Two students correctly predicted election results—when professional statisticians did not

They relied on past elections more than polls

Out of 193 student entrants into a prediction contest for the 2016 election, only two correctly forecasted that Donald Trump (R) would win.

Mary Beth Benzing and Sejla Palic, seniors at St. Lawrence University, predicted on October 30 that Trump would win 36 states and tie in New Hampshire. Their forecast ended up being off by three states—Trump won 33 states—but was nonetheless far more accurate than the other 191 entrants predicting Hillary Clinton's victory.

The students chalk up their success to past election results, which they more heavily took into account than polling results. 

Use the right data to make decisions

"Everyone was putting weight on 2016 polling, but we used a lot of data from 2012 because we thought it was most relevant," says Benzing.

Michael Schuckers, the girls' statistics professor, says he's glad his students succeeded, but is worried that so many professional statisticians failed.

"[Statisticians] have to reconsider how we model these sorts of data," says Schuckers.

"Polls are very wrong," says Yashelle Hunte, a student at Magruder High School in Rockville, Md. who incorrectly predicted Clinton would win. "People tell the pollsters how they're going to vote, but it's not set in stone. They go out to vote and change their minds" (Heim, Washington Post, 10/9).

One school has created a culture where numbers matter

Next in Today's Briefing

How colleges in 6 states are grappling with funding cuts

Next Briefing

  • Manage Your Events
  • Saved webpages and searches
  • Manage your subscriptions
  • Update personal information
  • Invite a colleague