A History of Presidential Pollsters

The Beginning: Roosevelt

Polling as a fledgling industry had its first major public relations coup in the form of the Gallup, Roper and Crossley successful prediction of FDR’s 1936 win in contrast to the failure of the Literary Digest’s readership survey’s prediction. It is perhaps unsurprising under the circumstances that Roosevelt would look kindly on the new methodology. His interest in using polling in elections, however, predated his second campaign, and his broad interest in the potential of polling for both campaigning and governing was extensive. His administration furthered the science of survey research and laid the foundation for the development of presidential polling. 

Emil Hurja, Roosevelt’s chief political advisor over his first two terms, was a proto-pollster who used a mix of approaches, some of which have been were utilized by other later researchers. He combined polling in key districts with other data to model the 1932 election. In an attempt to understand where the Literary Digest had gone wrong, he backweighted respondents in the data file to their reported vote in the past election and found that the non-random sample would still have produced accurate results under a proper weighting scheme, an approach now adopted by many online pollsters. 

As the threat of war grew, Roosevelt turned to Handley Cantril of the Office of Public Opinion Research in Princeton to provide him insights into American public opinion for effective governing. Cantril gathered and interpreted public polling for the president, often before release. But FDR’s team wanted to craft questions, not depend on the surveys others designed. Distrusting Gallup, a Republican, FDR turned to the loyal Democrat Cantril to serve as an intermediary, suggesting questions to Gallup and utilizing Gallup’s field organization for surveys. This awkward arrangement was discarded once a source of funding was found via Gerard Lambert, most known for the marketing acumen that had made a household name of his family business’s primary product, Listerine. Together Cantril and Lambert offered Roosevelt key information about American attitudes about the war.

FDR’s administration also instigated or encouraged the development of survey projects in the Office of War Information which hired major figures in polling including Elmo Roper and Samuel Stouffer and at the State Department, which assisted in the foundation of a new polling center, which would become NORC. Most of the polling professionals who built the field in the forties and fifties developed their skills and techniques during their time working for the government during the war. 

 

Dormancy and Revival: Truman and Eisenhower

Truman’s second term began with what might be polling’s greatest public failure: the projection of Thomas Dewey as the winner of the 1948 election. Perhaps unsurprisingly, Truman distrusted polls. He frequently spoke against them, and conventional wisdom among those who study presidential polling long held that polls were not used in his administration. However,  Brandon Rottinghaus’s review of the Truman archives found that some polling was consulted and even likely commissioned by Truman through the DNC, primarily in relation to campaign strategy. Nonetheless, Truman’s use of polling was extremely limited compared to Roosevelt’s, and no figure like Cantril or Hurja emerged to earn the title of “presidential pollster” to Truman.

Polls returned to the White House with Eisenhower, who used the media polls and his own internal polling for taking the pulse of the public. As with Roosevelt, major pollsters sent data directly to the White House. The Republican National Committee employed Batten, Barton, Durstein and Osborn for first the campaign, then a tracking poll during Eisenhower’s administration. Funding for polling came through the RNC. Party infrastructure became the primary way presidents paid for polls.

During Eisenhower’s administration, polling in the federal branch led to a crisis with Congress. The precipitating project was not presidential polling, but a 1957 State Department poll that asked about Americans’ attitudes toward US military involvement in Indo-China. The State Department had been conducting polls through NORC since 1943, but the operation had flown under Congress’s radar until a short article in the Washington Star referenced a polling result that set off the suspicions of Congressman Glenard P. Lipscomb, (R-CA). Upon discovery of State’s polling efforts, Congress bristled at the idea that unauthorized polling had been paid for through Congressionally-appropriated monies for what they saw as an illegal use of funds for propaganda.

The usurpation of the traditional Congressional role as the conduit of public opinion to the Executive branch seemed to be more relevant than concerns about propaganda. The House Government Operations Subcommittee on International Operations investigation made these conflicts overt, as quoted in Richard Eisinger in his 2003 book The Evolution of Presidential Polling:

Congressman Victor Know (R-VA): [T]he Secretary [of State] is limited by making use of the agency of the government to which public opinion is intended to be formally expressed, namely, the Congress.

H. Schuler Foster, chief of State Department’s Division of Public Studies: But, Mr. Knox, the primary purpose of Congress, of course, as you and I realize, is to enact laws , and not to convey from week to week, and day to day, the opinions of the American people on all the topics, foreign and domestic.

Knox: Well, now, Mr Foster, you are aware that the Secretary of State often appears before committees of Congress and testifies, which from the questions that are propounded to the Secretary and his answer to them, certainly gives him some indication as to what public opinion is, as far as Congress is concerned, does it not?

Foster: Yes, I would think it would be more accurate, Mr. Knox, to call that congressional opinion, or perhaps, committee opinion.

The State Department may have been duly chastised, but polling in the executive branch did not cease. The genie could not be put back in the bottle, and soon Congresspeople began to adopt polling for their own purposes.

Increasing Institutionalization: Kennedy, Johnson and Nixon

Louis Harris is often credited as the first true presidential pollster – advisor and pollster, deeply trusted by Kennedy. Harris’ role in Kennedy’s election campaign was the first of its kind, though not the last. Historian journalist David Halberstam wrote: "Never before in American history had a major presidential candidate so depended upon the advice and skills of a pollster, using polling much as an airline pilot uses radar to chart and comprehend what he can no longer see for himself." (2000) With Harris, polling became central to campaigning, and there it has remained.

Harris decided his public polling and his private clients created a conflict of interest he resolved by swearing off private political polling in 1963. One of Harris’s former assistants, Oliver Quayle, left to start his own firm as a result, and it was to him that Johnson turned to conduct polling when he became president. After a time, Quayle fell out of favor, perhaps because he was publicly vocal about his work for Johnson. Joseph Napolitan, a former political reporter turned public relations professional and coiner of the term “political consultant”, took over the role of pollster to the Johnson administration. Johnson utilized Napolitan’s polling, public polling reports, and any polling material he could receive from other Democrats. 

Harris decided his public polling and his private clients created a conflict of interest, so he resolved it by swearing off private political polling in 1963. One of Harris’s former assistants, Oliver Quayle, left to start his own firm as a result, and it was to him that Johnson turned to conduct polling both during the 1964 campaign and when he became president. After a time, Quayle fell out of favor, perhaps because he was publicly vocal about his work for Johnson. Joseph Napolitan, a former political reporter turned public relations professional and the man who coined the term “political consultant”, took over the role of pollster to the Johnson administration. Johnson utilized Napolitan’s polling, public polling reports, and any polling material he could receive from other Democrats. Johnson consulted polling to give a measure of public opinion on major issues that affected his campaigns and his presidency, including civil rights, particularly voting rights legislation; Medicare; the War on Poverty; federal funding for education; and the increasingly unpopular war in Vietnam.

The use of public opinion polling by presidents was accelerating quickly. Scholars Lawrence Jacobs and Robert Shapiro identified three areas of growth in presidential polling in the period from Kennedy through Nixon: “the enhancement of the White House's organizational capacity for collecting and analyzing public opinion information, the increased centralization of White House control over public opinion information, and the intensified political use of the apparatus.” (1995)

Certainly the sheer number of polls conducted for presidential candidates and presidents was skyrocketing. From examination of presidential archives, Jacobs and Shapiro identified 93 private polls conducted for Kennedy during his campaign and administration, 130 for Johnson and 233 for Nixon.

Nixon relied on multiple polling organizations. The majority of his polls were conducted by Opinion Research Corporation, under Harry O’Neill, and Chilton Research Services. He also utilized Richard Wirthlin’s Decision Making Information (DMI, later Wirthlin Worldwide) and Robert Teeter’s Market Opinion Research (MOR), among others. Wirthlin had started his firm in Los Angeles after getting his PhD from UCLA.  Teeter was in the process of turning a Detroit market research company into a Republican polling powerhouse. The different organizations were involved in different series and through different funding mechanisms, not all utilizing the RNC, allowing Nixon to maintain tight control over access.

Here is a link to a 1998 C-CPAN broadcast of a panel with Harry O'Neill, political scientist Cliff Zukin, and Gallup Organization Managing Editor David Moore, discussing comparisons of polling between President Nixon's and President Clinton's scandals.

Old and New Blood: Ford and Carter

After Nixon stepped down, Ford depended on polling from Robert Teeter, essentially extending the relationship that had existed with the Nixon White House. He also utilized Fred Steeper, who was a graduate  of University of Ann Arbor Michigan, where he had studied at the illustrious Institute for Social Research, before joining MOR under Teeter and then starting his own company, Market Decisions, Inc. 

The 1976 election brought a new name to the field. Carter had an existing relationship with D.C. Democratic pollster William R. Hamilton, but he also engaged the services of a newcomer, Patrick Caddell. Caddell had started his firm,  Cambridge Survey Research, as a senior at Harvard University and became celebrated as a wunderkind when his strategic messaging that painted Carter as a political outsider rocketed the candidate to the White House. Caddell was 26 years old at the time.

Caddell was at the center of Carter’s inner circle in the election and into his administration. Some researchers have pointed to Caddell as the first pollster to play a very public role as presidential advisor. Although Carter continued to consult polling, there are indications in the Carter Library archive that Caddell was unhappy with the frequency, which was, according to researchers Shoon Kathleen Murray and Peter Howard, on average every few months. In a November 6, 1979 memo to the president, Caddell wrote:

You may have been baffled by the disparity of recent polls, particularly the Yankelovich Time Magazine poll showing a distance of 10 points and the NY Times/CBS poll showing you trailing 54%-20%. Well, frankly, so am I. I have some tentative ideas – but since we have no data of our own (and have had none since early this year when we surveyed for the first time since early 1978 – Hamilton and other decided long ago that there was no need for polling, a decision that I would selfishly argue has been costly beyond measure) and since the news organizations are parsimonious in their willingness to release information, these conclusions are tentative at best.
 

Caddell made up for lost polling at the end of the election period. According to Mark Levy, in 1980 “From late August until election eve, Caddell conducted 133 separate surveys in 39 states, fielded 14 national studies, carried out 3 waves of open-ended qualitative surveys, and 2 before-and-after media market tests of campaign commercials.”

It wasn’t enough – perhaps election results showed Caddell was right to bemoan the failure to poll in 1979. But the next president would not make the same mistake. Reagan would never lack for polling.

Established Polling Operations: Reagan, Bush Sr and GW Bush

Reagan turned to two of Republican’s favorite pollsters to implement his massive polling program: Richard Wirthlin and, to a lesser degree, Lance Tarrance, who had been a Nixon political appointee to the Census and worked for DMI before starting The Tarrance Group. The Reagan polling operation was unprecedented, and included state and district-level polling, “brushfire” polls in response to emerging events, and the Eagle national tracking. 

George Bush’s polling operation was nowhere near as extensive as Reagan’s, but nonetheless, polling continued through the usual team of GOP pollsters.  Robert Teeter, who had polled for Nixon and Ford, played the primary role, but Richard Wirthlin also played a role. Fred Steeper of Market Decisions, Inc., a former employee of Teeter’s at MOR, also worked for Bush.

George W. Bush was vocal about his distrust of polling, but, as reported by Brookings, he used his father’s pollster Fred Steeper, as well as Jan vanLohuizen of Voter/Consumer Research. His expenditures for polling were substantial, but he maintained a certain distance from his polling operation, preferring to receive recommendations pulled from the data from his trusted advisor Karl Rove.

 

A Constant Polling Operation: President Bill Clinton

Bill Clinton did not come to his campaign or his presidency with the sort of legacy polling establishment that the Republican presidents of the previous two decades had had. He turned to Stan Greenberg of Greenberg Quinlan Rosner Research, a former Yale professor who had left academia to become a pollster focusing on progressive-liberal politicians and causes. Greenberg conducted a monthly tracking poll for Clinton, whose use of polling was on par with Reagan.

But Clinton also had a longstanding relationship with former Republican strategist and pollster Dick Morris. In 1994, when Clinton faced a disastrous midterm for the Democratic Party, the conservative Morris took on a greater role in Clinton’s circle. He brought in Mark Penn and Doug Shoen, Harvard dormmates who, like Caddell, began their polling work while still in school. Penn had, in fact, started his polling career in high school. Penn and Shoen replaced Greenberg and conducted a large polling operation that Clinton followed closely.

Recent Presidential Polling: Obama through Biden

While a few presidential pollsters had high public profiles, most have been far from household names. Their importance to White House operations in particular has been revealed through archival records, spending reports, and other research.  While the kind of deep dive into presidential polling that has been made facilitated by the archives at presidential libraries and other records has not yet been possible for the most recent presidents, contemporaneous reporting does provide some limited information about the polling operations from 2008 to today. 

Obama’s campaign polling operation included six pollsters, and during his administration he relied on Joel Benenson, Paul Harstad – both former journalists turned pollsters - and David Binder, a California-based political pollster. Like Bush, Obama preferred a hands-off approach to polling data, with David Axelrod serving as his intermediary.

Donald Trump’s primary pollster has been Tony Fabrizio, who had run polling operations for Republican candidates dating back to Bob Dole’s election. He had some reservations about accepting the position, because Trump “doesn’t like pollsters.” Nonetheless, Fabrizio has provided polling for all three of Trump’s presidential elections.

Joe Biden turned to longtime Democratic pollster John Anzalone of ALG Research and Impact Research, who had been worked on Obama’s and Hillary Clinton’s campaigns. The groups responsible for the polling aspects of Kamala Harris’ data operations have not yet been reported. 

This overview draws heavily from the Robert M. Eisinger’s 2003 book The Evolution of Presidential Polling. Other sources utilized and recommended for further research include:

Digital Collections:

Mark Penn digital polling data collection and oral history

Archival records at presidential libraries can be searched in the NARA catalog.

Other Sources

Altschuler, B. E. (1995). The First Modern Presidential Campaign: Polling the Primary Voters for JFK. American Review of Politics, 16, 185-200.

Altschuler, B. E. (1986). Lyndon Johnson and the public polls. Public Opinion Quarterly, 50(3), 285-299.

Bowers, D., Roberts, G., & Schulman, M. (2009). Harry O’Neill, 1929–2008. Public Opinion Quarterly, 73(3), 599-601. 

Brown, M. (1961). The demise of state department public opinion polls: a study in legislative oversight. Midwest Journal of Political Science, 5(1), 1-17.

Druckman, J. N., & Jacobs, L. R. (2006). Lumpers and splitters: The public opinion information that politicians collect and use. International Journal of Public Opinion Quarterly, 70(4), 453-476.

Druckman, J. N., & Jacobs, L. R. (2020). Who governs? Presidents, public opinion, and manipulation. University of Chicago Press.

Fried, A., & Harris, D. B. (2010). Governing with the Polls. The Historian, 72(2), 321-353.

Green, J. (2002). The Other War Room President Bush doesn't believe in polling-just ask his pollsters. Washington Monthly, 34(4), 11-16.

Halberstam, D. (2000). The powers that be. University of Illinois Press.

Heith, D. J. (1998). Staffing the White House public opinion apparatus 1969-1988. Public Opinion Quarterly, 165-189.

Holli, M. (2002). The wizard of Washington: Emil Hurja, Franklin Roosevelt, and the birth of public opinion polling. Springer.

Jacobs, L. R., & Shapiro, R. Y. (1995). The rise of presidential polling the Nixon White House in historical perspective. Public Opinion Quarterly, 59(2), 163-195.

Jacobs, L. R., & Shapiro, R. Y. (1994). Issues, candidate image, and priming: The use of private polls in Kennedy's 1960 presidential campaign. American Political Science Review, 88(3), 527-540.

Johnson, D. W. (2016). Democracy for hire: A history of American political consulting. Oxford University Press.

Levy, M. R. (1984). Polling and the presidential election. The Annals of the American Academy of Political and Social Science, 472(1), 85-96.

Rottinghaus, B. (2003). Reassessing public opinion polling in the Truman administration. Presidential Studies Quarterly, 33(2), 325-332.

Schneider, E. (2024, April 3) Biden campaign beefs up its polling operation. Politico. https://www.politico.com/news/2024/04/03/biden-campaign-polling-operation-00150337#:~:text=David%20Binder%2C%20who%20specializes%20in,day%2Dto%2Dday%20operations

Shapiro, R. Y., & Jacobs, L. R. (2001). Source material: Presidents and polling: Politicians, pandering, and the study of democratic responsiveness. Presidential Studies Quarterly, 31(1), 150-167.

Murray, S. K., & Howard, P. (2002). Variation in White House Polling Operations: Carter to Clinton. The Public Opinion Quarterly, 66(4), 527–558. http://www.jstor.org/stable/3078707

Smith, B. (2009, April 3). Meet Obama’s pollsters. Politico. https://www.politico.com/story/2009/04/meet-obamas-pollsters-020852

Smith, M.J. (2019, September 29) Patrick Caddell: The pollster who foretold an alienated electorate. Politico. https://www.politico.com/news/magazine/2019/12/29/patrick-caddell-pollster-obituary-086490

Sudman, S. (1982). The presidents and the polls. Public Opinion Quarterly, 46(3), 301-310.

Tenpas, K. D. (2003). Words vs. deeds. The Brookings Review, 21(3), 32.

Tenpas, K. D., & McCann, J. A. (2007). Testing the permanence of the permanent campaign: An analysis of presidential polling expenditures, 1977–2002. Public Opinion Quarterly, 71(3), 349-366.