We attended both political conventions (the Democrats in Boston and the Republicans in New York), and went away with the impression that the GOP was in a stronger position. The Dems strained to hide their true beliefs; most speakers in the convention hall relentlessly repeated the same strongerathomerespectedintheworld boilerplate. Even Al Gore, whose speech we attended in hope of seeing an angry outburst, was weirdly restrained. The Republicans seemed genuinely confident, while the Dems seemed to be trying to convince themselves they were going to win.
The Democrats made one big and erroneous assumption. We quoted Democratic weathervane Josh Marshall making the claim on June 30, 2004:
I take it as a given that virtually no Gore voters from 2000 will pull the lever for Bush. But how many lightly-committed Bush voters from 2000 will hold him to account if they believe he gambled big and gambled unwisely with America's honor and safety, and came up short? I think more than a few. And since there were more Gore voters than Bush voters last time anyway, well . . .
Yet we heard from many readers who said they had voted for Gore in 2000 and were switching to Bush this time around; we published a collection of their responses on July 1 and July 2. We also asked for Bush defectors, and published their contributions on July 7; few were actually enthusiastic about pulling the lever for Kerry. This was a self-selected group, but it was enough to suggest that paritisan Democrats were deluding themselves into thinking that their antipathy for Bush was a universal sentiment outside Republican partisan circles.
The best predictor of the outcome turned out to be not partisans on either side, but the bettors at Tradesports.com, an online gambling outfit that takes bets on political events. On Oct. 21, 2004, we produced this map based on the Tradesports futures prices:
This correctly predicted the outcome in 46 out of 47 states; New Mexico, which Bush carried, was the only exception. "If the other 47 states all follow Tradesports' expectations," we wrote, "Bush wins if he carries either Ohio or both Iowa and Wisconsin. Kerry needs Ohio along with either Iowa or Wisconsin." In the end, it came down to Ohio, since Iowa went for Bush and Wisconsin for Kerry, both by very narrow margins.
Tradesports turned out to be vulnerable to bad information; as this site shows, an election-evening Bush panic led to final results, as of 6:30 p.m. Election Day, wrongly favoring Kerry to carry Florida, Iowa, New Mexico and Ohio and thus win with 311 electoral votes. Betting patterns 12 days ahead of the election turned out to be a much better predictor than the early exit polls. |
