#3 Here's an NYT interview with Ray Fair (registration probably required) properly identified as being with Yale, not Harvard as in the Sun article. Note that Fair is a Kerry suporter as all the other academics probably are, especially in NY.
These models seem to focus on economic and polling data taken substantially before the election. The assumption is that people vote with their pocket books and know the candidates positions pretty well before the campaigning starts. This seems to be a reasonable asumption.
A truly interesting test of the model, unlikely to ever be made, would be for the leader in the forecast to stop TV advertising as soon as the model projections come out. This could save a lot of money and create a lot of positive publicity and press coverage.
Nonetheless, it is hard to imagine that the election results are impervious to exogenous variables in the interim such as a successful attack by Baathists freeing Saddam from captivity, a Chinese invasion of Taiwan or an Iranian nuclear attack on Israel.
OTOH, it's the economy, stupid! |