Thanks Kevin for the link. Your questions regarding the historical democratic bias of exit polls is an interesting one. I will check with the Roper Center to see if these unweighted data are available. The California State University is a subscriber to the Roper Center archives and according to the program manager for the CSU system, I can get the data in about 48 hrs, if they have it on file. I know the 2004 primary election data is there. Don't know about 2002.

I am taking another stats class next semester with Dr. Parker of SDSU who reviewed my work and is quoted in my post. For that class, I have to do a "project." Sounds like a good project to me! I'll look into it, but given that finals are in 3 weeks, I'll have to wait a bit.

This isn't going away.

BTW - your link to my "Exhibit" appears to be broken.

I fixed the Exhibit link.

In 8 of the last 12 presidential elections, the so-called "undecided" vote on the final day has broken to the Republican candidate, whether he be incumbant or challenger. This is perhaps some evidence of a systemic bias against Republican presidential candidates over a long period of time. That is a more likely explanation than undecided voters really breaking for the Republican so often.

- Lets see... a 3% Bush shortfall across the board....(scratching head)....could that be a hidden "poll tax" or could it have simply come out of the tip of some polling managers pencils...Maybe we should revisit some of the old "anti-trust" statutes...But I think if any of this is the case the Dem's got a quick lesson in economics... The market sets the prices, supply, and demand... In the end the American electorate wasn't "buying" Kerry but demanded Bush....The people who payed the highest price were the Dems, and the Red staters got the best of the deal...dare I say it....

- Bargain priced Bush....

A likely explanation for the 3% Bush underpoll is the habitual secrecy of conservatives who live in blue states, where anti-conservative bigots are very willing to use their social and workplace power to undercut their ideological opposites. I live in this environment, and can report a regular phenomenon. When I make pro-Bush, pro-Iraq-war, pro-gun-rights, pro-school-choice or other conservative remarks to people whose politics I am unsure of, but whom I suspect of being reasonable, at least half the time, the response I get is a conspiratorial WHISPER, often along the lines of: "you should hear what I have to listen to all day at work." If the exit pollers asked whether voters were homosexual, or drug users, do you think they would get an underpoll? A lot of conservatives are closeted in the same way.

This explanation is plausible even if the underpoll is consistent across red and blue states. What matters is the county level data, because blue-America really isn't a state level phenomenon but an urban-exurban phenomenon. Thus while Colorado, for instance, is clear majority red, Denver, Boulder, etc. are royal blue, where conservatives are likely to be closeted.

Conservatives are not bigoted the way leftists are, so the phenomenon does not work in reverse, leading to the net underpoll.

Want to know what Kerry said to Geraldo Rivera on why he didn't win? The Bin Laden tape that came out just before it.

Actually I'm surprised he got as many votes as he did but he still had to hold up the pomp and circumstance. Maybe now we can get back to business.

Cindy