The Book!

Friday, November 05, 2004

Democrats Robbed Again

The emerging party line among Democrats now seems to be that they were robbed again. The Diebold Corporation stole the election from them, be juking the vote on their voting machines in six (out of 88!) counties in Ohio.

Never mind that Kerry lost the popular vote big time. That only matters when it’s the other side. If a Democrat can win electorally while losing the popular vote, that’s just simple justice.

The only real evidence of fraud seems to be that the exit polls are at variance with the actual vote. And that the head of Diebold apparently supported Bush.

This is wildly improbable in its face. There are basically only four possibilities here, with regard to the variance between the exit polls and the actual vote:

1. the vote count was wrong, and there was wrongdoing involved.
2. the vote count was wrong, and it was honest error (snafu)
3. the exit polls were wrong, and there was wrongdoing involved.
4. the exit polls were wrong, and it was honest error.

Now, based on Occam’s Razor alone, 4 is the most probable answer, and 1, the Democratic position, is the least. Proposition 1 requires a vast conspiracy, proposition 4 requires only the usual level of random human imperfection.

But we have further evidence, making this Democratic claim, in a word, incredible.

First, the exit poll figures (or the vote) were off across the board by six percentage points or so; they were not off only in places where there were electronic voting machines, or any one type of voting. They were off in Democrat-controlled counties, and they were off in Republican-controlled counties. This means the accused voting machines cannot possibly be responsible for the discrepancy. Nor can any conceivable conspiracy have been big and diverse enough to have cooked it on the voting side.

Nor could any such conspiracy have confidently predicted it would finally come down to Ohio, and to six of its counties, and so have managed to plant their machines there, years ago.

It would have been theoretically possible, on the other hand, for a small group to cook the exit polls: they were all done by one firm. Making this a far more plausible solution, if there is a problem.

Now recall that the exit polls are also at variance with the regular polling in the days close to the election. These other polls, by contrast, predict the voting outcome rather accurately. So, you have, let's say, six polls plus the actual vote. All but one poll say one thing. One poll calls it differently. It’s a pretty slim premise that the one poll that got it wrong must be right, and all the others plus the actual vote wrong.

Second, we _know_ the exit polls are off. We know this because their demographics are skewed. They include far too many women to be representative of the group that actually voted. Therefore, the exit polls being off is zero evidence that there was anything wrong with the actual vote. This by itself reduces the real possibilities to propositions 3 and 4 above. The exit polls were wrong, and the only question is whether they were skewed deliberately, or in honest error.

Honest error seems the most likely answer. Especially since, according to Slate, "Today's exit polls were no more off the mark than were those of four, eight, or 12 years ago." - Slate. They were significantly off in 2000 as well—so far off that the job was given to a new firm this time.

This argues strongly to me for proposition four. However, we should also remember that:

Third, we actually know there was wrongdoing involved in the exit polling. This came, at a minimum, in their being publicly leaked. This violated the terms under which the exit pollers were employed, although they themselves were not necessarily the culprit. The leak prima facie aided Kerry, not Bush.

No comments: