October 14, 2009

But What Do the Polls Show?

How public opinion surveys came to play a major role in policymaking and politics

by Andrew Kohut, President, Pew Research Center

The following commentary is excerpted from The Politics of News: The News of Politics, 2nd Edition, with the permission of CQ Press.

In 1993 the Times Mirror Center, the forerunner of the Pew Research Center, began a series of comprehensive opinion surveys about foreign policy called “America’s Place in the World.” Over the years, the surveys have traced the currents in public opinion through the seemingly carefree days of the 1990s when “the United States had no enemies,” to the time when public concerns soared following the Sept. 11 attacks and Americans began a contentious debate about how best to deal with the terrorist threat.

The results of these surveys have been closely followed by the foreign policy community and well covered by the news media. Still, of all the briefings, press conferences and events associated with these surveys, one meeting in the mid-1990s stands out in my mind. I was introduced to my audience by Theodore Sorenson, once a principal adviser to and speech-writer for President John F. Kennedy. Sorenson remarked in his opening comments, “Now that we have to consider public opinion in the conduct of foreign policy, it’s worth listening to what Kohut has to say about what his polls show.”

Coming from a senior policymaker of another era, this introduction drove home to me quite clearly how much the role of public opinion had changed over the years. Polls now provide leaders with capital or impoverish them in their efforts to promote policies. Those who can back up their assertions by pointing to poll results find the going easier than leaders who cannot. In turn, news organizations cover policy initiatives differently when programs appear to have popular support compared with when they do not. As a result, the public has become a more important player in national affairs over the past three decades. It is not possible to find a major national policy initiative for which polling has not played a significant, even critical, role. In 1998, Kathleen Frankovich observed, “Polls have become even more important and necessary to news writing and presentation, to the point where their significance sometimes overwhelms the phenomena they are supposed to be measuring or supplementing.”1

The Emergence of Modern Polling

As is the case with so many big changes in modern society, many either credit or blame technology for the emergence of public opinion polling. The advent of inexpensive computing and low-cost communications was central to the rise in prominence of polls. As late as the 1960s most public opinion surveys were conducted by personal interviews. Telephone ownership did not become nearly universal until about the mid-1960s, and even at that point “long distance” telephone calls were expensive.

Personal interviewing required that polling organizations maintain national networks of interviewers across the country that carried out their surveys in randomly selected neighborhoods. Questionnaires were mailed to interviewers who would complete their assignments, and then mail them back. The entire procedure took about a month. It included printing the questionnaire, drawing maps for interviewers, mailing out, mailing back, and finally data processing on punch cards and on the slow computers of that era.

In the personal-interview era, just a few organizations, notably Gallup and Harris, had the facilities and national field staffs to conduct public opinion polls for news organizations. In “The Powers That Be,” David Halberstam reports a warning from outgoing president Lyndon B. Johnson to incoming vice president Spiro T. Agnew: “[W]e have in this country two television networks, NBC and CBS. We have two newsmagazines, Newsweek and Time. We have two wire services, AP and UPI. We have two pollsters, Gallup and Harris. We have two big newspapers: the Washington Post and the New York Times. They are all so damned big they think they own the country.”2

LBJ was correct for 1968, but things would soon change for the media and the pollsters. In the early 1970s AT&T began to offer discounted costs for nationwide telephoning on its WATS lines. At the same time, computing became less expensive, faster, and more efficient. These changes ushered in widespread use of nationwide telephone surveys, which were far less expensive than field interviewing and did not require an elaborate infrastructure. The ascendancy of the telephone surveys made it possible for the news media and others to conduct polls, but it is fair to say that the turbulence of the late 1960s and early 1970s made it imperative that news organizations better understand a nation that was experiencing extraordinary social and political change. The civil rights movement, race riots, the Vietnam War, the antiwar movement, the rise of the counterculture, and the women’s movement had changed the country and made its people far harder to understand than the American public of the 1950s.

The public was the story. No one understood that better than Phil Meyer. His pathbreaking book, “Precision Journalism: A Reporter’s Introduction to Social Science Methods,” spelled out this idea to news media willing to consider ways to improve their reporting on social changes they found difficult to understand. Meyer begins by detailing major stories that the media got wrong in coverage of the public’s reaction to the national turmoil of that era, underscoring the differences between what reporters were writing and what carefully conducted surveys were showing.3 Coverage of the Watts riots emphasized that the upheaval came at a time when relations between whites and blacks were worsening; polls showed just the opposite. Eugene McCarthy’s strong showing in the 1968 New Hampshire primary was interpreted as a manifestation of antiwar sentiment; University of Michigan surveys at the time found that hawks outnumbered doves among McCarthy’s supporters.4 And, although the headlines following the King assassination proclaimed the end of nonmilitancy among blacks, the polling showed more support for King’s philosophy, not less.

So conditions were right for the news media to embrace polling. And they did. The CBS/New York Times poll started regular news surveys in 1975. NBC’s first partner was the Associated Press, and it began polling in 1978. The ABC/Washington Post poll was launched in 1981. The impact of this on reporting of the findings of opinion polls is quite clear. Pioneering independent pollsters did ask a few questions of their national samples about Sen. Joe McCarthy in the 1950s and the Cuban Missile Crisis and civil rights movement in the 1960s. But they were, indeed, just a few questions. Gallup has been part of the public scene since the 1930s, and Roper and Harris followed in the 1940s and 1960s, respectively, but the intensive and routine coverage of public reactions to major national stories, political or otherwise, is a relatively new phenomenon.

President Jimmy Carter’s administration was the first to experience full bore the thorough polling scrutiny that has been the rule since. Polling covers not only the White House but also Congress and most other national institutions. Whether it is the economy, the budget deficit, health care, the environment or any number of domestic issues, polls explore and chart public reactions. National security and foreign policy, once the purview of the elites, are now subject to the scrutiny of the American public.

A few well-known cases illustrate the interplay among public opinion — as reported in national opinion surveys — public policy and politics.

The Public Restrains Reagan in Central America

In 1985 Ronald Reagan was riding high. He had been reelected in a landslide, the U.S. economy was on the mend. But the president faced a challenge in his own hemisphere from leftist insurgents in Central America. And with the debacle of Vietnam still fresh in their memory, Americans were wary of the president’s hard-line approach to the region. In summer 1986, Reagan’s overall approval rating was 63%; by comparison, just 34% approved of his handling of Nicaragua.5

The Reagan administration’s acknowledgment of public opposition to intervention in Nicaragua raised, at least for some, an enduring criticism of the role of public opinion polls: they subvert leadership. Mike Getler and David Ignatius, writing for the Washington Post, expressed the view that Reagan was pursuing “a statecraft driven by public-opinion polls as much as by a coherent strategy.” 6 They were not alone in charging that the Reagan administration was poll-driven but, writing at the time of revelations about the secret mining of Nicaraguan ports, they also pointed to an adverse consequence of the way the White House was coping with its lack of public support: “The administration, fearing a public backlash, has tended to plan its most important policies in secret, without adequate interagency discussion or expert advice.”7 This was, as we know, a prelude to the exposure of more serious Iran-contra dealings, which cast a shadow over the last three years of Reagan’s presidency.

The Public Is Persuaded to Go to War in the Gulf

When Saddam Hussein invaded Kuwait in August 1990, President George H.W. Bush and his administration had forgotten neither the lessons of Vietnam nor of the Iran-contra scandal. Polls showed the public ambivalent about the prospect of using military force to drive the Iraqis out of Kuwait. The news media polls found broad backing for sending troops to Saudi Arabia to protect the oil fields, but public reaction to a deeper involvement was decidedly mixed.In early August 1990 an ABC poll found the public opposed to bombing Iraqi military targets.8 During that same period, a Gallup poll showed the public divided about whether the Gulf was worth fighting for.9  Bush, however, masterfully took public opinion head on and built support for going to war.

The polls provided a track record of the impact of the two most significant steps the administration took to secure public approval. First, by seeking and obtaining a U.N. Security Council vote setting a deadline for Iraqi withdrawal from Kuwait, the administration transformed public opinion about the use of force, a transformation well tracked by Gallup’s CNN/ USA Today polls. Second, Bush’s desire to seek congressional approval bolstered the argument for going to war rather than waiting for economic sanctions to discourage Saddam. Over the course of the congressional debate, the ABC/Washington Post daily tracking poll found the percentage of respondents favoring the use of force immediately or within a month rose from 48% during the Jan. 2-6, 1991, period to 58% by Jan. 13.10

Ultimately the Gulf War enjoyed public support because it was brief and ended well. But it also illustrated the extent to which leadership could address and educate a public concerned about the use of force in an era in which the legacy of Vietnam was still much in evidence. And it showed the extent to which the media’s reporting of public opinion served as the backdrop for coverage of the debate about when and whether to go war.

The Public Saves President Clinton’s Job

Of all the opinions that polls have tracked in the modern era, none has been more remarkable than President Bill Clinton’s approval ratings rising on the news of allegations that he had carried on an affair with a White House intern, Monica Lewinsky. A Pew Research Center poll in mid-January 1998 found that 61% of its respondents approved of the way the president was handling his job.11 Two weeks later, Clinton’s ratings spiked to 71%, reflecting public outrage over the way the media had prejudged Clinton’s guilt.12 The same trend was recorded in Gallup and other national surveys. The Pew Research poll analysis found the public more discontented with the president’s accusers in the news media than upset by Clinton’s alleged misbehavior.13

The public’s unexpected rallying to Clinton’s side led to a transformation of the Washington establishment’s judgment of his political viability. Before news of Clinton’s polling boost, political insiders had all but written him off. Public support for the president allowed, if not encouraged, congressional Democrats to rally to his side.

The impact of Clinton’s standing in the polls along with growing antipathy toward the president’s accusers were also potent factors in the impeachment debate and the broader politics of that contentious midterm year. The public stood by Clinton through each chapter of the saga: his grand jury testimony, his admission of lying, the revelations of the Starr report, and ultimately the Republican vote to impeach him. He ended the year with a 71% approval rating. His party actually picked up eight seats in the House of Representatives — an unusual occurrence for a second-term president, let alone one about to be impeached. It is inconceivable to think that public opinion could have had such an impact in an era prior to the emergence of the media polls.

The Public Changes Its Mind About Privatizing Social Security

In December 2004 President George W. Bush proclaimed that he was “armed with political capital” that he had earned in his reelection victory, and he planned to spend some of it to reform the Social Security system. The main idea was to give younger workers the option of depositing part of their Social Security contributions into private accounts. The polls at the time indicated that the president might well be successful. Americans for years had given high priority to steps to secure the Social Security system, and polls found broad conceptual support for the idea of allowing younger workers to have a private accounts option.

Anticipating opposition from many quarters, including the powerful AARP, the president announced he would personally lead a campaign to assure public backing. As the White House revved up on the issue, however, the public’s response grew more negative. A March 2005 Pew Research survey noted that “Despite Bush’s intensive campaign to promote the idea, the percentage of Americans who say they favor private accounts has tumbled to 46% in Pew’s latest nationwide survey, down from 54% in December and 58% in September. Support has declined as the public has become increasingly aware of the president’s plan. More than four-in-ten (43%) say they have heard a lot about the proposal, nearly double the number who said that in December (23%).”14

The reaction of the Bush administration to the feedback from the polls was a lack of reaction. The president continued to promote the plan in meetings across the country. As a result, growing numbers of Americans became aware of the idea, fewer supported it, and more expressed overall disapproval for Bush. He was spending his political capital, but not getting much for it.

The administration’s continued push on this issue began a process in which the public reconsidered Bush more generally. His personal favorable ratings tumbled, and the percentage of people regarding Bush as a strong, trustworthy leader who could get things done declined dramatically. A September 2005 Washington Post article quoted conservative commentator Bill Kristol, editor of the Weekly Standard, as saying, “The negative effect of the Social Security [campaign] is underestimated. Once you make that kind of mistake, people tend to be less deferential to your decisions.” 15

This effect came home most dramatically in a Pew Research survey in spring 2005 that showed public opposition to proposals with Bush’s name specifically attached to them, even when the public favored their thrust. For example, the poll found broad support for Bush’s suggestion of limiting the growth of Social Security benefits for wealthy and middle-income retirees, while keeping the current system intact for lower-income people. By a margin of 53% to 36% the public liked the idea — as long as the plan did not have the president’s name on it. When Pew Research tested it as a “Bush proposal,” public reaction to the same idea was quite different: 45% in favor, 43% opposed.16

Lessons and Limits

These cases should not be viewed as a celebration of the power of public opinion or the importance of polls. Rather, they illustrate the extent to which public views have played a central role in the course of national affairs since the 1980s. They also provide an opportunity to consider how the emergence of an empowered public has altered the relationship both between the people and the press and the people and its leaders.

As to the people and the press — or, more broadly, the media — the polls have made clear in reactions to the Clinton case, as in many other instances, how great a capacity the public has for ignoring the media. The Clinton episode is certainly not an isolated incident of public opinion polls coming in with a different verdict from the one proclaimed by the media. One stunning turnabout occurred in the 1988 presidential campaign, when candidate Bush picked Dan Quayle as his running mate. The press let out a hoop and a holler from the convention in New Orleans, predicting that the junior senator from Indiana’s reputation as a “lightweight” and his National Guard service during the Vietnam era would doom Bush’s chances of winning the fall election. The polls came in quickly and in unison — they said yes, the public did not have a high regard for Quayle, but his presence on the ticket made not a whit of difference to potential support for Bush’s candidacy.

Another high-stakes example of the public ignoring press exhortations occurred in 1995 following the Republican takeover of Congress. While the media, and the political community more generally, were extolling the new House Speaker Newt Gingrich’s political success and hyping how conservative the country had turned, the polls came in very quickly to say wait — this is not the case. We voted against the Democrats. We did not vote for undermining the school lunch program, shutting down the Department of Education, or weakening the Environmental Protection Agency, and the like.

So, although polling has provided the press with a more comprehensive and accurate portrayal of public opinion, at critical junctures it also serves as reality check for the media when they prematurely and inaccurately pronounce what conventional wisdom is emanating from the American public.

Time and again, the polls have illustrated that at no point does the public suspend judgment of its leaders. In the Clinton years, the public gave its assent to many major policies, including quiet approval of welfare reform and grudging approval to NAFTA. Yet the polls also illustrated the public’s capacity to ultimately reject a major reform proposal even though it addressed an area of great concern. In September 1993, when the Clinton health care plan was first announced, the polls initially found at least tepid support for it. But public reactions changed from cautious enthusiasm to strong rejection over a six-month period as the public heard more about the details.

The case of G. H. W. Bush’s marshalling of public opinion during the buildup to the Gulf War reminds us that the public responds to leadership — give people a rationale for national sacrifice by seeking international and congressional endorsements, and they will follow. But it also reminds us that leaders cannot own public support; they can only rent it. Bush had the highest approval ratings of any president in polling history in March 1991, yet he lost a bid for reelection 18 months later.

What can be concluded about public opinion itself in the modern era? Yes, through media polls, public opinion has become an omnipresent factor in national affairs. As we have shown, it has a direct bearing on relations between the people and the media and between the people and their leaders. But, what inferences can be drawn about the nature of the role of public opinion in national affairs, as a consequence?

First, the public plays a passive, not active role in shaping public policy. The polls show assent or opposition to policies that the media know or suspect are on the agenda of national leaders. For any number of reasons, the news media are unlikely to conduct polls on policy options that are not under current consideration. One reason is they do not want to be accused of manufacturing news through polling. Even so, this is one of the frequent complaints about civic journalism, which is primarily concerned with local affairs. But polling that asks ordinary citizens about policy options they have not heard about often produces dubious results. Polls provide a good sounding board for public reactions, but they are not a fount of specific suggestions for public policy.

Second, polling has influenced the techniques and strategies of leadership, rather than hamstrung leaders. Coping with what the polls show about public opinion is one of the things that leaders now have to do. Yes, previous generations of leaders could not ignore public opinion altogether; certainly not on the biggest questions of the day. FDR knew he had to bring the public around to enter World War II. But such instances were the exception, not the rule. Harry Truman did not have to worry about constant reminders that he lacked public support for the Marshall Plan or the Berlin airlift or other early Cold War policies. In four years (1950-1953), the Gallup poll asked only 135 questions about Korea. In contrast, Gallup alone asked 1,021 questions about Iraq just in the three-and-a-half years following the 2003 launch of that war.17

Modern leaders cannot avoid public opinion as measured in the polls. They need to know how to use polls or they end up being used by them, to quote historian Garry Wills.18 In addition, the registration of public support or lack of it can become an issue in itself. Backers of President Clinton in 1998 could point to public opinion polls to show how wrongheaded Republicans in Congress were when it came to attempting to remove the president from office because of his affair with Monica Lewinsky. At almost the same time, the Clinton administration had a public relations problem with the findings of polls showing lack of support for the interventions it would undertake in the Balkans and Haiti.

Third, although the prominence of polls gives greater voice to the people, other competing voices can still trump public opinion or, at a minimum, tone down the impact of what the polls are showing. A notable example is found in the general public’s consistent calls for greater gun control. After the Columbine school shooting, support for greater restrictions reached a crescendo.19 Yet, broad public opinion was no match for the power of the National Rifle Association. The NRA can marshal voting support for candidates of its choice far more effectively than can supporters of gun control, despite their far larger potential constituency. Similarly, public opposition to free trade is often blunted by effective lobbying by business interests, which in the modern era are often more effective than labor unions and other like-minded groups.

Fourth, the great empowerment of public opinion raises questions about the potential for manipulation. The American public is noted for the limited attention it pays to public affairs.20 Critics of the public’s judgments charge that an ill-informed public is easily pushed and pulled by advocates.

It is certainly not unreasonable to think that the public is susceptible to undue persuasion on occasion, but there is a long history of failed attempts to manipulate public opinion. Perhaps the best way to think about public opinion and its relationship to politics and policymaking is that the American public is typically short on facts, but often long on judgment.


1. Kathleen Frankovic, “Public Opinion and Polling,” in The Politics of News: The News of Politics, ed. Doris Graber, Denis McQuail, and Pippa Norris (Washington, D.C.: CQ Press, 1998).
2. David Halberstam, The Powers That Be (New York: Knopf, 1979), 596.
3. Philip Meyer, Precision Journalism: A Reporter’s Introduction to Social Science Methods (Bloomington: Indiana University Press, 1979), 1–3.
4. Ibid,
5. Gallup Organization survey, July 11–14, 1986, based on in-person interviews with a national sample of 1,539 adults.
6. David Ignatius and Michael Getler, “Reagan’s Foreign Policy: Where’s the Rest of It?” Washington Post, November 16, 1986.
7. Ibid.
8. ABC News/Washington Post survey, August 8, 1990, based on telephone interviews with a national sample of 769 adults.
9. Gallup Organization survey, August 23–26, 1990, based on telephone interviews with a national sample of 1,010 adults.
10. Andrew Kohut and Robert C Toth, “The People, the Press, and the Use of Force,” The Aspen Strategy Group, August 14-19, 1994, Aspen, Colo. (Washington, D.C.: The Aspen Institute, 1994).
11. Pew Research Center for the People and the Press survey, “Spending Favored over Tax Cuts or Debt Reduction,” January 23, 1998.
12. Pew Research Center for the People and the Press survey, “Popular Policies and Unpopular Press Lift Clinton Ratings,” February 6, 1998.
13. Ibid.
14. Pew Research Center for the People and the Press survey, “Bush Failing in Social Security Push,” March 2, 2005.
15. Peter Baker, “Bush Continues Social Security Campaign; Polls Show President’s Roadshow Failing to Drive Up Support for his Plan,” Washington Post, May 20, 2005.
16. Pew Research Center for the People and the Press survey, “Economy, Iraq Weighing Down Bush Popularity,” May 19, 2005.
17. Gallup question numbers come from a Roper Center for Public Opinion Research Ipoll search of Gallup surveys, between the identified dates. Terms searched were Korea and Iraq.
18. Garry Wills, “Read Polls, Heed America,” New York Times magazine, November 6, 1994, 49.
19.Andrew Kohut, “Gore, Bush and Guns,” New York Times, May 12, 2000.
20. Pew Research Center for the People and the Press survey, “The Age of Indifference,” June 28, 1990.