I agree that it's ridiculous to try to claim that a discrepancy between the early leaked exit polls and the final vote tally is proof of fraud. However, the final exit polls are adjusted to reflect the final vote count, so of course they are going to show Bush gaining enough strength in some subgroups to account for the official margin of victory. Had Kerry won 51-48, instead of the other way around, the subgroup numbers would have reflected that. Long-time Democratic pollster Mark Blumenthal gives a good description of exit poll methodology for those who don't understand what's going on here. Scroll down to the heading EXIT POLLS: WHAT YOU SHOULD KNOW.
http://www.mysterypollster.com/main /
The problem that remains is that apparently the final predictive poll numbers, which are released around the time the polls close and are waited according to turnout, apparently also showed Kerry ahead by something like 51-48.
Officials with some of the newspapers that subscribed to the service said the ultimately misleading polling data forced them to scramble late at night to change some articles. The presumption of a Kerry victory built a head of steam late in the day, when the national survey showed the senator with a statistically significant lead, one falling outside the survey's margin of error.
"The last wave of national exit polls we received, along with many other subscribers, showed Kerry winning the popular vote by 51 percent to 48 percent, if true, surely enough to carry the Electoral College,'' Steve Coll, managing editor of The Washington Post, wrote in an online chat with readers Wednesday.http://www.nytimes.com/2004/11/05/politics/campaign/05p... If that is the case, then either there would appear to be a serious problem with the exit poll methodology or a serious problem with the vote counting, or both. Unfortunately, as far as I know, those numbers are not publicly available anywhere, although they were presumably disseminated to numerous news agencies. Mitofsky, the agency which does the polling, tries to explain the discrepancy as a problem with the exit polls themselves, but it looks to me like they are just presuming there was a fair count of the votes and trying to make excuses. It looks to me that the explanation is at least as likely to be massive fraud.
Edited for trivial typing error, and another