According to Professor Freeman, the discrepancies between the exit polls and the vote tally are an anomaly even if you examine just the key battleground states of Pennsylvania, Ohio and Florida. The likelihood that in Ohio Kerry's early exit polls would differ from the final vote tallies was less than one in 100; in Pennsylvania, it was slightly more than one in 100; and in Florida, less than two in 100.

According to Dr. Freeman, the odds against any two of these three combinations of votes occurring together are over 5,000 to one, and the odds against all three occurring together are 662,000 to one.

He concludes that "as much as we can say in social science that something is impossible, it is impossible that the discrepancies between predicted and actual vote counts in the three critical battleground states of the 2004 election could have been due to chance or random error."

Exit polls are not just opinion polls. They are polls of people who actually showed up to vote, taken just after the voting, and weighted to take into account any preponderance of one group.

Professor Freeman's paper points out that exit polls are used to check and verify the validity of elections in several countries, including Germany and Mexico. When exit polls contradicted the claim that Edward Shevardnadze had won election in the former Soviet state of Georgia, he was forced to resign under pressure from the U.S. Recently in Ukraine, widespread allegations of vote rigging were supported by discrepant exit polls, partly funded by the U.S. government, and a newly "elected" government was toppled.

Freeman's initial report was reinforced by the release on March 31, 2005, of a more detailed study of the exit polls by the group US Count Votes.

The authors who provided the research and other evidence for this report included faculty experts in statistics and mathematics from a number of universities.

In their 27-page report, the professors noted numerous problems with the 2004 national election: "Voting machine shortages; ballots counted and recounted in secret; lost, discarded, and improperly rejected registration forms and absentee ballots; touch-screen machines that registered Bush when voters pressed Kerry; precincts in which there were more votes recorded than registered voters; precincts in which the reported participation rate was less than 10 percent; high rates of "spoiled" ballots and undervotes, in which no choice for president was recorded; a sworn affidavit by a Florida computer programmer who claims he was "hired to develop a voting program with a 'back door' mechanism that would undetectably alter vote tallies."

"Under such circumstances," the authors pointed out, "we must rely on indirect evidence" such as exit polls, or analysis of election result data "as a check of the overall integrity of the official election results. Without auditability or transparency in our election systems, the role of exit polls as a trigger for further scrutiny is of paramount importance." They also point out that the accuracy of the exit polls when predicting the outcome of Senate races was not questioned.

The US Count Votes report's authors conclude that "if the discrepancies between exit polls and election results cannot be explained by random sampling error; the 'Reluctant Bush Responder' hypothesis is inconsistent with the data; and other exit polling errors are insufficient to explain the large exit polling discrepancies, then the only remaining explanation that 'the official vote count was corrupted' must be seriously considered."

In this country, immediate investigation is most urgent in the four states where the swing from exit poll results to the published vote tally also swung from Kerry to Bush: in Ohio, Florida, New Mexico and Iowa. Precincts and counties in all four of those states reported problems with the voting process and with vote counting. These four states control 59 electoral votes, more than enough to change the election outcome.

OPTICAL SCAMMING—For Florida counties, the biggest difference in the 2004 election was not between "red" and "blue" jurisdictions but between touch-screen and optical scan voting machines. Fifteen Florida counties used touch-screen voting machines, produced by the company Electronic Systems and Software (ES&S) or by another company, Sequoia Voting Systems. The other 52 counties used paper ballots. But the paper ballots, traditionally the safest way to vote in a supervised setting, were not counted manually. Instead, the paper ballots were processed by optical scanning equipment manufactured by ES&S, Sequoia, or by a third company, Diebold, Inc.

Optical scanning in voting has been used for years, mostly without the careful checking that supermarket scanners occasionally get. Surveys indicate that supermarket optical scanners get the price wrong about 5 percent of the time.

Following the election, a mathematician and researcher named Kathy Dopp tabulated differences between touch-screen counties and op-scan counties. The differences reveal a simple and clear pattern. In touch-screen counties the county's vote for president almost always went for the candidate of its majority party. In op-scan counties, the county's vote for president mostly went against the candidate of its majority party.