Stolen Election?

The current issue of Harper's has an article by Mark Crispin Miller about irregularities in the 2004 election. I am not convinced that the election was stolen, and neither, for that matter, is Miller. It's entirely possible that the election went off perfectly fairly, and also entirely possible that some votes were stolen but that the outcome of the election would not have changed. I'm also not interested, even if the election does turn out to have been compromised, in seeing Bush step down or something similar happening. But if something goes wrong with a voting system, we need to know about it, so the system can be fixed. That's why it's disturbing is that the vast catalog of irregularities seemed to merit virtually no investigation by the press.

Congressman John Conyers did mount an investigation, which Republicans refused to participate in. (His report is here). Miller's piece is only available to subscribers, but an audio stream of a forum on Ohio is available here.

But is there really any credence to these allegations? The evidence seems to be as follows:

1. We know that in the past people have tried and probably succeeded in altering the outcmes of US presidential elections. (There is good evidence, for example, that JFK did not really win Illinois.)
2. The outcome of the 2004 election was completely at odds with exit polls in many swing states. Exit polls are generally very reliable--international election monitors often use them to detect evidence of vote fraud. We have a large enough sample of exit polls and election results to conduct a statistical analysis of the 2004 election, and the results are not reassuring. The chances that all the swing states where exit polls showed Kerry winning would instead yield victories for Bush are miniscule--by some estimates they are on the order of 16 million to one. Colin Shea at Freezerbox was one of the first writers to discuss this, here and here". Statistician Steven Freeman also discussed it, here. Reporter Russ Baker critiqued Freeman (and other allegations of fraud)here and Freeman responded here. (In that exchange I think Freeman got the best of Baker; Baker, who plainly knows nothing about statistics, relies on an anoynmous source who tells him Freeman is "all wrong." Why would anyone need an anonymous source to critique a statistics exercise?!? The rest of Baker's critique is worth reading, though.)
3. We also know that some other weird things happened in Ohio. For example, counties that voted Democratic for every other office--from sheriff to alderman to what-have-you chose Bush or third party candidates for President. Socialist-turned-Neocon Christopher Hitchens, of all people, summarized this evidence in Vanity Fair.
4. Lastly, enough strange things happened in the 2000 election to make people suspicious.

Does all this add up to fraud? I don't know, and I'm not really in a position to judge. My curiousity is related more to why the regular media did not actively pursue some of these questions. While some of fraud allegations are pretty far out there, not all of them are. Concerns about election fraud are not prima facie evidence of wingnuttery--it's not a JFK or 9/11 conspiracy theory, after all. Election fraud is a real crime that happens in many places. Of course, it is also hard to pull off. That's the big argument against such allegations: how'd they do it? And can it be explained just as easily by various mechanical failures/incompetence?

To which the rejoinder is that if these discrepancies CAN be explained by incompetence, we need to know about that, too. And so an investigation is probably still in order. For me, the main thing that sticks in my craw is the exit poll discrepancy. If anyone knows of good debunkings of the Freeman-type analyses I'd love to see them.

UPDATE: Aha! I found one - pollster Mark Blumenthal has a nice analysis of Freeman's work that closes a lot of doors. Read it here.

UPDATE 2: Apparently I'm very late to this game (what I get for relying on Harper's, I suppose.) There's an entire Wiki on the 2004 election disparitie here.

PREVIOUS | INDEX | NEXT