Bush Stole the Election

Well, maybe not, but humor me for a moment. I have to say I was somewhat disappointed that Kerry conceded so easily. When I vote for a candidate, I expect him or her to press on till there's clearly no hope of victory, and I think Kerry gave up too soon. I'll admit that I'm probably particularly easily annoyed right this second since Donna Frye, a write-in candidate for mayor in San Diego, basically won and then bailed saying that she really didn't want to be mayor anyway. It would have been nice for her to point that out before I wasted my precious vote on her. Oh well.

So what's the evidence that Bush stole the election? Well, I'll have to admit it doesn't look terribly solid, but it is ample enough to make me think that it is a possibility, even if somewhat remote. The evidence consists mainly of a couple of statistical studies, coupled with the fact that electronic voting machines can be fairly easily hacked.

A study by Steven Freeman at the University of Pennsylvania looking at differences in exit polling results and counted votes states 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 [Ohio, Florida, and Pennsylvania] of the 2004 election could have been due to chance or random error."

Certainly, the exit polling methodology could be the problem (as opposed to the actual voting) and admittedly, Freeman concludes with

"Systematic fraud or mistabulation is a premature conclusion ..."

but still, I do wonder why there is such a significant discrepancy between voting and the exit polls. Perhaps we'll find out eventually, but until we do, we can only conclude that there is a discrepancy, not that there is any certainty of fraud.

The second statistical study is from UC Berkeley and concludes that irregularities associated with electronic voting machines may have awarded 130,000 - 260,000 or more excess votes to Bush in Florida. Unfortunately, the operative word is "may" in that there "may" also have been zero excess votes awarded to Bush. Also, there is the minor detail that Bush won Florida by almost 400,000 votes so even if the study is right, and the upper estimate is correct, Bush would have still carried Florida. In addition, there are a number of problems with this study, the most significant being that the authors needed to use a very complicated multivariate interaction model which, as Michael McDonald, a professor at George Mason University concurs, "lacks theoretical grounding", which in laymans terms means "what the heck were they thinking when they decided to use that bizarro model?" Basically, a simple, single, linear independent variable didn't show any discrepancy at all, so they had to try other models with more degrees of freedom to fit the data, whether or not those more complicated models made any sense. So I don't find this study particularly convincing, but the authors do, so maybe there's more to it than I (or Professor McDonald) think. Perhaps the authors will update their paper and explain why they think the model they used was a good choice and why it doesn't overfit the data and why the single independent variable model shows absolutely nothing of interest.

The evidence that the electronic voting systems can be hacked is fairly solid and disconcerting. It's especially disconcerting because there's no paper trail, so it can't be independently audited. Here in San Diego, they have electronic tabulators that count paper ballots that we mark with pens. Ballots that have errors are kicked out immediately and destroyed and the voter is told to go try again. I like our system. It can be counted by hand if need be, yet is completely automated. Voting systems without an audit trail just seem like a bad idea to me at this point in the technological development.

What will make it extremely difficult for the Kerry camp to make any headway, is that there is no hard evidence that anybody actually did commit massive voter fraud. It's not good enough to show that it could be done, it has to be shown that it actually was done, that fraud was actually committed by someone, and that the fraud was as massive as required to swing the election to Kerry. Even statistics with more confidence and fewer variables with the only possible explanation being vote fraud might not be enough, someone has to be caught who committed material fraud.

Because we can't just recount the votes. The electronic machines will give the same results. And we can't just say that there seemed to be fraud, let's give 500,000 votes to Kerry to distribute how he sees fit. There would actually have to be a re-vote. It would have to be in every state, not just ones that Kerry lost. That means Kerry might pick up Ohio but lose New Jersey instead. Or he might squeak out a win. Then what? Best two out of three? Best three out of five? You can sure the Republicans will find an anomaly or two in the second election and get a re-re-vote and perhaps a re-re-re-vote and a re-re-re-re-vote if necessary.

No, as much as I would like to see Kerry as President come January, I think it best to let this election go at this point (so maybe I shouldn't blame Kerry for conceding - but that was then, this is now). Then we should concentrate on making the voting systems better. Unfortunately, that's left to each State and I'm happy with mine so I'll have to count on Ohioians and Floridians to step up to the plate and fix their own.