The Light That Died

http://www.buzzflash.com/alerts/04/11/The_unexplained_exit_poll_discrepancy_v00k.pdf suggests that the chance of the exit polls in all US states being so far in disagreement with the election results is two hundred and fifty million to one. That's right, you have to run 250,000,000 US federal elections before you see such a result occur once.

When you have multiple counties in Florida that are registered 80+% Democrat, but vote massively for Bush, or when Florida votes 70% for a liberal wage referendum, but not 70% for the Democratic party, not to mention the astonishing exit-poll irregularities - check the histograms on UsaSystemsCatastrophe - the BurdenOfProof is on those who say the election was legit. If they can't prove it, do it over.

Is there evidence that voters registered Democrat in the South never vote for Republican candidates? Recall that the South was until recently the stronghold of conservative Democrats.

Doesn't explain why the pro-Bush swing correlates directly with the kind of voting machine used - see http://residentbush.com/Aftermath-2004_Florida-Results.htm for data.

[Some discussion moved to ModestElectronicVoteProposal]

But to take a bunch of less than stellar pollsters and conclude that this proves something, well, that doesn't seem very reasonable to me. If you ask me, my bets are that they screwed up, and they don't want the incident to be investigated so that they can cover their asses. It can also be a possibility that there's nothing to investigate as they haven't kept good enough records.

The probability can be determined as a matter of simple statistics; hypothesis has nothing to do with it (except the hypothesis that exit polls constitute a properly random sample). To be fair, it should be noted that the probability of the exit polls agreeing with the results exactly is pretty low as well; there are simply so many possibilities (disagree by one vote, by two etc.) However, I find the discrepancies quite alarming, personally, especially considering that such discrepancies have not occurred in the past. Given all the other election irregularities that have been turned up, it is very difficult not to be a ConspiracyTheorist.

If this were an isolated incident in an isolated geographic location, you might have a point. But when it's in multiple counties in the two main battleground states, and a massive swing, and correlated closely with the particular kind of voting machine used, then you need to account for the whole phenomenon, not just the exit poll mismatch. How else can you explain these results than systematic electoral fraud?

Hot news: a "Liberal" candidate came out ahead in the first round of the presidential election - see http://seattletimes.nwsource.com/html/nationworld/2002087786_ukraine11.html for details of the results and associated allegations.

The electoral colleges and first-past-the-post system (and the resulting two-party nature of the elections) are bad enough; having to go out and actively register to vote is pretty retarded; allowing electronic voting machines that do not leave a paper trail to be used in "swing states" is quite disgusting, especially given the allegations of collusion between the machine's manufacturers and one of the parties.

I've been hearing variations on "how can such a prosperous nation be so stupid?" There seems to be a consensus that the USA "doesn't deserve" its freedoms and prosperity because it refuses to conform to the standards of [insert country] or the standards of [international group]. It's possible that as the USA moves closer to compliance with [country] and/or [nation group] things deteriorate in the USA, rather than the other way around.

It's possible that wings grafted to pigs will allow them to fly. Let's try not to be purposefully obtuse here. The point is that significant statistical experience in math has been applied here to show that the difference between the exit polls and the vote totals really are skewed. Trying to redirect the discussion to other topics doesn't change that. Try not to use DisproofByFallacy, please.

CategoryVoting