Republicans overwhelmingly thought troops should be kept in both countries, but Democrats were less sure. Fifty-three percent of all Democrats thought troops should be kept in Afghanistan, but only 25 percent felt that way about Iraq.45 The fact that the war in Afghanistan has lasted longer shows that simple duration of a conflict does not determine the level of support. Most important, the connection between terrorism and the war in Afghanistan has not been questioned. Moreover, the cost in American lives and dollars has been less in Afghanistan than Iraq, and news coverage of successes and failures has been more extensive in Iraq. With the Obama administration’s emphasis on Afghanistan as the right place to fight terrorism, partisan-shaped opinions were once again evident as a 2011 survey showed that while 72 percent of the public supported Obama’s Afghanistan withdrawal plan, only 50 percent of Republicans liked the idea compared to 87 percent of Democrats and 74 percent of independents.46

Table 6-9

Sources: CBS News Poll, April 6–8, 2004, data provided by Inter-university Consortium for Political and Social Research,; 2008 American National Election Studies, available at

Up until Vietnam, bipartisanship was touted as the hallmark of American foreign policy—that “politics stops at the water’s edge.” While strictly speaking this was never true, the division between internationalist and isolationist views on the United States’ role in the world tended not to follow party lines. Today, the partisan polarization extends also to international issues, especially as they concern one or the other party’s handling of them. As Table 6-10 demonstrates, although there is widespread bipartisan agreement on general principles, such as preventing the spread of nuclear weapons, there is disagreement on the role of the United Nations, on combating global hunger, and on assessing presidential performance on foreign policy issues.

Table 6-10

Source: American National Election Studies, available at

As noted in previous sections, fairly strong and consistent relationships exist between partisanship and domestic issues and, increasingly, foreign policy issues. A leading assumption is that partisan identification provides guidance for the public on policy matters—that is, most Americans adopt opinions consistent with their partisanship. It also is likely that policy positions developed independently of one’s partisanship, but consistent with it, will reinforce feelings of party loyalty or that attitudes on issues will lead to a preference for the party most in agreement with them. Furthermore, issue preferences inconsistent with party loyalty can erode or change it. For any particular individual, it would be extremely difficult to untangle the effects of partisanship and policy preferences over a long period of time.

Even though on many issues most partisans of one party will hold a position different from that held by the majority of the other party, considerable numbers of people with issue positions “inconsistent” with their party identification remain loyal to that party. To account for this, it is variously suggested that (1) issues are unimportant to many voters; (2) only the issues most important to individuals need to be congruent with their partisanship; (3) individuals regularly misperceive the positions of the parties to remain comfortable with both their party loyalty and their policy preferences; or (4) the positions of each party are ambiguous or dissimilar enough in different areas of the country that no clear distinction exists between the parties. Undoubtedly, all these explanations have some degree of truth, and no one should expect to find extremely strong relationships between partisanship and positions on particular issues. Nevertheless, as an indication of the recent increased polarization of the parties, racial issues and cultural issues have joined traditional domestic economic issues to clearly differentiate Democrats from Republicans. While this has led those with consistently liberal or consistently conservative positions to be increasingly polarized from each other, it has also led to a widening “middle” of ideological libertarians, populists, and moderates—people with dissimilar views to each other and the parties in government in Washington, DC. These individuals, who have a set of views that match Republicans on one issue dimension and Democrats on another, are less likely to participate in political activities, are less partisan, and are more likely to split their tickets when voting. As noted earlier, there are times when these voters (like populists in 2016) do meaningfully change their typical electoral behavior in ways that can affect the outcome of an election.


A political ideology is a set of fundamental beliefs or principles about politics and government: what the scope of government should be, how decisions should be made, what values should be pursued, and so on. In the United States, the most prominent ideological patterns are those captured by the terms liberalism and conservatism. Although these words are used in a variety of ways, generally liberalism endorses the idea of social change and advocates the involvement of government in effecting such change, whereas conservatism seeks to defend the status quo and prescribes a more limited role for governmental activity. Another common conception of the terms portrays liberalism as advocating equality and individual freedom and conservatism as endorsing a more structured, ordered society. These two conceptions of liberalism and conservatism clash when considering differences in people’s preferences about economic and social issues. That is, modern conservatives generally prefer that the government stay out of the deep management of economic issues while preferring that the government be more activist in regulating social issues. For example, conservatives generally want the government to relax business regulations and to outlaw abortion and define marriage traditionally. Contemporary liberals generally prefer more aggressive business regulations but want a government that stays out of abortion and does not define marriage.

“Ideological” views on social and economic issues are not always joined in the political thinking of Americans. That is, some Americans have views on economic issues, relating to the government’s role in managing the economy, that are generally liberal or generally conservative, and some Americans have opinions on social issues, relating to the government’s weighing in on what is morally right or wrong, that are broadly liberal or broadly conservative. However, there is no law or general psychological principle that requires Americans to have views that are consistently liberal or conservative across both dimensions. What’s more, some people have attitudes that are moderate on these issues; still others just don’t care. Thus, the ideological landscape of American public opinion certainly contains those who are either liberal or conservative across both economic and social issues, but it also contains those who are conservative economically and liberal socially (libertarians), those who are liberal economically and conservative socially (populists), and those who fall somewhere in between (moderates).47 Usually, liberals and conservatives do the most participating in elections. In 2016, populists were uniquely important, both in the Republican primaries and in the general election itself.

While the experienced politicians running for the Republican presidential nomination in 2016 ran traditional campaigns that courted traditional primary voters, Donald Trump charted a different course. Aside from his behavior on Twitter and in the news media, Trump courted voters by espousing nationalist issue preferences that would “make America great again.” Eric Oliver and Wendy Rahn showed that Trump supporters were especially strong with respect to their national identity.48 Carmines and colleagues examined the support for nationalist policies among populists, libertarians, moderates, liberals, and conservatives, finding that populists held the strongest nationalist attitudes in 2008 (believing that immigrants take jobs, favoring torture, believing that Blacks have too much influence, and favoring less outsourcing).49 In 2016, conservatives and populists tied for holding the highest percentage of nationalist views. Interestingly, populists typically are more likely to vote for Democrats in presidential elections, as the ratio of Democratic populists to Republican populists is 55 to 27 percent. However, in the 2016 primaries, populists participating in the Republican primaries were the group that was most supportive of Trump’s candidacy for the Republican nomination.

Despite the ambiguities of ideological definitions, most commentators on the American political scene, as well as its active participants, describe much of what happens in terms of liberalism or conservatism—and with good reason. The news media routinely chronicle political debate as a contest between Republicans and Democrats in government. At the federal level, there is a clear “left–right” divide between Republicans, who are on the ideological right, and Democrats, who are on the ideological left. Indeed, Republicans and Democrats in the U.S. Congress are now more polarized in their roll call voting behavior than they were in the years immediately following the Civil War!50 Furthermore, most political history and commentary treats the Democratic Party as the liberal party and the Republican Party as the conservative one. Although considerable ideological variation remains in both parties, the trend in recent years is toward greater ideological distinctiveness between the two parties at both the elite and mass levels.51 Candidates of both parties attempt to pin ideological labels on opposing candidates (usually candidates of the other party, but sometimes within their own). In recent years, liberal has been portrayed more negatively than conservative, and some candidates for office portray themselves as progressive or use other such terms to avoid the liberal label.

Analysts of American political history pay special attention to those rare periods when a single-issue dimension dominates the public’s views of governmental policy. Periods such as the Civil War or the New Deal revealed deep divisions in the public, paralleled by a distinctiveness in the issue stands of the political parties. Electoral realignments of voters are forged by unusually strong issue alignments, and during such times a close correspondence can be expected between attitudes on the relevant issues and partisanship.

At other times, highly salient issues may capture the attention of the public, but they often cut across, rather than reinforce, other issue positions and party loyalties. If the parties do not take clearly differentiated stands on such issues and if party supporters are divided in their feelings toward the issues, party loyalty and the existing partisan alignment are undermined. On the other hand, if parties do take clear but competing positions on such issues, they run the risk of alienating some of their supporters. In a complex political system, such as that of the United States, new, dissimilar issue divisions accumulate until a crisis causes one dimension to dominate and obscure other issues. An example of these kinds of issues—and their electoral consequences—that we will explore in this chapter is issues related to nationalism.

The most consistent and the most distinctive ideological difference between the parties emerged during the New Deal realignment (see also Chapter 5). It focused on domestic economic issues, specifically on the question of what role the government should take in regulating the economy and providing social welfare benefits. These issues still underlie the division between the parties. Since the 1930s, the Democratic Party has advocated more governmental activity, and the Republican Party has preferred less. Historically, American political parties have not been viewed as particularly ideological, in part because other issues—such as racial or social issues—have cut across the economic dimension and blurred distinctions between the parties.

For example, in the 1940s and 1950s, the Republican Party was at least as liberal on race (i.e., supportive of civil rights legislation) as was the Democratic Party, with its strong Southern base. Similarly, in the 1970s, the two parties were both divided internally on the issue of abortion. Currently, immigration has the potential to be a crosscutting issue, though it seems to divide Republicans more than it does Democrats. What makes the two parties especially interesting today is that they are in the rare historical position of being quite consistent and ideologically distinctive. The two parties have become more ideologically polarized over a broader range of issues, and their supporters seem to have sorted themselves out, as increasing numbers of Democrats have liberal positions on social and economic issues, while more and more Republicans have conservative positions on them.52

So, a political ideology is a set of interrelated attitudes that fit together into some coherent and consistent view of or orientation toward the political world. Americans have opinions on a wide range of issues, and political analysts and commentators characterize these positions as “liberal” or “conservative.” Does this mean, then, that the typical American voter has an ideology that serves as a guide to political thought and action, much the same way partisanship does?

When Americans are asked to identify themselves as liberal or conservative, most are able to do so. The categories have some meaning for most Americans, although the identifications are not of overriding importance. Table 6-11 presents the ideological identification of Americans over the past five decades. A consistently larger proportion of respondents call themselves conservative as opposed to liberal. At the same time, about a quarter of the population regards itself as middle-of-the-road ideologically. The question wording provides respondents with the opportunity to say they “haven’t thought much about this,” and fully a quarter to a third typically respond this way, though that number dropped to 15 percent in 2020. The series of liberal and conservative responses has been remarkably stable over the years. Ideological identification, in the aggregate, is even more stable than party identification. The slight drop in the percentages saying they “haven’t thought about” themselves in these terms, and the corresponding increase in the proportion of those calling themselves conservatives in 1994 and beyond, may be a response to the heightened ideological rhetoric of the 1994 campaign and the increased polarization of the political parties. The overall stability of these numbers, however, should make one cautious of commentary that finds big shifts in liberalism or conservatism in the American electorate.

Table 6-11

Source: American National Election Studies, available at

Table 6-12 shows the relationship between ideological self-identification and party identification. Democrats are more liberal than conservative, Republicans are disproportionately conservative, and independents are fairly evenly balanced. But conservatives are three times as likely to be Republicans as Democrats, and liberals are far more likely to be Democrats. The relationship between ideology and partisanship is shown in the low coincidence of liberal Republicans and conservative Democrats. Also, the electorate tends to perceive the Democratic Party as liberal and the Republican Party as conservative. Those who see an ideological difference between the parties believe Republicans are more conservative than Democrats by a ratio of more than four to one.

There is also a relationship between social characteristics and ideological self-identification. Self-identified liberals are most frequent among better-educated whites who claim no religious affiliation or a non-Christian affiliation. Figure 6-5 also shows that a plurality of Hispanics (35 percent) self-identify as liberal, a significant change from 2016, while 29 percent identify as conservative. Self-identified conservatives are most common among white “other” Christians who went to college, but they are also found in substantial numbers among white “Other” Christians with no college and among Catholics. The major impact of education is to reduce the proportion of respondents who opt for the middle category. The better educated are more likely to call themselves either liberal or conservative than the high school educated. This tendency would be even greater if those who offered no self-identification were included.

Table 6-12

Source: 2020 American National Election Studies, available at

“No opinion” and “haven’t thought about it” responses omitted.

Three bar graphs of ideological identification by race, ethnicity, religion, and education in 2020.Description

Figure 6-5 Ideological Identification, by Race, Ethnicity, Religion, and Education, 2020

Source: 2020 American National Election Studies, available at

Approximately three-quarters of the public identifies itself as liberal or conservative. Do these individuals use ideological orientation to organize political information and attitudes? Does political ideology play a role for Americans similar to the role of partisanship as a basic determinant of specific political views? Analysis has usually centered on two kinds of evidence to assess the extent of ideological thinking in the American electorate—the use of ideological concepts in discussing politics and the consistency of attitudes on related issues—suggesting an underlying perspective in the individual’s approach to politics.

Data reported in The American Voter, by Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes, showed that few members of the electorate discussed their evaluations of the parties and the candidates in ideological language; only 12 percent did so in 1956.53 Other scholars have contributed similar analysis of subsequent years.54

It is possible that individuals may simply be unsophisticated in the verbal descriptions of their feelings about politics and political parties. Their ideology may guide their political decisions, but they may be unable to articulate it. In that case, the individuals’ attitudes toward public issues might be expected to show a degree of coherence and consistency; they would arrive at those positions by applying a common underlying set of political ideals. If individuals are liberal on one issue, one would expect them to be liberal on other related issues; if they are conservative on one issue, they likely are conservative on others. The most sophisticated analysis of ideological perspectives and consistency in issue positions, usually called issue constraint, was carried out by Philip E. Converse using data from 1956, 1958, and 1960.55 He found that the strength of the relationship among domestic issues and among foreign policy issues was about twice as strong as that between domestic and foreign issues. By normal standards, even the strongest relationship among domestic issues did not suggest particularly impressive issue consistency, but consistency does increase when survey questions reference which party holds a particular position on an issue. Additionally, the degree of issue constraint on various policy matters increased in 1964 and remains at this higher level to the present.

As might be expected, the degree of consistency among attitudes on different issues varies with the level of education of the individual; the more educated are substantially more consistent in their views than the less educated. However, increasing levels of education do not appear to account for the increase in issue constraint. The work of Norman Nie, Sidney Verba, and John Petrocik (see The Changing American Voter) shows convincingly that interest in politics is more critical. In other words, as members of the public become more concerned with issues and more attentive to political leaders, they perceive, and reflect, a higher degree of issue coherence. Almost certainly, the electorate has the capacity for greater issue constraint than it has shown. However, the exercise of the capacity depends much more on political leaders and events than on the characteristics of the electorate. When political leaders use ideological terms to describe themselves and the clusters of issues that they support, the electorate is capable of following suit.

The degree of issue constraint also depends on the range of issues considered. The American Voter documented a coherent set of attitudes on welfare policies and governmental activity, even in the 1950s. When the analysis moves to more disparate issues, such as support for welfare policies and civil liberties, the relationship weakens substantially. It can be argued that little relationship should be expected between positions in these different issue areas because they tap different ideological dimensions with no logical or necessary connections among them. For example, there is nothing logically inconsistent in a person’s opposing government regulation of business and believing in racial equality. In the first half of the twentieth century, internationalist views in foreign policy were considered the liberal position and isolationist attitudes conservative. However, the Cold War and Vietnam did much to rearrange these notions as liberals argued against American involvement and conservatives became more aggressive internationalists. In considering the question of issue constraint, two points should be kept in mind: (1) The meaning of the terms liberal and conservative changes with time, as do the connections between these ideologies and specific historical events, and (2) analysts, in studying issue constraint, invariably impose on the analysis their own version of ideological consistency, which, in light of the ambiguities surrounding the terms, is likely to be somewhat artificial.

A less strenuous criterion than issue constraint for assessing the impact of ideology on political attitudes is simply to look at the relationship between individuals’ ideological identification and their positions on various issues taken one at a time. Here, substantial relationships are found. The relationship between ideological identification and liberal views on various policy matters over the past thirty-six years is shown in Table 6-13. Between one-fourth and one-third of the people sampled did not have an ideological position or did not profess attitudes on these issues and are, therefore, missing in the analysis. Nevertheless, the data in Table 6-13 document strong, consistent relationships between ideological identification and many issue positions. The data do not, however, demonstrate that ideology determines issue positions.

If everyone had a strong ideology, attitudes would be determined by that ideology. To a considerable extent, this appears to happen to the most politically alert and concerned in the society, but this group is a small minority of the total adult population. To the extent that the major American political parties are ideologically oriented—by following the parties or political leaders in these parties—Americans have their opinions determined indirectly by ideology. American political parties are often characterized as nonideological, and the country has passed through substantial historical periods when the parties have seemed bent on obscuring the differences between them. At other times, such as 1964 and since the early 1990s, political leaders were more intent on drawing distinctions between themselves and the opposing party in ideological terms. At these times, the public responds by appearing more ideological as well.

Whether tightly constrained or seemingly built at random, ideology and issue preferences have to come from somewhere. For years, the dominant explanation was that preferences, ideology, and partisanship came from a process of socialization from one’s parents, schools, religion, and socioeconomic status. In recent years, scholars have started to demonstrate that while socialization certainly affects our opinions, ideology, and partisanship, biological factors do as well. In one study, from John Alford, Carolyn Funk, and John Hibbing, researchers compared issue preferences of identical twins to those of fraternal twins, finding that identical twins (ones from the same zygote, sharing 100 percent of their DNA) had views that were more alike than fraternal twins (from two zygotes, sharing roughly 50 percent of their DNA).56 Jeffrey J. Mondak has shown how people’s personality characteristics inform their ideology and preferences,57 and other scholars such as Peter Hatemi, Kevin Smith, James Fowler, Rose McDermott, and Hibbing have provided evidence on matters as diverse as the role that genes play in one’s likelihood of identifying as a partisan (but not with which party), how physiological responses to threatening and disgusting images correlate with liberalism and conservatism, and how genes are related to voter turnout.58 These studies have elicited plenty of controversy, with some scholars arguing that such studies are dangerous, whereas other scholars question the precision of measurements connecting deep-seated biological orientations with expressed political attitudes and behaviors.

Table 6-13

Source: American National Election Studies, available at

Note: The numbers in the table are the percentages (in each group) taking the liberal position on each issue.


The study of public opinion is of obvious relevance to public officials and political journalists who wish to assess the mood of the people on various topics, but the extent to which decision makers are influenced by public opinion on any particular policy is almost impossible to determine. Although policy makers must have some sense of the public mood, no one supposes that they measure precisely the attitudes of the public or are influenced by public opinion alone.

Political analysts and public officials both have difficulty assessing the likely impact of public opinion as measured by public opinion polls because the intensity of feelings will influence the willingness of the public to act on their views. Public officials who value their careers must be conscious of the issues that raise feelings strong enough to cause people to contribute money, to campaign, and to cast their ballots solely on the basis of that issue. As a result, public officials may be more responsive to the desires of small, intense groups than to larger, but basically indifferent, segments of the public.

In American society, public attitudes toward policies usually can be described in one of two ways: (1) as permissive opinion, whereby a wide range of possible governmental activities are acceptable to the public, or (2) as directive opinion, either supportive or negative, whereby specific alternatives are definitely demanded or opposed. Ordinarily, policy alternatives advocated by both political parties are within the range of permissive opinion, a situation that does not create highly salient issues or sharp cleavages in the public, even though political leaders may present their positions dramatically. Only when many people hold directive opinions will the level of issue salience rise or issue clashes appear among the public. For example, widespread directive support exists for public education in the United States. Most individuals demand a system of public education or would demand one were it threatened. At the same time, permissive support is evident for a wide range of policies and programs in public education. Governments at several levels may engage in a variety of programs without arousing the public to opposition or support. Within this permissive range the public is indifferent.

Occasionally, the public may out-and-out oppose some programs and form directive opinions that impose limits on how far government can go. For example, the widespread opposition to busing children out of their neighborhoods for purposes of integration has perhaps become a directive, negative opinion. Political analysts or politicians cannot easily discover the boundaries between permissive and directive opinions. Political leaders are likely to argue that there are supportive, directive opinions for their own positions and negative, directive opinions for their opponents’ views. One should be skeptical of these claims because it is much more likely that there are permissive opinions and casual indifference toward the alternative views. Indifference is widespread and, of course, does not create political pressure. It frees political leaders of restrictions on issue positions but, on balance, is probably more frustrating than welcome. When comparing the general ideological orientation of states to the general public policies produced by state legislatures, Robert Erikson, Gerald Wright, and John McIver have found that states with more liberal publics have state legislatures that produce more liberal public policies and vice versa.59 At the national level, James Stimson provides evidence of federal representation, showing that as the overall public mood becomes more liberal, liberals are more likely to get elected and produce left-leaning policies, and that as public mood swings to the conservative side, conservatives are more likely to enter office and produce more conservative policies.60

Study Questions

· What are some of the ways that people organize their attitudes about politics?

· What are some challenges involved with the measuring and interpretation of public opinion?

· How do Americans’ attitudes array across various economic, social, racial, and international issues, and how have these changed over time?

Suggested Readings

Althaus, Scott. Collective Preferences in Democratic Politics: Opinion Surveys and the Will of the People. New York: Cambridge University Press, 2003. A compelling demonstration that measures of public opinion do not represent the public equally.

Carmines, Edward G., and James A. Stimson. Issue Evolution: Race and the Transformation of American Politics. Princeton, NJ: Princeton University Press, 1989. A fascinating account of the role of racial issues and policies in American politics in recent decades.

Converse, Philip E. “The Nature of Belief Systems in Mass Publics.” In Ideology and Discontent, edited by David Apter, 206–261. New York: Free Press, 1964. A classic analysis of the levels of sophistication in the American public.

Erikson, Robert S., Michael B. MacKuen, and James A. Stimson. The Macro Polity. Cambridge, England: Cambridge University Press, 2002. A methodologically sophisticated study of representation in the United States.

Hibbing, John R., Kevin B. Smith, and John R. Alford. Predisposed: Liberals, Conservatives, and the Biology of Political Differences. New York: Routledge, 2013. An in-depth look at the evidence supporting the biological roots of the notion that political opponents experience and respond to the world differently.

Hochschild, Jennifer L. What’s Fair. Cambridge, MA: Harvard University Press, 1981. An intensive, in-depth study of the beliefs and attitudes of a few people that deals with traditional topics from a different perspective.

Kellstedt, Paul. The Mass Media and the Dynamics of Racial Attitudes. New York: Cambridge University Press, 2003.

Kinder, Donald R., and Nathan P. Kalmoe. Neither Liberal nor Conservative: Ideological Innocence in the American Public. Chicago: University of Chicago Press, 2017. A provocative analysis of public opinion arguing that Americans are not ideological but group-centric in their political decision making.

Lewis-Beck, Michael S., William G. Jacoby, Helmut Norpoth, and Herbert F. Weisberg. The American Voter Revisited. Ann Arbor: University of Michigan Press, 2008. A rich reanalysis of the themes from the classic work, using mainly 2000 and 2004 data.

Stimson, James A. The Tides of Consent: How Public Opinion Shapes American Politics. Cambridge, England: Cambridge University Press, 2004. A complex analysis of changing policy views and their impact.

White, Ismail K., and Chryl N. Laird. Steadfast Democrats: How Social Forces Shape Black Political Behavior. Princeton, NJ: Princeton University Press, 2020. Rich, innovative analyses of why Black Americans’ social expectations of other Black Americans encourage Democratic Party support.

Internet Resources

The website of the American National Election Studies,, has extensive data on the topics covered in this chapter. Go to the Resources menu and click on “Guide to Public Opinion and Electoral Behavior,” then scroll down to “Ideological Self-Identifications” for ideological items in a number of election years. Scroll to “Public Opinion on Public Policy Issues” for a wide range of attitudes from 1952 to the present. Each political item is broken down by an extensive set of social characteristics.

The Roper Center for Public Opinion Research,, at the University of Connecticut has an enormous collection of survey data on American public opinion from the 1930s to the present.

The General Social Survey,, has been collecting public opinion data since 1972 on a wide range of topics. Some of these data can be analyzed online at


1. Jianing Li, Michael W. Wagner, Lewis A. Friedland, and Dhavan V. Shah, “When Do Voters Support Black Lives Matter or the Green New Deal?,” Washington Post, December 8, 2020,

2. Lydia Saad, “U.S. Debt Ceiling Increase Remains Unpopular With Americans,” Gallup Politics, July 12, 2011,

3. Art Swift, “Americans Split on Whether NAFTA Is Good or Bad for U.S.,” Gallup Politics, February 24, 2017,

4. Philip E. Converse, “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. David Apter (New York: Free Press, 1964), 206–261.

5. John R. Zaller, The Nature and Origin of Mass Opinion (New York: Cambridge University Press, 1980).

6. Jianing Li and Michael W. Wagner, “The Value of Not Knowing: Partisan Cue-Taking and Belief Updating of the Informed, Uniformed and Ambiguous,” Journal of Communication 70, no. 5 (2020): 646–669.

7. Edward G. Carmines and James A. Stimson, “The Two Faces of Issue Voting,” American Political Science Review 74 (1980): 78–91.

8. Paul M. Sniderman and John G. Bullock, “A Consistency Theory of Public Opinion and Political Choice: The Hypothesis of Menu Dependence,” in Studies in Public Opinion: Gauging Attitudes, Nonattitudes, Measurement Error, and Change, ed. Willem E. Saris and Paul M. Sniderman (Princeton, NJ: Princeton University Press, 2004), 337–357.

9. Brett Creech, “Are Most Americans Cutting the Cord on Landlines?,” Beyond the Numbers 8, no. 7 (May 2019), U.S. Bureau of Labor Statistics, accessed July 6, 2021,

10. “In U.S., Decline of Christianity Continues at Rapid Pace,” Pew Research Center, October 17, 2019,

11. “2020 Time Series Study,” American National Election Studies, accessed July 6, 2021,

12. “2020 Election Surveys,” Elections Research Center at the University of Wisconsin–Madison, accessed July 6, 2021,

13. Drew DeSilver, “Q&A: After Misses in 2016 and 2020, Does Polling Need to Be Fixed Again? What Our Survey Experts Say,” Pew Research Center, April 8, 2021,

14. “How Does Gallup Polling Work?,” Gallup, accessed July 6, 2021,

15. In the 2006 General Social Survey, for example, 38 percent said “too much” was being spent on “welfare”; however, 70 percent said “too little” was being spent on “assistance to the poor.” See

16. Suzanne Mettler, The Submerged State: How Invisible Government Policies Undermine American Democracy (Chicago: University of Chicago Press, 2011).

17. “1992 Time Series Study,” American National Election Studies, accessed July 6, 2021,

18. “2008 Time Series Study,” American National Election Studies, accessed July 6, 2021,

19. “KFF Health Tracking Poll: The Public’s Views on the ACA,” Kaiser Family Foundation, June 3, 2021,

20. This view is most prominently associated with Paul M. Sniderman and Edward G. Carmines, Reaching Beyond Race (Cambridge, MA: Harvard University Press, 1997). See also Paul M. Sniderman, Philip E. Tetlock, and Edward G. Carmines, Prejudice, Politics, and the American Dilemma (Stanford, CA: Stanford University Press, 1993).

21. A strong statement of the symbolic racism position is in Donald R. Kinder and Lynn M. Sanders, Divided by Color (Chicago: University of Chicago Press, 1996).

22. This position is most commonly associated with Lawrence Bobo. See his “Race and Beliefs About Affirmative Action” in Racialized Politics, eds. David Sears, Jim Sidanius, and Lawrence Bobo (Chicago: University of Chicago Press, 2000), Chapter 5.

23. This position is represented by Jim Sidanius and Felicia Pratto in Social Dominance: An Intergroup Theory of Social Dominance and Oppression (Cambridge, England: Cambridge University Press, 1999).

24. Edward G. Carmines and James A. Stimson, Issue Evolution: Race and the Transformation of American Politics (Princeton, NJ: Princeton University Press, 1989).

25. Martin Gilens, Why Americans Hate Welfare (Chicago: University of Chicago Press, 1999).

26. Kinder and Sanders, Divided by Color.

27. Paul M. Kellstedt, The Mass Media and the Dynamics of American Racial Attitudes (New York: Cambridge University Press, 2003).

28. Ismail K. White and Chryl N. Laird, Steadfast Democrats: How Social Forces Shape Black Political Behavior (Princeton, NJ: Princeton University Press, 2020).

29. Brian F. Schaffner, Matthew MacWilliams, and Tatishe Nteta, “Explaining White Polarization in the 2016 Vote for President: The Sobering Role of Racism and Sexism,” paper prepared for presentation at the Conference on The U.S. Elections of 2016: Domestic and International Aspects, January 8–9, 2017, IDC Herzliya Campus, accessed July 6, 2021,

30. Doug McLeod, “Five Problems With How the Media Cover Protests,” Poynter, June 25, 2020,

31. Amber Boydstun, Making the News: Politics, the Media, and Agenda Setting (Chicago: University of Chicago Press, 2013).

32. Brian Montopoli, “Most Say U.S. Will Always Face Terrorism Threat,” CBS News, September 9, 2011,

33. “America Speaks Out About Homeland Security Survey,” conducted by Hart and Teeter Research Companies for the Council for Excellence in Government, February 5–8, 2004, Roper Center for Public Opinion Research,

34. Pew News Interest Index Poll, January 4–8, 2006, Roper Center for Public Opinion Research.

35. “America Speaks Out About Homeland Security Survey,” Roper Center for Public Opinion Research.

36. CBS News Poll, April 2005, Roper Center for Public Opinion Research.

37. “CBS News Monthly Poll #5, March 2003” (ICPSR 3787), Inter-university Consortium for Political and Social Research, April 29, 2009, version,

38. CBS News/New York Times Poll, April 23–27, 2004,

39. “2008 Time Series Study,” American National Election Studies.

40. John E. Mueller, War, Presidents, and Public Opinion (New York: Wiley, 1973), 208–213.

41. These data are available at See “President Bush: Job Ratings,”, accessed July 6, 2021,

42. “CBS News Monthly Poll #5”; and CBS News/New York Times Poll, April 23–27, 2004. Data provided by the Inter-university Consortium for Political and Social Research,

43. Jeffrey M. Jones, “Americans Divided on How Well Iraq War Is Going for U.S.,” Gallup Politics, August 5, 2010,

44. Gallup/USA Today Poll, July 2008, Roper Center for Public Opinion Research,

45. Pew Research Center for the People and the Press, September 2008,

46. Lydia Saad, “Americans Broadly Favor Obama’s Afghanistan Pullout Plan,” Gallup Politics, June 29, 2011,

47. Edward G. Carmines, Michael J. Ensley, and Michael W. Wagner, “Political Ideology in American Politics: One, Two, or None?” The Forum 10, no. 4 (2012): 1–18.

48. J. Eric Oliver and Wendy M. Rahn, “Rise of the Trumpenvolk: Populism in the 2016 Election,” The ANNALS 667, no. 1 (2016): 189–206.

49. Edward G. Carmines, Michael J. Ensley, and Michael W. Wagner, “Ideological Heterogeneity and the Rise of Donald Trump,” The Forum 14, no. 4 (2016): 385–397.

50. Nolan McCarty, Keith T. Poole, and Howard Rosenthal, Polarized America: The Dance of Ideology and Unequal Riches (Cambridge, MA: MIT Press, 2006). Updated data can be found at

51. Marc J. Hetherington, “Resurgent Mass Partisanship: The Role of Elite Polarization,” American Political Science Review 95, no. 3 (2001): 619–656.

52. Matthew Levendusky, The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans (Chicago: University of Chicago Press, 2009).

53. Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes, The American Voter (New York: Wiley, 1960), 249.

54. John C. Pierce, “Ideology, Attitudes, and Voting Behavior of the American Electorate: 1956, 1960, 1964” (PhD dissertation, University of Minnesota, 1969), 63, Table 3.1; Paul R. Hagner and John C. Pierce, “Conceptualization and Consistency in Political Beliefs: 1956–1976” (paper presented at the annual meeting of the Midwest Political Science Association, Chicago, 1981); and Michael Lewis-Beck, William G. Jacoby, Helmut Norpoth, and Herbert F. Weisberg, The American Voter Revisited (Ann Arbor: University of Michigan Press, 2008), Chapter 10. See also Norman H. Nie, Sidney Verba, and John R. Petrocik, The Changing American Voter (Cambridge, MA: Harvard University Press, 1976), Chapter 7.

55. Converse, “Nature of Belief Systems in Mass Publics.”

56. John R. Alford, Carolyn L. Funk, and John R. Hibbing, “Are Political Orientations Genetically Transmitted?,” American Political Science Review 99 (May 2005): 153–167.

57. Jeffery J. Mondak, Personality and the Foundations of Political Behavior (New York: Cambridge University Press, 2010).

58. P. K. Hatemi and R. McDermott, “The Genetics of Politics: Discovery, Challenges and Progress,” Trends in Genetics 28, no. 10 (2012): 525–533; John R. Hibbing, Kevin B. Smith, and John R. Alford, Predisposed: Liberals, Conservatives, and the Biology of Political Differences (New York: Routledge, 2013); James H. Fowler and Christopher T. Dawes, “In Defense of Genopolitics,” American Political Science Review 107, no. 2 (2013): 362–374.

59. Robert S. Erikson, Gerald C. Wright, and John P. McIver, Statehouse Democracy: Public Opinion, and Policy in the American States (Cambridge, England: Cambridge University Press, 1993).

60. James A. Stimson, Public Opinion in America: Moods, Cycles, and Swings, 2nd ed. (Boulder, CO: Westview Press, 1999).

Descriptions of Images and Figures

Back to Figure

In all eight graphs, the horizontal axis ranges from 0 percent to 100 percent in increments of 10, and the vertical axis shows different attitudes. All data are approximate.

The data from the first graph are tabulated below.

The data from the second graph are tabulated below.

The data from the third graph are tabulated below.

The data from the fourth graph are tabulated below.

The data from the fifth graph are tabulated below.

The data from the sixth graph are tabulated below.

The data from the seventh graph are tabulated below.

The data from the eighth graph are tabulated below.

Back to Figure

In both the graphs, the horizontal axis shows different social compositions, and the vertical axis shows percentage and ranges from 0 to 75 in increments of 25. The data from the graphs are tabulated below.

Back to Figure

In both the graphs, the horizontal axis shows frequency of church attendance, and the vertical axis shows percentage, ranging from 20 to 70 in increments of 10. The data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 0 to 45 in increments of 5. The vertical axis shows different levels of concerns. The approximate data from the graph are tabulated below.

Back to Figure

In all graphs, the horizontal axis shows different social compositions, and the vertical axis shows percentage, ranging from 0 to 80 in increments of 20. The data from the graph are tabulated below.

If you find an error or have any questions, please email us at Thank you!