Why Nobody Wins With Post-Debate Polls

Last night's debate approached "Super Bowl status" in terms of viewership. But there's one big difference between a debate and the Super Bowl -- at the end of the Super Bowl, everybody knows who won.

Not so with a debate.

Which means it's up to the media to pick a winner. For journalists writing op-ed pieces, this is a relatively straightforward assignment. They watch the debate and crank out a few hundred words giving their opinion on the candidates.

The bigger challenge is for journalists who need to put their finger on the pulse of the electorate, and tell the public if the debate made a difference in the election or not.

But with this mad rush to crown a winner, too often speed trumps accuracy, no pun intended. Journalists and media outlets search for any sign that a candidate has won, and then (often mistakenly) imply that the results extend to the general electorate.


  • A headline in the Charlotte Observer claimed "Clinton loses ground among some voters in swing state." Yes, Clinton lost ground among some voters. Four of them, to be exact. The Observer and McClatchy had a focus group of 21 people watching the debate, and "four who were up for grabs before the debate moved away from her by the end."
  • The Drudge Report post-debate poll had Trump up 82% to Clinton's 18% as of this morning. We voted. And then we refreshed our browsers and voted again. And again. See the problem? But it's not just multiple voters that are a potential issue. As Philip Bump pointed out in The Washington Post, online polls are "open to anyone, meaning that anybody with an Internet connection can go and cast a vote." If you live in Russia, says Bump, you can vote in a U.S. post-debate poll. Same if you're a 12-year-old kid.
  • We didn't hear about Trump's proposed border wall last night, but we did see that the Mexican peso gained ground against the U.S. dollar, "indicating that traders think Clinton is in charge," according to Business Insider. Is that really what it means? Or could it mean that traders believed Trump's claim that more U.S. jobs were moving to Mexico? Looking at the numbers can't tell you what the traders were actually thinking.

These types of mistaken (if well-intentioned) interpretations of data occur almost every day. Whether it's a new study about vaping that's based on the actions of 16 people, or just stunningly bad graphs, misleading data is all around us.

If you really want to know who won the debate, wait a few days until the major (scientific-based) polls have a chance to weigh in. Until then, proceed with caution.