After listening to portions of the House debate and sampling some of the discussion here, it strikes me that the statistical evidence against the Ohio election is getting short shrift. Could it be that innumeracy is a major factor in our failure to address this matter?
Innumeracy is the mathematical analog of illiteracy: a failure to learn enough of the subject to be basically competent in it. Innumeracy is widespread, and when it comes to statistics, very few people indeed appreciate the issues.
We all know that there are "lies, damned lies, and statistics" and we all share a healthy skepticism about statistics. It's too easy for a charlatan to twist the numbers around to prove anything he wants. But we must remember that such "proof" works only on the innumerate; a competent statistician can slice through most of the claptrap in seconds.
Although statistics can be and often has been used to obscure the truth, statistics can also shed light on the truth, when it is properly applied. Which brings me to the many statistical analyses that I have seen, both of the Florida results and the Ohio results. I look over all those analyses and I come to the conclusion that they provide proof of anomalous results. Some of the studies have been flawed, and critics have pointed out the flaws, but the great bulk of the studies clearly show that something is very fishy about both elections. To the best of my knowledge, the rebuttals to these studies are few, and focus on minor details. I have not seen any broad rebuttal to the studies that impress me the most.
These studies do not prove fraud -- that requires some information on the motivations of people, which statistics cannot illuminate in this case. But they do prove that these elections do not conform to conventional expectations of electoral probity.
Yet these statistical analyses have gotten little attention, both here and in Congress. It seems to me that most people simply wave aside this data because they're in no position to evaluate it competently. But this is our most powerful data. I do not find all the anecdotal evidence prepared by the Conyers group to be convincing. Sure, there were problems in Ohio, but there were problems everywhere. The anecdotal evidence just doesn't amount to what I consider to be "proof". The statistics do.
We therefore come to the sad state where the nation has all the evidence it needs to conclude that the electoral results in Ohio are unreliable -- yet very few people appreciate that evidence.