Exit polls as audits?

I might've given up on the exit poll discrepancies as a sign of something bigger going on here, but Keith Olbermann hasn't. He's pointing to two extremely contradicting studies coming out of two respected universities -- and no, not Bob Jones University.

On the one hand, there's the voting project from CalTech and MIT, which says that while the incorrectness of national exit polling can't be explained by a "margin of error," "on a state-by-state basis, it actually was within that margin." That was my reaction, too, when looking at the exit polls from later in the day, where the discrepancies were rarely larger than 3 percent.

But then Keith sites a study from a professor at the University of Pennsylvania that exit polls are "used as 'audits' on the elections in places like Germany and Mexico." This report "suggests the actual statistical odds that the exit polling was that wrong in the battleground states were 250,000,000 to one."

Now those are some odds.

The study from Penn is by Steven F. Freeman, a "visiting scholar from the Center for Organizational Dynamics" and can be downloaded here. Freeman's report says he's "on the faculty of the University of Pennsylvania: his areas of expertise include resilience, innovation, and research methods. He obtained his Ph.D. from MIT."

In his report, Freeman says he has "tried to jusstify the discrepancy as a legitimate issue that warrants public attention. The unexplained discrepancy leaves us with two broad categories of explanations: the polls were flawed or the count is off.The most important investigations concern verification of the tallies and allegations of fraud on one side; and examination of the exit poll's methodology and findings on the other. Some useful statistical analyses would compare the 'shift' in battleground states vs. non-battleground states, and in states, counties and precincts where safeguards ar estrong vs. those where they are suspect. Obviously, if the polling consortium would release their data, that would allow us to do more definitive analyses." Freeman makes the right point that "so many people suspect misplay undermines not only the legitimacy of the President, but faith in the foundations of the democracy."

But the Caltech-MIT Voting Technology Project says in their report that in examining the exit poll discrepancies -- and only the exit poll discrepancies -- that there "is no evidence based on exit poll results to conclude that there was fraud in the 2004 Presidential election." In addition, Caltech-MIT also has their own report explaining the Dixiecrat effect in Florida. So Caltech and MIT are disputing the already debunked "Florida Democrats voting for Bush" theories and the exit poll discrepancies -- but no other debunking is going on here.

No debunking of Ohio. Or Broward. Or New Hampshire. Or anywhere else.

I'm not sure if Freeman and Caltech-MIT were using the same exit poll results in their comparisons. Remember that the exit polls from earlier in the day showed much different results than those from the late afternoon.

We'll see where it goes from here. I don't know about you, but I'm setting my Tivo so that Keith Olbermann's Countdown is on season pass...