We are not going to stop loving political polling

By Richard Barry

In noting how badly pre-election polls underestimated the Conservatives' performance in the U.K. general election and overestimated Labour's result, Nate Silver at FiveThirtyEight writes that "it's becoming harder to find an election  in which polls did all that well."

He sites four other recent examples of polling problems: underestimating the size of the "no" side in the Scottish independence referendum; underestimating the GOP's performance in 2014 U.S. Senate races; badly underestimating Likud's performance in recent Israeli legislative elections; and underestimating Obama's performance in the 2012 general election by about three points nationwide. Silver points out that if that error went the other way, Romney would have won the popular vote and perhaps also the Electoral College.

Silver suggests some possible reasons for the problem:

Voters are becoming harder to contact, especially on landline telephones. Online polls have become commonplace, but some eschew probability sampling, historically the bedrock of polling methodology. And in the U.S., some pollsters have been caught withholding results when they differ from other surveys, “herding” toward a false consensus about a race instead of behaving independently.

This is not the first time we have heard concerns raised about the accuracy of polling, but we continue to rely on them somewhat uncritically and base our analysis on what we think they tell us.  

Well, public polling companies aren't going to go out of business anytime soon, but when a politician finds himself or herself on the lousy end of a survey, maybe we shouldn't laugh when she or he says that the only poll that counts is the one held on Election Day.

Having said that, I love polls and will continue to treat them like Gospel truth because they give us so much to talk and write about, and isn't that the point?

And besides, who likes going into Election Day not knowing the likely outcome? Way too nerve-racking.