Scott's Pet Peeve #6914

Scott Hardie | May 10, 2016
It's campaign season again (ha! I kid! campaign season never ends), so I've been reading a lot of polls and analysis of polls and analysis of analysis of polls. No, it's not all Five Thirty Eight.

Anyway, a point keeps getting made in the various columns and blogs that I read: This far out from election day, polls are a poor indicator of what will happen at the ballot box, so they have no value. It's true that a lot can change in the months before Election Day, and anyone can pull off a surprise win that defies the polls, even polls taken the day before. It even happened a couple of times in the recent primaries.

But the false assumption that irritates me is that a poll's only value is predicting what will happen later in the election. That kind of speculation is fun, but to me the primary value of a poll is to quantify each candidate's level of acceptance/preference with the American people right now. If I hear that, say, Clinton is up five points and Trump is down five points, my thought isn't, will this hold true in November; my thought is that's how people feel about them right now. Polls are especially useful at showing what happens after different events in the news, from external events like terrorist attacks to internal events like Trump saying some other crazy thing that revs up his base. I want the voters' temperature taken because I like knowing how they feel today, and I don't expect the numbers to stay static in the future. Why does anybody else?

Samir Mehta | May 11, 2016
[hidden by request]