Thursday, November 18, 2004

Some More on the Berkeley Study

I have been cruising the net in search of discussions about this study, and I realized that it's important to point out a few things:

1. This is pretty strong evidence for something being wrong with the electronic voting machines, about as strong as the evidence we have that smoking causes cancer. Those who argue that this is not proof should also then accept that we don't know if smoking causes cancer. The only alternative explanation would be if there is something else about the areas with electronic voting machines that correlates with their introduction, but I have not seen any good suggestions on that. The idea that this increase could be caused by a higher Jewish vote for Bush doesn't work, because there are not enough Jewish voters in those areas to cause the effect that was observed, though it's always possible that some fraction of the effect is due to increased Jewish support of the wingnuts. This is something that could be studied quite easily, to find out if that is the case.

2. A reporter supposedly said that this study doesn't matter because it will not change the Florida results. That is an inane comment for two reasons: First, election transparency matters whether the audits would change results or not. Second, this study cannot be seen as meaning that all other reasons for increases in Bush votes have been explained. Consider what the study does: it studies whether the counties with voting machines differ from those who used punch cards. It does NOT consider whether the vote in the punch card areas is valid or not. All the findings mean is that there is something about the voting machine areas which is odd. It doesn't say that the rest of Florida is perfectly ok, and neither does it say the reverse. That remains an open question.

3. The conclusions the authors of the study made is that there are good grounds for auditing the vote in at least some parts of Florida. They do not say anything about whether the difference they observe is due to fraud or not. Of course, the problem is that the election people in Florida are not exactly neutral auditors of their own problems. There is also the other problem that there may be no way of auditing machines without paper trails...