The Exit Polls That Refuse to Die

Cannonfire summarizes a thread on Democratic Underground that is worth pondering:

Here's just a bit of it:

The exit pollsters, as we all should know, did not merely ask folks whether they pulled the lever for Bush or Kerry in 2004. The polls included a number of other questions. Specifically, respondents were asked to divulge the recipient of their vote in the year 2000.

In a preliminary exit poll released on CNN at 12:22 a.m., the results for the query about 2000 were 41% Bush, 38% Gore.

Two hours later, the final exit poll was released. At that time, the respondents said that they had voted in 2000 in a ratio of 43% Bush, 37% Gore.

Have you found the oddity yet? Feel free to re-read the last two paragraphs. And when you do, ponder this little factoid:

Al Gore WON the popular vote in 2000!!!

How, prithee, can the NEP and the SSRC (not to mention Dick Morris and innumerable other GOP propagandists) ask us to believe that the exit polls were skewed in favor of John Kerry? If such weighting existed, then the question about the 2000 race would have resulted in a demonstrable preference for Al Gore.

If you scroll further down in the DU forum, you will note that one reader suggested that 2004 respondents may have lied about who they voted for in the year 2000. But this "explanation" explains exactly nothing. If, as alleged, the exit polls were weighted in favor of the Democrats, why would Kerry supporters make false claims about having voted for George W. Bush in the previous cycle? Logic and experience tell us that people are usually reticent to mention that they once voted for a candidate who has since left them feeling disenchanted and ill-used.

But once we allow ourselves to consider the possibility that the exit polling was actually weighted in favor of Bush -- well. Much is explained.

Obviously, any party attempting to rig the election would also have to think seriously about ways to shade the exit polls.