Saturday, January 16, 2010

Beware of people bearing polls

The entrail-readers have already begun issuing their interpretations of the Coakley-Brown race. Liberal commentators (like, say, my dad) often claim that because Obama hasn’t been liberal enough, it’s responsible for the problem Coakley has. Conservative commentators claim exactly the opposite. Centrist commentators think he should have been more centrist. Everyone has their pet theory, and the problem with polls is that they can support any kind of interpretation you want to draw.

Well, not any kind. You can believe people’s opinions in polls as to how they’ll vote, but not why they’ll vote in a particular way. It’s really difficult to do why just on polls alone, because then you’ll discover that the American people are—at least in the ways they answer poll questions--…confused about politics and their ideology.

Let’s take this Suffolk poll analyzed by a fivethirtyeight.com post (you’ll note that all the post does is analyze who’s ahead). According to this poll, there’s a non-trivial percentage of Massachusetts voters who believe these things: they like Obama, Scott Brown, Martha Coakley, Bill Clinton, support the Massachusetts health care law, believe it unaffordable, oppose the national health care law and believe it unaffordable. I’d describe that as confusion. And you can probably look at the crosstabs of any poll and come to similar conclusions.

The point here is to be wary of people bearing polls to support their own ideological proclivities. The cynical explanation is that they’ve got an agenda to shill for. But I suspect part of it is a failure of recognition: the entrail-readers, who are ideological, look at other people and see other ideologues like themselves, when those other people tend to say “Yes” to every question so they can move on from political matters. (Here’s a good essay on undecided voters by Chris Hayes.) It’s perilous then to look into their minds and assume a similar political junkie-dom and infer complicated exaplanation. People are strange, and explanations are difficult.

If I might do a little entrail-reading myself, here’s my explanation for why the race is as close as it is: Scott Brown might win because everyone thinks Martha Coakley will win. Question 28 asks, who do you think will win the race (in pollerese): 60% believe Coakley will win. Well, I suspect a pretty large percentage of those Coakley-will-win voters are casting Brown votes as an eff you to the powers that be. Please don’t take this explanation as an explanation.

No comments:

Post a Comment