Polls I Like are Professional Because I Like Them; Polls You Like Are Unprofessional Because You Like Them

by Mike on October 22, 2012

in Politics,Uncategorized

It’s that maddening time of year when about 312 political polls a day are released.  Although a properly conducted poll can be a useful tool to gauge the mood of the electorate at a specific point in time, most of those who think they follow political news closely usually react depending on the results they want to see.

Liberals will jump all over Rasmussen because it’s a “right wing” pollster while ignoring the fact that it has been one of, if not the most accurate pollster of the day.  Some conservatives will jump all over Gallup numbers showing Romney with a big lead while ignoring the fact that the once great polling organization predicted the wrong popular vote winner in 2000 (there was actually nothing wrong with the final Gallup poll, it was within the margin of error; it’s just that the headline was a bit awkward), and then unscientifically allocated NEARLY ALL undecided voters to John Kerry in 2004 to create a tie and thereby avoid risking two “embarrassing” results in a row.  Gallup is still a fine poll to if you want to see a trend, but they have some work to do if they want to regain their credibility.

The easiest way to misread a poll is to assign value to one reporting what you want to see and discounting those who don’t.  Even many professionals do that.  Take Nate Silver, the most overrated analyst in political polling today.  His m.o. os to analyze political polls according to his secret “model” and then assign each candidate a percentage indicating that candidate’s probability of winning.  That’s actually the first red flag because a scientifically-valid poll is a snapshot of the electorate’s mood at a point in time.  Taken together, those snapshots can also present trends, but no legitimate pollster would ever interpret a poll to suggest that a candidate has an “x percent” chance of winning; that’s not what a poll does.  Today, at National Review’s website, Josh Jordan does an excellent job of dissecting Silver’s model, subjective and shifting weighting, and explaining that Silver’s model is really little more than Silver crediting the poll results he likes and discounting those he doesn’t.  Not to pick on the guy, but the MSM really puts Silver on a pedestal he doesn’t deserve.

Polls can be useful, but you really need to scrutinize the internals.  Too often, when in the fog of political war, we see what we want to see.  Well, some of us.  We’ve always been Rasmussen and Mason-Dixon fans on this site, regardless of whether the news is good or bad.  If either of those companies lose that edge, then of course, we’ll reevaluate.  Honesty requires it.

Also, be sure to check out Ace of Spades, (HT for the link to Jordan).  He makes a lot of really good points about Silver as well as the general relationship between state-level and national polling.

Leave a Comment

You can use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Previous post:

Next post: