Of Polls and Pundits

Polls seem to be in the news a lot lately. I had a post not too long ago where I groused a bit about the quality of the sampling data in the polls, but grousing about polls in general seems to have spread to many places.

In the past couple of weeks, there has been some significant movement in the polls. Mr. Obama got a small polling bounce after the Democratic Convention. That dissipated after about a week to a tie. Now it seems that there has been some movement towards Mr. Obama, enough for the polling aggregate site Real Clear Politics to move Ohio to "Lean Obama".

The reaction to these polls has been heavy on the extreme ends. Supporters of Mr. Obama are crowing about the polls and all but declaring the race over. Supporters of Mr. Romney are dismissing the polls, with some even declaring that the polls are all rigged and only those by Rasmussen are worth anything.

The answer is probably somewhere in middle. No one can know what the final polling sample is and much of the statistical manipulation of the raw data is driven by the internal biases of the pollster. So it is probably true that many pollsters either actually think (or at least hope) that turnout will be similar to that of 2008.

On the other hand, reducing your scope to a single pollster (because you like their results), you blind yourself to broader trends. Even with a skewed poll, if the skew is relatively consistent, you can observe movements that generally reflect reality.

The only poll that really matters is the one that is counted on the night of November 6th, so any handwringing, knashing of teeth, or premature celebrating before then is just wasted effort. There are three possiblities of where this race is right now:

1. The majority of polls are right and Mr. Obama is winning the race handly.
2. The majority of polls are a bit off and the race is a statistical tie.
3. The majority of polls are way off and Mr. Romney is winning the race handly.

As I reminded Daughter X after one of her soccer games when she talked about how her team won the third quarter, it doesn't matter who wins a quarter. It only matters who's ahead at the end of the game. The real problem is that we are not sure exactly of what the field is like to determine how the final score will be resolved.

In 2008, there were 7% more Democrats at the polls than Republicans (39-32) with Mr. Obama also winning Independent voters. In 2010, Democrats and Republicans came to the polls in equal numbers (35-35). My own personal take is that Democrats will end up having a 2 or 3 point advantage in turnout, but I could be completely wrong about that. That also doesn't tell you if the votes will match the partisan split or to whom the Independent vote will break.

The best anyone can do is to sit back, relax, watch trends (both short and long-term) and go vote for their preferred candidate on the 6th.