Scientific Election Polling - 1936
A Crucial Turning Point in Presidential Political Polling
May 11, 2008
One of the most significant phenomenons in the history of presidential elections is the development and success of political polls. Today, mathematicians and statisticians develop entire careers calculating percentages of opinion favoring or rejecting one candidate or another. George Gallup, a name almost synonymous with opinion polls did a great deal to change the methods of presidential political polling in the United States.
Gallup Transformed the Way of Presidential Polling in 1936
George Gallup made a name for himself in the 1936 presidential election. At the time, the Literary Digest magazine, was respected as the top pollster in the U.S. They predicted incorrectly that Alf Landon would defeat Franklin Roosevelt in his re-election bid. Roosevelt won in a landslide that gave him a higher percentage of popular and electoral votes than he received in 1932. Gallup confidently predicted that the Literary Digest would get it wrong and F.D.R. would win. He got it right in a big way.
Although the Literary Digest had correctly predicted the outcomes of the previous five presidential elections, George Gallup understood the flaws in their polling method. The Digest had relied on the technique of conducting straw polls. Such polls basically queried a given number of people, posed the questions of who is and is not favored, and tabulated the responses. This seemed logical because straw polls were used for over 100 years in predicting the results of U.S. elections.
However, Alf Landon lost the election and the Digest lost most of the credibility they had established. Straw polls also lost status as a respected method for measuring public opinion in presidential elections. In this election, George Gallup altered the course of “business as usual” in polling methodology. Gallup’s newly formed polling service, the American Institute of Public Opinion, would grow to become a significant fixture in discerning public opinion.
A More Scientific Method of Opinion Polling
From the responses of approximately 3,000 to 5,000 surveys, Gallup’s minor polling service, predicted accurately. However, the Literary Digest received 2.3 million completed surveys from the 10 million surveys that had been sent out that year. This represented about one third of all households in the U.S. at the time. But, Gallup understood that large samples would not guarantee accurate predictions.
George Gallup, as well as other pioneers in the polling industry, utilized a method known as “quota sampling” which was a technique to obtain answers to surveys from selected groups of people representative of the population. In the 1936 election, Gallup sent hundreds of interviewers all over the country to obtain answers from different types of people. For example, they would interview so many men, so many women, groups based upon race, and based upon income level or social class, until their specific quotas were reached. With the results of these interviews, Gallup sparked a change in the way future political polls would be conducted.
Polling pioneers like Gallup, Archibald Crossley and Elmo Roper, utilized a statistical law that states that a random selection of the population will not require a large number of people selected in order to obtain an accurate assessment of the opinions of the entire populace. With a more scientific approach to opinion polling and a significantly smaller sample, Gallup and his peers successfully introduced the American people to a scientific method of polling public opinion. The industry grew significantly in importance from this turning point and eventually developed into an entrenched institution in the United States.
Sources:
American Government and Politics Today: The Essentials Bardes, Shelley, Schmidt, 2008, Thomson Wadsworth
