Surveys of Public Attitudes toward the Arts

What Surveys Tell Us about the Arts' Political Trials

Paul DiMaggio and Becky Pettit

The NEA has been mired in controversy for most of the past decade, during which time it has lost much of its appropriation and even more of its autonomy, as Congress has directed larger chunks of its annual appropriation to the states, earmarked other moneys for special purposes, and effectively placed off limits NEA fellowships for most kinds of artists.

That many legislators — including House members who must face the voters every two years — have felt free to bash the Endowment with impunity suggests that the public does not support federal arts subsidies. Yet at regular intervals — most recently in 1996, with the release of the latest Americans and the Arts poll by Louis Harris — we are told that Americans love the arts and want the government to support them, even if it means paying a few dollars more in taxes for that purpose.

The Endowment has displayed some resilience and Congress has evinced no small measure of internal division. Last summer, the House of Representatives voted to eliminate the National Endowment for the Arts, whereas the Senate voted to increase its appropriation. This spring, the House's Republican majority underscored its commitment to abolishing the agency — the destruction of which had reached the top of the organized religious right's short list of legislative priorities; whereas, this summer, it became clear that the House would, for once, pass the Endowment's budget with minimal cliffhanging.

How can we understand this ambivalent performance, this oscillation between efforts to kill the NEA and renewed (if diminished) support? Political scientists tell us that legislators ordinarily follow the lead of public opinion, rarely defying the firm preferences of their constituents. How then can we square Congress's shoddy treatment of the NEA with arts advocates' assertions — backed by twenty years of polling data — that Americans love the arts?

• • •

We addressed this puzzle with the help of a new resource, an annotated directory of more than two dozen publicly available sets of data containing information on North Americans' participation in and attitudes towards the arts.1 One of us (Becky Pettit) had developed this resource in 1997 for Princeton's Center for Arts and Cultural Policy Studies, a research institute located in the University's Woodrow Wilson School. With support from the Pew Charitable Trusts, we conducted original analyses of thirteen of these data sets and reviewed the results of more than fifty polling questions from the Roper Center Archive to understand better how residents of the U.S. view the arts, what attitudes they hold towards public subsidy of artists and arts organizations, and how their views have changed over time.2

The results suggest that neither arts advocates nor the NEA's opponents can claim a clear decision before the court of public opinion. Americans think very highly of the arts in general: Majorities of close to ninety percent routinely agree that the arts are vital to the good life, that they are important to the development of children, and that they enhance the quality of communities.

But Americans' attitudes toward government arts programs are considerably more ambivalent. Support for some form of public spending on the arts is substantial — about two thirds of U.S. adults — and has been remarkably stable throughout sharp fluctuations in the NEA's political fortunes. But evidence also suggests that people are somewhat more supportive of local or state than of federal programs and of aid to museums and libraries than of assistance for artists. Moreover, whereas support for federal government aid to the arts is broad but shallow, about fifteen to twenty percent of the public have opposed federal arts programs with fierce conviction. Over the past decade, by attacking the NEA at its most vulnerable points, the right has created a powerful, if small, coalition of Christian conservatives and Republican partisans committed to ending thirty years of federal support for the arts. What the Endowment's enemies have been unable to win in the court of public opinion, they have won (in large measure) in the stadium of political strategy.

We think that arts advocates might have done better if they had used survey research more effectively. Opinion surveys have been less useful than they should have been due to the advocacy agendas of many study sponsors, an imbalance between resources devoted to data collection and those invested in data analysis, and a failure to monitor change by asking enough of the same questions on repeated surveys. After reviewing what we have learned from a quarter century of surveys, we make some suggestions about how funding for such research can be more effective.

A Reservoir of Good Will

Whenever Americans are asked about the value of the arts in a general fashion — i.e. without considering a price tag or comparing their value to something else — they express strong support. This is evident in surveys that ask people whether they want their children exposed to the arts (well over eighty percent do) or whether it is important to have arts education in the schools (almost everyone agrees). It is easy, of course, to concur with very broad statements about the arts' value to a complete education. But most survey respondents also support including specific courses in the curriculum and say that they would pay more taxes to keep the arts in the schools. Americans are equally united in believing that the arts are good for their communities, a position that about ninety percent of survey respondents endorse.

Government Aid to the Arts: Broad Soft Support vs. Narrow Intense Opposition

By contrast, public funding of the arts is more controversial and attitudes are more complex, as they require survey respondents to balance their largely positive views of the arts and their largely negative attitudes towards government. There is no single answer to the question “how much do Americans support public arts programs?” because polling results are influenced strongly by question wording and related factors. (For example, when asked if they approved of federal funding for the arts, two out of three respondents agree. But when asked in sequence if they approve of public arts funding at the city, county, state, and federal level [or at all levels], respectively, fewer than two in five say “yes” to federal support.

Opinion is most malleable when people lack firm convictions. (If surveys asked about attitudes towards cannibalism, question wording wouldn't matter much: almost everyone would oppose it.) Direct polling questions confirm that public policy towards the arts simply isn't very salient to many U.S. voters. Fewer than forty percent of U.S. adults in the mid-1990s described their opinions about government arts support as “strong;” thirty-seven percent told pollsters they knew too little to have an opinion. In 1995, as the House debated the NEA's demise and arts advocates and conservative direct mailers struggled to win the public to their side, two thirds of Americans reported that they “had not heard or read anything lately” about the agency. In other words, most people's attitudes towards arts policy are ill-formed, weakly held, and therefore easily swayed by vivid images or vigorous arguments. Given this, it is not surprising that the NEA has become a political football.

On the other hand — and this is the good news for arts advocates — most Americans have endorsed public funding of the arts over the past twenty-five years. Ongoing surveys have found that between two thirds and three quarters of survey respondents favor some form of government aid and between fifty and sixty percent favor a federal presence. And remarkably enough, these figures have remained stable even as the NEA has experienced more than a decade of controversy and cutbacks.

Why, then, isn't such support readily converted into legislative success? In large part because it is as weak as it is broad. When survey respondents are given the chance to choose between “strongly support” or “somewhat support” public funding, most choose the latter. If one compares only the proportion of respondents who describe themselves as “strongly” in one camp or another, the number of opponents is almost as high as the number of arts advocates. (This comparison is politically important, because only citizens with strong convictions are likely to take the trouble to sign a petition or contact an elected representative about an arts-related subject.)

Second, response patterns change when questions specify whether “public funding” occurs at the federal, state, or local level. People in the U.S. favor local government arts support more strongly than federal, making it easy for the Endowment's congressional opponents and advocates to compromise by earmarking more of the NEA budget for the states.

Third, the level of support expressed for arts funding is vulnerable to question formats that ask respondents to consider the cost of arts funding, compare the arts to other public purposes, or take into account broader efforts to cut the federal budget. If you ask people if they believe that government should support the arts, most will tell you that they do. If you ask them if they are willing to pay more taxes so that the federal government can spend more money on the arts (as opposed to health care or child welfare), they may respond very differently. In particular, when questions follow a lead-in that alludes to ongoing efforts to reduce the federal deficit (as they did at several points during the Reagan presidency and again in 1995), much larger percentages favor cuts than if such a deficit frame is absent.

Fourth, although majorities say that they favor federal aid to the arts, only small percentages of respondents advocate increasing the level of support — and this has been true since the early 1980s, even though the actual amount spent on the arts has declined substantially during this period. Significantly, among people with strong opinions the NEA's opponents have a decided edge. In 1990, for example, only one in forty people wanted “much more” spending on the arts, whereas almost one in five (eighteen percent) wanted “much less.” Such figures may be enough to embolden congressional conservatives, especially in the absence of an equally firm countervailing force of comparable size.

What Portions of the Public Does the Arts' Constituency Comprise?

We also explored differences between groups in their attitudes towards the arts. As with the views of the public as a whole, few changes were evident from the 1970s to the present. Differences between groups have persisted for the past twenty-five years.

Almost all the surveys revealed that women are significantly more supportive of government cultural subsidies and of the arts in education than men. Young people proved more sympathetic to government funding of the arts than older people: in one study, respondents younger than twenty-five favored federal support by a ratio of three to one, whereas opinion among people over fifty was evenly split. (If young voters retain their positive views as they age, the arts could face a kinder political environment in years to come.)

African Americans report significantly more positive attitudes towards the arts than their white counterparts. The difference is visible across many kinds of attitudes, but is especially strong in responses to questions about government support for the arts, which African American respondents are more likely to favor.

The number of years that a person has gone to school is the best predictor of his or her attendance at arts events, and education is strongly associated with many positive attitudes towards the arts as well. The strength of education's association with support for public funding is surprisingly modest, however, and appears to have declined in recent years.

The relationship between income and attitudes towards the arts is similarly complex, because economic success increases the value that people place on the arts, but militates against support for public expenditures. Respondents with higher incomes are much more likely than others to vouch for the importance of the arts for children and the community, but are no more likely than others to endorse federal spending that supports the arts.

Not surprisingly, people who attend arts events and visit museums hold more positive views towards the arts than people who do not. Indeed, the more kinds of arts events one participates in personally, the more likely one is to approve of government spending on the arts. But attitudes towards public funding reflect much more than self-interest, for the relationship between attendance and opinions, while statistically significant, is not enormous. Surveys also find a lot of support for the principle of public patronage among nonattenders, and identify a sizable minority of active attenders who do not favor a government role.

Cross-Pressured Supporters and Mobilized Opponents

These findings illuminate the problem that arts advocates have faced in mobilizing support for their cause, and the success that the NEA's opponents have experienced in gaining support for theirs. To favor federal aid to artists and arts organizations, one must believe that the arts are valuable and be disposed to endorse an activist federal government. The natural constituency for public arts programs should be the people who attend arts activities most frequently and value the arts most highly. But because such people are drawn disproportionately from among well-educated, high-income Americans who tend to be skeptical of federal initiatives, they are difficult to recruit to efforts to defend the NEA.

By contrast, opposition to federal funding tends to come from men and women who are fiscally conservative and who tend not to frequent arts events. The campaign against the NEA has worked (in so far as it has) by using institutions to which people are committed — the Republican party and predominantly white Evangelical Christian churches — to strengthen these tendencies among fiscally conservative Republicans and Evangelical Protestants who are suspicious of artists and the arts.

Why do we believe this? The General Social Survey (probably the highest quality ongoing sample survey of social attitudes in the U.S.)3 asked people whether they thought the federal government should spend “much more,” “more,” “less,” “much less,” or “about the same” amount of money on the arts — in 1985 (before the brouhaha over Endowment support for controversial photographers and performance artists), in 1990 (at the peak of the debate), and again in 1996. We explained people's support for increases and cuts in federal arts spending statistically with the help of logistic regression analysis, a method that lets one assess the impact of particular attributes or identities on which people differ while “holding constant” the influence of other characteristics.

The results are striking. In 1985, only two things predicted which respondents wanted “much less” arts funding: their education (people with B.A.s were less likely to advocate big cuts) and their attitudes towards the federal budget (people who wanted to cut other programs wanted to cut the arts as well). In 1990, these factors still mattered, but religion had come to the fore as an important predictor of support for large reductions to (or elimination of) federal arts programs, with regularly church-going Catholics and Evangelical Christians almost twice as likely to take this position as other Americans to whom they were similar with respect to age, race, education, income, fiscal philosophy, and political party affiliation.

By 1996, with the Serrano controversy in the past, Catholics no longer opposed the NEA significantly more than anyone else. Evangelical Christians remained the group most strongly favoring cuts — but now they were joined by self-styled “strong” Republicans, who had not been particularly antagonistic to the NEA (as a group) during the Reagan and Bush administrations. What is more, the commitment of college graduates was wavering by 1996. Consequently these two factors — Evangelical religion and Republican politics — were the most important (aside from general attitudes toward budget-cutting) in predicting opposition to the NEA. Moreover, each had an independent effect. Evangelicals were more likely to want big cuts than other people, regardless of their party affiliation; and Republicans were more likely to favor large reductions whether they were religious or secular. By fusing a coalition of fiscal conservatives, Republican partisans, and Evangelical Christians, the NEA's opponents constructed a potent pressure group whose relatively small size — fewer than twenty-five percent of the voting public — belied its visibility and clout.

The lesson we take from this account is not that federal arts support is doomed or its opponents invincible. Rather our conclusion is that public opinion is uncertain, protean, and up for grabs; that strategy matters; and that good information is necessary to develop good strategy.

Problems with Surveys — and How to Solve Them

We think that arts advocates can get better information than they have had available so far and that they can and should use it more wisely. Good information cannot in itself win political battles, but it is one ingredient in a successful strategy and one over which the arts' defenders wield substantial control.

Survey research is an indispensable tool for understanding public opinion, because it is the only way to generalize from the man or woman in the street to the U.S. population as a whole. Moreover, only sample surveys can provide reliable information about how the views of different groups within the population — Democrats and Republicans, men and women, arts attenders and stay-at-homes — differ.

Up to now, surveys of public sentiments towards the arts have been less helpful to those who wish to understand public opinion than they should be. Vague or leading questions, unnecessary change in question wording from one survey to another, failure to establish what terms mean to respondents or how much respondents know or care about the arts, all reduce the value of the research that has been done.

One problem is that most such surveys have been sponsored by arts advocates whose primary interest is in “demonstrating” that the public is behind their cause. Once a few sound-bites about the percentage of Americans who feel the arts are important or think that government should help the arts hit the evening news, the studies are usually abandoned.4 Few researchers have been motivated to concentrate on what one must do to improve public understanding by thoroughly probing survey results or by asking the same questions over time to monitor change.

Moreover, arts advocates have often sought to employ polls to preach to the converted rather than to change opinion. As advocacy purposes often have been understood in the arts, it is less important to understand the public's “real” opinions, than to produce usable opinions — opinions that can decorate newspaper editorials and public addresses. Arts studies often produce results pleasing to arts advocates by avoiding questions that would enable one to understand the ambivalence, multiple perspectives, and indifference that much of the public feels when (and if) they think about the arts.

Arts supporters have paid a high price for this naivete. If one wants to shape public opinion, one has to understand it first. People in the opinion-making business would never settle for a survey designed merely to make their candidate or product look good. Instead, they recognize that the purpose of research is to inform the design of strategies of persuasions by understanding the views of the people they want to reach. Given the threats that the arts now face, advocates require information that will enable them to develop persuasive messages that can reinforce positive elements and deaccentuate negative components in the public's view of the arts, and guide efforts to target those messages most efficiently.

How can research on public attitudes towards the arts contribute to this end? The problem isn't financial: The field needs wiser, not larger, investments in data gathering and analysis. Here are a few suggestions:5

Monitoring trends. The first priority is monitoring trends in a few key attitudes as reliably as possible. The most cost-efficient way to do this is to buy time on a reliable, professionally administered, national survey that is conducted at regular intervals with consistent methods, like the University of Chicago's General Social Survey or the University of Michigan's National Election Study. Because hundreds of social scientists and their students use the data from such studies in research and classes, analysis of the data at little additional cost is guaranteed. Repeated every three years or so, such baseline research would require minimal annual investment.

Ensuring utilization. The greatest waste in survey research is the frequent imbalance between investment in data collection and investment in data analysis. Arts advocates could stimulate high-quality research by following up data collection with a few well-written RFPs for sophisticated analytic reports on topics of substantial importance. (The NEA Research Division has used this approach successfully to ensure utilization of its Surveys of Public Participation in the Arts.) The size of each grant should be small (under $10,000) — sufficient incentive to attract talented researchers who are interested in the problem already, but too small to attract researchers who are only in it for the money. Such a program would ensure that investments in research yield the maximum impact.

Harnessing expertise. Although many problems raised in survey implementation can be resolved by working with a high-quality survey-research provider, an ongoing program of research into attitudes towards the arts would benefit greatly if it were advised by a survey workshop — a set of scholars including specialists on survey design, cognitive psychology, and attitudes towards the arts that would meet once or twice a year to advise study sponsors and generate ideas that would make the research effort more effective. Such a workshop could play several roles — counseling research sponsors on decisions about survey methods, reviewing proposed specifications for potential contractors, drafting and revising survey questions to meet the most rigorous standards of quality, and developing inexpensive, experimental studies to inform survey design or address questions that come out of the larger effort.

Probing Deeper: Small-Scale Experiments. It is often difficult to know what survey respondents have in mind when they choose an option on a questionnaire. This is particularly the case when the subject about which respondents are being asked is not very salient to them. In addition to big national surveys, then, people interested in public opinion require ways to get behind the brute data that surveys generate and explore the meanings and subtle understandings that lay behind people's responses.

One approach is in-depth interviewing or focussed interviewing — for example, to learn if people understand words like “arts” and “culture” in the same way as people who write survey questions. Small, inexpensive, experimental telephone surveys can also inform the production of an ongoing survey, by testing question design. (Ordinarily, they draw an inexpensive telephone sample, divide it into two portions, and ask a different version of a question to each.)

The very malleability of attitudes towards the arts suggests that many people may entertain several competing sets of opinions about the arts, or several partially contradictory images or understandings of what art is and does. When this is the case, people's ambivalence can be evoked by small-scale experimental telephone surveys that vary the lead-in or “framing” to an attitude question, in ways that help us understand how, for example, attitudes towards public funding are linked to concerns such as (respectively) arts education or erotic artistic content. A few small studies could reveal a lot about the cognitive structures that underlie and condition the sentiments that people espouse.

• • •

Research on attitudes towards the arts reveals much support alongside substantial ambivalence and indifference. Residents of the United States believe that the arts are important at the same time they are suspicious of government initiatives. Most of them haven't given public policy towards the arts a great deal of thought, making attitudes malleable. At the same time, aggregate opinion towards public support for the arts has been very stable, with most people endorsing federal, state, and local arts programs without wanting to spend any more money on them.

Understanding — and therefore acquiring the ability to influence — opinion requires a coherent program of research, informed by thoughtful deliberation within the field, but carried out by many investigators in many cities. Traditional surveys represent only one way (albeit probably still the most valuable way) to understand people's attitudes towards the arts. They can be complemented, informed, and rendered more efficient by the judicious use of other techniques ranging from interviews and focus groups to small-scale telephone-interviews.

Such survey techniques help us to see attitudes not simply as separate bits of data but as interlocking narratives — and thus enable us to understand the ambivalence, uncertainty, and complexity of the beliefs and opinions that drive people's political behavior. The better that arts advocates understand these attitudes and opinions, the more prepared they will be to convert the public's vast reservoir of good will towards the arts into political support capable of influencing the arts' political environment.

Paul DiMaggio is research coordinator, Princeton University Center for Arts and Cultural Policy Studies and chair, Princeton University Sociology Department. He has written widely on organizational studies and on arts and cultural policy. Becky Pettit is a doctoral candidate at Princeton in sociology and the Office of Population Research. In addition to her work in the arts, she is completing a dissertation on the effects of planned residential mobility on the achievement and welfare of children from low income families.


  1. Becky Pettit, Resources for Studying Public Participation in the Arts: An Inventory and Review of Available Survey Data on North Americans' Participation in and Attitudes toward the Arts, Working Paper #2 (April 1997) of the Princeton Center for Arts and Cultural Policy Studies. Available for $10 (plus $2 shipping and handling) from the Center (Woodrow Wilson School, Robertson Hall, Princeton University, Princeton, New Jersey 98544); also available online at the Center's Web site. Support for the inventory project came from the Andrew W. Mellon Foundation and the Nathan Cummings Foundation.
  2. The results of this work are available in two papers. The first, Becky Pettit and Paul DiMaggio, Public Sentiments towards the Arts: A Critical Reanalysis of 13 Opinion Surveys (Working Paper #5 of the Princeton Center, March 1998) is available through the Center (see note #1 for address) for $12 (plus $2 shipping and handling). The second, unchristened at this writing, should be available by or soon after the publication of this essay.)
  3. The General Social Survey is a full-probability personal interview survey designed to monitor change in social characteristics and attitudes. It has been conducted almost annually since 1972 by the University of Chicago's National Opinion Research Center. The survey regularly asks people how they feel about government spending for a range of specific purposes. In 1993 and again in 1998 it included special questions about people's attitudes towards the arts and culture.
  4. Louis Harris' Americans and the Arts series is an admirable exception in two respects: first, the surveys in this series have included similarly phrased questions about attitudes towards federal, state, and local arts spending and about willingness to pay more taxes for the arts since 1980; and, second (and most important), survey data (except for the 1996 survey) have been archived and made available to other researchers for secondary analyses.
  5. These recommendations are developed at greater length in the last section of “Public Sentiments toward the Arts” (see note 1).