The Worst U.S. Presidential Election Prediction in History

Mr. Beat briefly tells the story of the worst presidential election prediction in history.

Produced by Matt Beat.

Sources/further reading:

Moon, N. (1999). Opinion polls - History, theory and practice. Manchester: Manchester University Press.

Join this channel to get access to perks:

For business inquiries or to send snail mail to Mr. Beat:

How to support and donate to my channel:

Subscribe to @iammrbeat & hit the notification bell 🔔

Join for great perks on Patreon

Donate to Mr. Beat on Paypal

Buy Mr. Beat a coffee

Subscribe to my second channel: The Beat Goes On

Patreon for The Beat Goes On

Connect with me:

@beatmastermatt on Twitter

Merch:

Affiliate Links:

Useful Charts

In 1936, President Franklin Roosevelt ran for re-election. His main opponent? Alf Landon, the Governor of Kansas.

Throughout the year, a magazine known as The Literary Digest polled its readers, asking who they were voting for. Roosevelt or Landon? They sent out an impressive 10 million ballots and got back more than 2 million responses. According to the ballots, Landon would win the popular vote 57% to 43%. Indeed, they predicted Landon would win in a landslide.

Uhhh, but no. Absolutely not. In fact, the exact opposite happened. Roosevelt beat Landon in one of the most lopsided presidential election victories in American history.

The actual result? Roosevelt won 62% of the popular vote with Landon only getting 37%. The Literary Digest overestimated Landon’s support by 20 percentage points.

So why did The Literary Digest get it so wrong? Mainly because there was a participation bias, meaning the volunteers who actually responded were just generally not a good sample of the entire population. For example, it was mostly wealthier voters who responded, when most Roosevelt voters were not wealthy.

The Literary Digest lost a lot of credibility due to this election, and would stop publishing four years after the election.