In Our Opinion
George Gallup and the Iowa Polling Method
Story by Mary Gottschalk
The first recorded presidential straw poll took place nearly 200 years ago when a Pennsylvania newspaper predicted (incorrectly) that Andrew Jackson would best John Quincy Adams in the 1824 election. More than a century later, another presidential straw poll predicted (also incorrectly) that Alf Landon would beat Franklin Roosevelt in 1936.
George Gallup believed such straw polls were unreliable gauges of how the country as a whole would vote. In the view of this young pollster from Iowa, selecting respondents based on attributes such as phone number, voter registration, or the willingness to mail a survey back tended to over-represent higher-income households.
(He might argue today that the self-selected participants in the recent Republican straw poll — those able and willing to drive or be bussed to Ames — are unlikely to mirror the views of all American voters.)
Gallup instead used door-to-door interviews based on random sampling (i.e., everyone has an equal opportunity to participate) of American voters — and correctly projected Roosevelt’s election in 1936.
His work proved that a small, scientifically drawn sample would more accurately mirror the population than a large, unscientifically chosen group. Gallup’s approach fundamentally changed the way that public opinion research has been conducted ever since.
“The Iowa connection seems important,” notes Ann Selzer, who conducts the Iowa Poll for the Des Moines Register and was the only professional pollster to correctly call the Obama win in the Iowa caucuses.
“Growing up in an agricultural community, Gallup would have understood about sampling of crop yields and that ‘volunteer corn’ (corn that has sprung up spontaneously from old seeds) can materially distort your yield estimates.”
In Gallup’s view, statistically reliable polling is key to the democratic process. “If government is supposed to be based on the will of the people,” he said, “somebody ought to go out and find out what that will is.”
A foundational principle of his work was that public opinion pollsters — and the organizations that employ them — must be independent of the issues or candidates that are the subject of the polls.
Born in Jefferson, Iowa, in 1901 and educated at the University of Iowa, George Gallup conducted his first published scientific survey in Des Moines. His first political poll addressed a statewide Iowa election.
Gallup, who seemed determined to challenge the status quo from an early age, began his career in journalism first at his high school paper and then as editor of the Daily Iowan.
At the University of Iowa he made a name for himself with a series of controversial editorials on such varied topics as the high proportion of unattractive women in the college, the self-complacency of fraternity men, the value of sex education, and the need for radical thinkers in a university setting.
By the time he was out of college, however, he seemed to be as interested in evaluating readers’ opinions as he was in influencing them. Between his junior and senior years, when he worked on a survey of newspaper readership, he started thinking about how to make surveys better.
Industry lore has it that he undertook a number of informal surveys during his school years. His first formal attempt at a statistically reliable survey — what he called the “Iowa method” — formed the basis of his 1928 Ph.D. dissertation on newspaper readership.
His findings (for example, that more people read the comics than the headlines) challenged conventional wisdom about the preferences of readers of both Des Moines newspapers, the Register and the Tribune.
He continued to do readership surveys while teaching journalism at Drake University in Des Moines. In Gallup’s view, nothing he ever did “generated as much instantaneous response” as a 1930 article he wrote on readership surveys.
That article led to an invitation to establish a market research department for Young & Rubicam, the largest advertising agency in the country, based in New York.
Not long after moving to the East Coast, he was drawn back to his Iowa roots by the electoral campaign of his mother-in-law. Ola Babcock Miller was running for secretary of state, a race she was not expected to win.
Gallup says he undertook “crude samples” of “what various and sundry groups of people thought.” His mother-in-law’s victory triggered Gallup’s next endeavor — using polls to capture changes in public opinion.
By 1934 Gallup believed that his methodology was reliable enough to warrant a national survey of public opinion on a weekly basis. In 1935 the American Institute of Public Opinion — an impressive name for what started out as a one-room office in Princeton, New Jersey — began publishing Gallup’s polling results on a variety of social, economic, and political topics in a syndicated newspaper column, “America Speaks.”
Gallup was not the only pollster to predict Roosevelt’s upset victory over Landon in 1936; however, he was the only one with the courage to publish his polling results before the election. The prediction thrust him onto the national stage. Several foreign languages now include gallup as a verb meaning “to take a poll.”
While Gallup’s key principles — random sampling and independence — continue to guide professional polling today, survey techniques have changed. Gallup’s door-to-door surveys — the “gold standard” of polling, according to Frank Newport, editor in chief at Gallup — are no longer feasible, due in part to cost and to growing security concerns of individual householders.
Most published polls today rely on phone surveys, an approach that requires new sampling procedures to ensure that cell phone-only households — typically younger, poorer, and less likely to be married than the population as a whole — are adequately represented.
Pollsters are exploring the possibilities of the Internet, but Newport notes that roughly 30 percent of the population does not have easy online access, and even for those who do, there is no systematic way to generate a random sample of Internet/email users. “We’re still reaching for the gold standard.”
Gallup in Hollywood
Audience response was one of the many types of market research that fascinated Gallup. Throughout the 1940s Gallup conducted some 5,000 surveys for movie producers, including big-name studios such as Disney and Columbia. His research allowed producers to test the box office appeal of story ideas and proposed stars, as well as to measure the effectiveness of publicity. He also developed the invitational preview, a research technique in which members of a target demographic group complete a questionnaire on their responses to an advance screening of a film.
Caucus Upon Us
At press time, the 2012 Iowa caucuses are likely to be here sooner than ever before, currently set for Tuesday, January 3, 2010. The first poll numbers predicting popularity started arriving two years ago.
George Gallup was an innovative thinker who refused to accept conventional ways of doing things. Just how much his unique perspective was stimulated by his unconventional childhood home we don’t know — but the eight-sided house on South Chestnut Street offers visitors an opportunity to ponder. Listed on the National Register of Historic Places in 1985, the octagonal Gallup House — a Craftsman/Victorian hybrid — was designed, so the builders said, to provide good air circulation and be safer in high winds.
The turn-of-the-20th-century home (Gallup was born in its parlor shortly after completion in 1901) was rennovated between 2007 and 2009. Numerous artists and craftspeople from Iowa and around the country added thier inspired works to the project. Learn more about these artisan furnishings on the Gallup House website.
The Gallup House
Available for retreats, conferences, receptions, and overnight stays.
703 South Chestnut Street, Jefferson
A Lengthened Shadow
Des Moines’ Sticks, Inc.,