George Gallup

George Gallup

George Horace Gallup, the son of George Henry Gallup, a dairy farmer, was born in Jefferson, Iowa, on 18th November, 1901. While he was at school he started up his own newspaper.

Gallup studied at the University of Iowa and his doctoral thesis in 1928 involved developing techniques to measure newspaper readership. After leaving university he moved to Des Moines where he served as head of the Department of Journalism at Drake University.

In 1931 he moved to Evanston, Illinois, as a professor of journalism and advertising at Northwestern University. The following year he joined Young and Rubicam, an advertising agency in New York City. One of his first tasks was to set up a copy research department to determine what kinds of advertising were reaching people.

Gallup's mother-in-law was active in the Democratic Party and in 1932 he carried out a political survey to discover her chances of being elected secretary of state for Iowa. In 1935 he formed his own polling company, the American Institute of Public Opinion. Gallup later explained: "The American Institute of Public Opinion, a non-partisan fact-finding organization which will report the trend of public opinion on one major issue each week". According to Michael Wheeler, the author of Lies, Damn Lies, and Statistics: The Manipulation of Public Opinion in America (2007): "Gallup did not start polling professionally until 1935, but everything he did before then, the journalism and the advertising in particular, had a strong influence on his orientation as a pollster. Much of Gallup's success is attributable to his understanding of what sells newspapers, as well as his own gift for self-promotion. When Gallup began, there were no pollsters as such. He did not pursue a career; rather, he created one."

Gallup's first survey, carried out in October, 1935, concerned the policies of President Franklin D. Roosevelt. Gallup argued that he collected this information "by means of personal interviews and mail questionnaires from thousands of voters located in every state in the union... persons in all walks of life have been polled in order to obtain an accurate cross section." Gallup's first poll showed that 60% of those questioned thought that the government was spending too much money on "relief and recovery". A second survey showed that Roosevelt's support had declined dramatically since the 1932 Presidential Election. These poll results were published in leading newspapers.

The 1936 Presidential Election was the first where polls were published in newspapers during the campaign. The Literary Digest sent out 10,000,000 questionnaires to voters. Gallup used a very different strategy: "The sampling procedure described is designed to produce an approximation of the adult civilian population living in the United States, except for the persons in institutions such as prisons or hospitals... The places (where people were interviewed) were selected to provide broad geographic distribution within states and at the same time in combination to be politically representative of the state or group of states in terms of three previous elections."

The Literary Digest predicted that Alfred Landon (57%) would defeat Roosevelt (43%). Gallup's survey showed Roosevelt winning with 55.7% of the vote. Although this was 4% less than Roosevelt actually achieved, Gallup's more scientific method was deemed to be a success. It later emerged the magazine reported its results based on only 15% of the questionnaires set out. It seems that Landon's supporters were much more willing to send back their questionnaires and therefore distorting the results.

Gallup was a member of the Republican Party and he was accused of under-reporting Roosevelt's support in order to influence the final result. Gallup defended himself by arguing that "we pride ourselves on being fact-finders and scorekeepers, nothing else." To demonstrate his neutrality he said that he never voted in elections where he had carried out surveys: "It would be as if the referee at the next Princeton-Yale game announced loudly that he was in favour of Princeton. Everything he would do would be a little suspect."

In 1939 Gallup employed David Ogilvy to work for the Audience Research Institute. Ogilvy later claimed that it was the luckiest break of his life "as it furnished him with immeasurably useful knowledge about the techniques of marketing research, as well as about what made United States citizens really tick". The following year Ogilvy was recruited as an agent by William Stephenson, the head of the British Security Coordination (BSC). As William Boyd has pointed out: "The phrase (British Security Coordination) is bland, almost defiantly ordinary, depicting perhaps some sub-committee of a minor department in a lowly Whitehall ministry. In fact BSC, as it was generally known, represented one of the largest covert operations in British spying history... With the US alongside Britain, Hitler would be defeated - eventually. Without the US (Russia was neutral at the time), the future looked unbearably bleak... polls in the US still showed that 80% of Americans were against joining the war in Europe. Anglophobia was widespread and the US Congress was violently opposed to any form of intervention." An office was opened in the Rockefeller Centre in Manhattan with the agreement of President Franklin D. Roosevelt and J. Edgar Hoover of the FBI.

One of Ogilvy's tasks was to persuade Gallup from publishing polls considered harmful to the British. As Richard W. Steele has pointed out: "public opinion polls had become a political weapon that could be used to inform the views of the doubtful, weaken the commitment of opponents, and strengthen the conviction of supporters." William Stephenson later admitted: "Great care was taken beforehand to make certain the poll results would turn out as desired. The questions were to steer opinion toward the support of Britain and the war... Public Opinion was manipulated through what seemed an objective poll." According to Thomas E. Mahl, the author of Desperate Deception: British Covert Operations in the United States, 1939-44 (1998): "BSC persuaded Gallup... to drop the results of questions that reflected poorly on the British cause."

Michael Wheeler, the author of Lies, Damn Lies, and Statistics: The Manipulation of Public Opinion in America (2007) has argued: "Proving that a given poll is rigged is difficult because there are so many subtle ways to fake data... a clever pollster can just as easily favor one candidate or the other by making less conspicuous adjustments, such as allocating the undecided voters as suits his needs, throwing out certain interviews on the grounds that they were non-voters, or manipulating the sequence and context within which the questions are asked... Polls can even be rigged without the pollster knowing it.... Most major polling organizations keep their sampling lists under lock and key."

David Ogilvy worked closely with Hadley Cantril, who was secretly working for President Franklin D. Roosevelt. In a secret report written by a team of BSC agents that included Roald Dahl and H. Montgomery Hyde, pointed out: "As the campaign against Fifth Columnists in the US continued, BSC was able, through the Gallup Poll, to see how its progress was affecting American public opinion. The results, as polled by Gallup, were most gratifying. On 11 March, only 49% of the American people thought that Britain was doing her utmost to win the war. On 23 April, this proportion had jumped to 65%, although no important naval or military victory had occurred during this period to influence the public in Britain's favour. Gallup's assistant (Ogilvy), who eventually joined the staff of BSC, was able to ensure a constant flow of intelligence on public opinion in the United States, since he had access not only to the questionnaires sent out by Gallup and Cantril and to the recommendations offered by the latter to the White House, but also to the findings of the Survey Division of the Office of War Information and of the Opinion Research Division of the US Army."

Gallup was a close friend of Thomas Dewey and tried very hard to make him the Republican Party candidate in the 1944 Presidential Election. The British Security Coordination secret report claimed that the "Gallup Poll did not prove a reliable guide to the Presidential election of 1944. This was so, largely because Gallup is himself a Republican and a staunch supporter of Dewey. As William Stephenson learned, there was little doubt that Gallup deliberately adjusted his figures in Dewey's favour in the hope of stampeding the electorate." Ernest Cuneo told Stephenson: "Dewey is one of Gallup's principal clients... Dewey is calling up Gallup so often they have to have a clerk to answer him."

Gallup was also accused of trying to get Thomas Dewey elected against Harry S. Truman in the 1948 Presidential Election. Truman lost all nine of the Gallup Poll's post-election surveys. In late September, Dewey had a 17 point lead. According to Albert E. Sindlinger, who worked for Gallup, claimed that "before the 1948 election Dewey and Gallup were on the phone constantly. Dewey was looking for a handle on public opinion and he turned to George Gallup." Sindlinger says Gallup deliberately rigged the polls to favour Dewey. "Gallup's sample excluded people who hadn't voted before. I found that they were heavily pro-Truman, but Gallup just didn't count them." Sindlinger added: "We'd set up the headlines and draft the story, and then we would go out and do the surveys to fill in the gaps. If the results squared with our story, we'd congratulate ourselves on how smart we were. But if they didn't, then the data would be adjusted, supposedly because there was something wrong with the sample."

Gallup was severely embarrassed by Truman winning with 49.6%of the vote compared to 45.1% for Dewey. Sindlinger believes that Gallup's biased polls helped to defeat Dewey as it made the Republicans over-confident. Sindlinger admits that during the campaign he came across a lot of people who said they would not bother to vote because Dewey was a certainty: "pollsters may deny it, but if you look at the evidence it's overwhelmingly clear that polls do influence people."

A number of his subscribing newspapers threatened to cancel their contracts because Gallup's polls did not reflect the result. Gallup replied that scientific surveys be expected to take into account "bribery of voters" and "tampering with ballot boxes"? By 1950 Gallup's market research business picked up. Gallup later explained: "The fact that the polls would recover was in my mind absolutely inevitable. The one thing that sustained me was the fact that no one had ever found a better system for understanding public opinion and I didn't think anyone ever would."

David Ogilvy published a book, Confessions of an Advertising Man in 1963. It sold over 400,000 copies and made him the only advertising figure whose reputation went far beyond that of the industry. It included a great deal about his work with Gallup. For example: "Dr. Gallup is a fountain of useful information on how people react to different kinds of commercials. He tells us that commercials which start by setting up a problem, then wheel up your product to solve the problem, then prove the solution by demonstration, sell to four times as many people as commercials which merely preach about the product. Dr. Gallup also reports that commercials with a strong element of news are particularly effective. So you should squeeze every drop of news value out of the material available for your commercials..... Dr. Gallup has discovered that the kind of photographs which win awards from camera clubs - sensitive, subtle, and beautifully composed - don't work in advertisements. What do work are photographs which arouse the reader's curiosity... He glances at the photograph and says to himself, What goes on here? Then he reads your copy to find out. This is the trap to set."

Michael Wheeler, the author of Lies, Damn Lies, and Statistics: The Manipulation of Public Opinion in America (2007) has argued that it was generally accepted by politicians that Gallup was secretly working for the Republican Party. He points out that Charles Colson worked for Richard Nixon in the 1968 Presidential Election: "Colson also hoped to develop a fruitful relationship with the Gallup organization, but gave that a much lower priority than his courting of Harris, because Gallup was already thought to be sympathetic. Gallup has always been tagged as a Republican partisan - his protestations of neutrality notwithstanding - in part because he overestimated the Republican vote in each of the first four presidential elections in which he polled, and in part because many of his polling associates have worked for Republican candidates. One of his oldest associates, Claude Robinson, ran Nixon's polling operation in 1960, and Robinson's firm, Opinion Research, polled for Nixon in 1968. During that 1968 campaign, Nixon had a source within the Gallup organization who provided advance word on when the surveys were going to be taken. This allowed Nixon to time his activities so that they would have the maximum impact on Gallup's polls."

George Horace Gallup died of a heart-attack at his summer home in Tschingel on 26th July, 1984.

Primary Sources

(1) Michael Wheeler, Lies, Damn Lies, and Statistics: The Manipulation of Public Opinion in America (2007)

Gallup did not start polling professionally until 1935, but everything he did before then, the journalism and the advertising in particular, had a strong influence on his orientation as a pollster. Much of Gallup's success is attributable to his understanding of what sells newspapers, as well as his own gift for self-promotion. When Gallup began, there were no pollsters as such. He did not pursue a career; rather, he created one.

(2) David Ogilvy, Blood, Brains and Beer: The Autobiography of David Ogilvy (1978)

I had been moonlighting as advisor to the British government on American Public Opinion, but it was time I played a more active part... I could not have had a better boss than Dr. Gallup. His confidence in me was such that I do not recall his ever reading any of the reports I wrote in his name. Once he had worked out the methodology of the research, he lost interest and moved on to something new.

(3) Roald Dahl, H. Montgomery Hyde, Giles Playfair, Gilbert Highet and Tom Hill, British Security Coordination: The Secret History of British Intelligence in the Americas, 1940-45 (1945)

As the campaign against Fifth Columnists in the US continued, BSC was able, through the Gallup Poll, to see how its progress was affecting American public opinion. The results, as polled by Gallup, were most gratifying. On 11 March, only 49% of the American people thought that Britain was doing her utmost to win the war. On 23 April, this proportion had jumped to 65%, although no important naval or military victory had occurred during this period to influence the public in Britain's favour.

Gallup's assistant, who eventually joined the staff of BSC, was able to ensure a constant flow of intelligence on public opinion in the United States, since he had access not only to the questionnaires sent out by Gallup and Cantril and to the recommendations offered by the latter to the White House, but also to the findings of the Survey Division of the Office of War Information and of the Opinion Research Division of the US Army. The mass of information which BSC collected in this way was obviously of interest to London. But it was most immediately useful in helping the British Information Services, the Embassy and the Consulates throughout the country to plan effective counter-measures against anti-British propaganda in the United States. The BSC reports were described by one Department of the Embassy as "the most reliable index of Anglo-American relations available".

Gallup himself was by no means unreservedly pro-British, but BSC's contact was able to dissuade him from publishing the results of certain polls which would have had a damaging effect on British prestige. It would have been unfortunate, for instance, if Gallup had released to the hundred or more newspapers which published his findings the fact that only 50% of the British people believed that Britain was doing her utmost to win the war and only 54% believed that America was doing hers. Yet these were the results of a poll conducted by Gallup's representative in England in 1942. Nor, again, could it have proved other than harmful had it become generally known that a large number of Americans were in favour of immediate self-government for India and of the formation of a Palestinian army.

(4) David Ogilvy, Confessions of an Advertising Man (1963)

When I worked for Dr. Gallup, I was able to demonstrate that moviegoers are more interested in actors of their own sex than in actors of the opposite sex. True there are a few exceptions to this rule: the female sex-kittens find great favour with male moviegoers and the lesbian stars do not appeal to men. But in general, people take more interest in movie stars with whom they can identify. In the same way, the cast of characters in most people's dreams contain more people of their own sex than of the opposite sex. Dr. Calvin Hall reports that "the male-female character ratio in male dreams is 1.7 to 1."

(5) David Ogilvy, Confessions of an Advertising Man (1963)

Dr. Gallup is a fountain of useful information on how people react to different kinds of commercials. He tells us that commercials which start by setting up a problem, then wheel up your product to solve the problem, then prove the solution by demonstration, sell to four times as many people as commercials which merely preach about the product. Dr. Gallup also reports that commercials with a strong element of news are particularly effective. So you should squeeze every drop of news value out of the material available for your commercials.

(6) Michael Wheeler, Lies, Damn Lies, and Statistics: The Manipulation of Public Opinion in America (2007)

Charles Colson also hoped to develop a fruitful relationship with the Gallup organization, but gave that a much lower priority than his courting of Harris, because Gallup was already thought to be sympathetic. Gallup has always been tagged as a Republican partisan - his protestations of neutrality notwithstanding - in part because he overestimated the Republican vote in each of the first four presidential elections in which he polled, and in part because many of his polling associates have worked for Republican candidates. One of his oldest associates, Claude Robinson, ran Nixon's polling operation in 1960, and Robinson's firm, Opinion Research, polled for Nixon in 1968. During that 1968 campaign, Nixon had a source within the Gallup organization who provided advance word on when the surveys were going to be taken. This allowed Nixon to time his activities so that they would have the maximum impact on Gallup's polls.