Sunday, September 23, 2012

Correcting the Polls

Over and over I hear the same complaint about the polling data being foisted off on the public by the polling companies. Every time the problem resolves to the same thing: Oversampling of Democrats. Folks there is an obvious solution to this. When a pollster calls, answer the phone, and answer the questions honestly. Except for the one when they ask you about your party affiliation.

Tell them you're a Democrat. Not just any generic Democrat, but a fervent, dedicated, full-of-kool-aid Democrat. They might have some difficulty reconciling some of your answers with this, but that's someone elses job, not yours.

Push polls can be especially fun:
Q: Whom did you vote for in the last elestion, Candidate D or candidate R?
A: Candidate R.

Q: If the election were held today, would you vote for candidate D or candidate R?
A: Candidate R.

Q: If you knew that candidate R squashed baby ducks with a hammer and drowned kittens before cooking and eating them would this influence your feelings for him?
A: Not at all.

Q: Do you consider yourself a D or an R or something else?
A: A D of course.

Q: Do you consider yourself a strong D, a middling D or a moderate D?
A: Strong D.

Q: What is the likelihood that you'll be voting in the upcoming election?
A: I voted here, in Chicago, and in Florida in the last election, and I have no intention of shirking my civic duty in this one.

So when they over sample D's in this poll, your opinion will be counted. It also helps support the narrative that the D+6 bias in party identification we saw in 2008 is still there, and 2010 didn't happen.

No comments: