Straw Vote

One of the early types of preelection polls, the straw vote, got its name from the farmers' trick of throwing a handful of straw in the air to see which way the wind was blowing. For more than a century, newspaper and magazine straw polls were the leading predictors of major elections.

The first published presidential poll was in the Harrisburg Pennsylvanian on July 24, 1824. The publication's [Page p468]straw vote among people in Wilmington and Newark, Delaware, showed a preference for Andrew Jackson over John Quincy Adams. Jackson did win the popular vote but lost to Adams in the second presidential election decided by the U.S. House. (See President, nominating and electing.)

In 1866 the Cleveland Leader reported that a straw vote taken on a train showed passengers favoring Congress over President Andrew Johnson in their conflicting views on Reconstruction. Two years later the conflict peaked when Johnson became the first president to be impeached.

By 1896 straw polling had become fairly scientific. For the presidential election that year the Chicago Tribune polled railroad and factory workers and found 80 percent supporting Republican William McKinley over Democrat William Jennings Bryan, who had been expected to win the labor vote.

Another Chicago paper, the Record, conducted an even more scientific poll. It mailed ballots to all 300,000 registered voters in the city and, based on the responses, predicted McKinley would win in Chicago with 57.95 percent of the vote—matching almost exactly the actual vote of 57.91 percent.

Basically there were three ways to conduct a straw vote: by printing a ballot in the newspaper; mailing ballots to all registered voters or to a random sample of them, such as every twelfth voter; or sending canvassers into communities to have residents fill out ballots then and there. Of the three methods, the first was the least reliable because it was subject to ballot stuffing: people could buy extra papers and “vote” many times. Mailed ballots were safer but they could be counterfeited, and sometimes were. Personal canvassing was the most reliable method.

Newspapers became quite skilled in taking straw votes. Polls by the Cincinnati Enquirer, for example, mirrored actual results in presidential elections from 1908 to 1928.

The most famous straw poll, however, was by a magazine, the Literary Digest. It began presidential polling in 1924 and gained a reputation for uncanny accuracy by correctly predicting that election and the next two, including Franklin D. Roosevelt's unexpected defeat of incumbent Herbert Hoover in 1932.

The Digest poll was a massive undertaking, requiring millions of ballot mailings. The mailing list, however, was derived mostly from telephone directories and motor vehicle registrations, which were unrepresentative of the U.S. population in the Great Depression year of 1936 because only the wealthier households had phones or cars. The Digest suffered a major embarrassment that year with its erroneous prediction that Kansas governor Alfred M. Landon would unseat Roosevelt.

In the same election a newcomer, George Gallup, gained credibility at the Digest's expense with his prediction that Roosevelt would win a second term. Although Gallup underestimated Roosevelt's 60.8 percent vote—the biggest presidential landslide up to that time—his methodology worked and his reputation was established.

...