Thursday, October 11, 2012

Polls: Random Noise

Everybody, at one time or another, has freaked out over the polls. Needlessly. That doesn't mean that Obama did himself any favors by trying to take a nap during the presidential debate nor Romney by dissing half the country.
  • It does mean that opinion polls also register a willingness to talk. Only 9% of people contacted by pollsters actually respond. Ninety-one percent just hang up. It doesn't take much movement in the hang up numbers because of momentary disappointment or enthusiasm to influence the polling results. 
  • The best polls have a margin of error plus or minus 3% meaning a poll number of 47% is really anywhere between 44 and 50 percent.
  • Order of questions makes a big difference in results. If they ask "who won the debate?" first it will influence how some people answer "who are you voting for?"
But the biggest reason the polls are random noise is that most of the states don't matter. Romney has a huge lead in the Confederacy and the Grange states. Obama has a similar massive lead on the left and right coasts. Those leads could be by ten, twenty, even thirty points and it doesn't make a difference. What matters are those in between states - the Great Lakes, southern Rockies, and Florida.

And I have to believe those poor people are being drowned in campaign calls and polling calls. I'd be shocked if the hang up rate in Ohio is under 95%. Each pollster is dialing 20,000 Buckeye phones just to get enough responses to make up a polling result. And there are over a dozen pollsters in the field harassing Ohioans.

Bottom line. It is a close election where Get Out the Vote and Voter Suppression programs will make the difference.

No comments: