letter

From the New York Times Book Review.

To the Editor:

It was heartening to see Michael Kinsley give America's flawed election system the attention it deserves ("Election Day," Nov. 5). However, when discussing "Was the 2004 Presidential Election Stolen? Exit Polls, Election Fraud, and the Official Count," of which I am a co-author, Kinsley writes that "the authors offer no particular reason to believe the random exit polls and disbelieve the actual vote."

Kinsley apparently associates "random" with meaningless, when in fact randomly selected samples are a fundamental principle of statistical inference. The 2004 national exit poll sampled more than 100,000 respondents. The sampling error on such a poll is well under plus or minus 1 percent.

Our book is a systematic investigation of the only two hypotheses that could explain the seven-percentage-point (nine-million-vote) disparity between the exit poll results and the official count. Either Bush voters participated in the polls at a far lower rate than Kerry voters, or the count was corrupted. No data has ever been released that would support the thesis that Bush voters responded to the exit polls at a lower rate. Indeed, the data indicate that where Bush voters predominated, response rates were slightly higher than where Kerry voters predominated.

This disparity between the exit polls and the official count is significantly higher in the election's 11 battleground states, and higher yet in the critical states of Pennsylvania, Florida and Ohio. And, in contrast to the seven-percentage-point disparity in precincts where the votes were counted by machines, no disparity is found in those few remaining precincts where votes were cast on paper and counted manually.

Steven F. Freeman
Philadelphia

The letter writer is the co-author of Was the 2004 Presidential Election Stolen?: Exit Polls, Election Fraud, and the Official Count.

The answer, by the way, is yes.