Thursday, December 09, 2004

Exit Polls in Colorado

I'm religiously reading all the election information on the internet and will post something more about it soon. But this little bit is quite fascinating, and it's beginning to answer an important question about the possible inaccuracy of the exit polls in the 2004 U.S. presidential race: Were other races equally inaccurately predicted in these polls? If there was something really wrong with the polling method, one would expect that the answer to this question would be yes.

The difficulty in studying this question is that the raw exit poll data has not been made available. But one website has obtained some exit poll data on other races. In particular, the state of Colorado has late exit poll information not only on the presidential race but also on the senatorial race.

It turns out that the late exit polls predicted the senatorial race perfectly, whereas they overstated Kerry's vote percentage by almost two percent. Curious.

Of course, the exit poll data on this site may not be valid. We really should be given the late exit poll data by the company that has it. Keeping it secret serves no useful objective.