Wednesday, January 05, 2005

MSM duped by statistical nuance.

The main stream media (MSM) have been licking their wounds, and hurling insults at the blogoverse, since they dropped the ball on the Presidental Elections Nov. 2. Liberals still insist that something funky happend when all the exit polls showed that Kerry was owning Bush, only to have the real vote turn out quite differently. Now, Slate is reporting that the polling company, Edison/Mitofsky, was quite up front about the fact that they were weighting the exit polls. This is something I pointed out prior to the election, thought I didn't have a good forum for mass communication (ie. my blog), to explain the grossly different poll results. Statistitions use statistical models to analyse their data. They take the raw numbers, weight certain values to represent the overall population (as oposed to the sample) and then crunch the numbers. If, as the case seems to be, the statistician assumes that the voting public is more liberal or conservative (such as say, assuming that women will tend to vote Kerry) than it really is, then the numbers will be skewed. For instance;

As late as 7:33 P.M. on Election Day, Mitofsky and Lenski were apparently telling their clients (NBC, CBS, CNN, AP, etc.) that after "weighting" Kerry was beating Bush by 9 points among women and losing by only 4 among men. By 1:24 P.M. the next day revised results revealed that, in fact, Kerry won women by only 3 points while Bush won men by 11 points.

The statisticians who work the numbers are supposed to make educated assumptions based on other trends. I seriously wonder where these guys got their models. I have a feeling a couple of them might be on the unemployment lines this New Year.

Tip O'the Hat to Ace for this one.