1936 Presidential Election

The Literary Digest was the most widely-read weekly magazine in America in 1936, with a circulation of over a million. It had correctly called the winner of every presidential election since 1916. For 1936, it mailed out 10 million postcards and received responses from 2.4 million people — the largest pre-election poll ever conducted. It predicted Landon would win with 57% of the vote. FDR won 61%.

The disaster had a simple cause: the Digest drew its sample from telephone directories and automobile registration lists. In Depression-era America, owning a telephone or a car was a marker of middle-class or upper-class status. The poll systematically excluded the working class and the poor — precisely the voters who were most enthusiastic about FDR’s New Deal. Self-selection also played a role: Landon supporters, motivated by their intense opposition to FDR, were more likely to mail back the postcards.

George Gallup, by contrast, used a much smaller sample of only a few thousand respondents selected to represent the population proportionally. He correctly predicted FDR’s landslide — and also correctly predicted the Literary Digest would get it wrong. The contrast between Gallup’s accurate small sample and the Digest’s catastrophically wrong huge sample established the principle that representative sampling matters far more than sample size. The Literary Digest folded within two years. Modern polling was born.

By 1936 the New Deal had transformed the relationship between Americans and their federal government. The Civilian Conservation Corps had employed 500,000 young men. The Works Progress Administration was building roads, schools, and public buildings across the country. Social Security had been enacted. The National Labor Relations Act had protected workers’ right to unionize, and union membership was surging. While unemployment remained high at around 16%, it had fallen sharply from the 25% of 1932 and most Americans credited FDR with the improvement.

Alf Landon, the moderate Republican Governor of Kansas, was not an unreasonable man. He accepted most of the New Deal while criticizing its inefficiency and cost. But “a more efficient New Deal” was not a compelling alternative to the actual New Deal. FDR had rebuilt the Democratic base into a governing majority that included organized labor, African Americans in the North (shifting Democratic for the first time in large numbers), Southern whites, urban Catholics and immigrants, and Western farmers.

FDR’s post-election overconfidence led to his biggest mistake: the 1937 court-packing scheme, in which he proposed to expand the Supreme Court with additional justices who would uphold New Deal legislation (the Court had struck down several programs). The plan was seen as an attack on judicial independence and failed in Congress, fracturing the New Deal coalition and energizing the opposition. The 1938 midterms saw significant Republican gains, ending the peak New Deal era.

The 1936 election is the benchmark for electoral landslides. When politicians speak of “the biggest win since Roosevelt,” they almost always mean 1936. The 523-8 Electoral College margin has never been matched in a contested race. Lyndon Johnson’s 486-52 win in 1964 and Richard Nixon’s 520-17 win in 1972 came close but not quite.

The Literary Digest catastrophe remains the canonical example of how large sample sizes cannot compensate for biased sampling methods. The lesson — that the representativeness of a sample matters more than its size — is foundational to modern polling methodology and appears in virtually every introductory statistics textbook. Polls that overrepresent one demographic group will always produce distorted results, regardless of how many people are sampled.

The 1936 coalition also represents the high-water mark of a political alignment that has been slowly dissolving ever since. The combination of Southern whites, Northern labor, African Americans, and urban Catholics that delivered FDR’s landslide began fracturing with Truman’s civil rights moves in 1948, accelerated with the Civil Rights Acts of the 1960s, and is now completely rearranged. Understanding the New Deal coalition requires understanding the 1936 election that proved its strength — and the structural tensions that would eventually pull it apart.