WE'VE MOVED! IsraPundit has relocated to Click here to go there now.
News and views on Israel, Zionism and the war on terrorism.

November 20, 2004


The "voter fraud" explanation that many Leftists give in explanation for their loss of the recent U.S. Presidential election has already been pretty thoroughly debunked but Leftists are still not giving up on it. Given their own efforts to rig the vote by enrolling dead people etc., one can understand their conviction that there was foul play. It's another example of their chronic "projection" (seeing their own faults in others). And I am sure that one place where the fraud explanation is almost universally treated as gospel would be the People's Republic of Berkeley. So it is no surprise that a group of Berkeley sociologists have done a statistical analysis (PDF) which they believe offers proof that voting-machine fraud took place in the Florida voting.

The method they adopted is amusing, however. I myself taught sociological statistics at a major Australian university for many years so maybe I can explain simply what they did. They took a large number of things that normally predict the vote and combined them to produce an estimate of what the vote SHOULD have been. They then show that this estimate of the Republican vote was lower than the official Republican vote in precincts where voting machines were used. They then conclude that the official figures were "rigged" because they diverged from the theoretical figures.

To show how ridiculous this procedure is we only have to ask why the USA had an election at all if the result was predictable with such certainty? Surely it would have been easier to leave the whole business of electing the President to the sociologists and their computer models! The plain fact, however, is that there are many things -- most of them not easily meaasurable -- that could have influenced the vote but which were not included in our sociologists' model. And even the data that were used are subject to error.

What one of the omitted factors could have been is suggested by econometrician Newmark's Door. He points out that there is a large Jewish population in the "suspect" counties and that Jews who quietly switched to Bush because of his support for Israel and his strong opposition to Islamic terrorism could account for at least part of the "wrong" voting. And it may be noted that the percentage of Jewish voting for Bush is one of those variables that could be particularly hard to estimate. Given the strong historic tradition of Democrat voting in the Jewish community, Jews who did switch to Bush might not be keen to go public about it.