As a supplement (or possible corrective) to my last post, this article by David Glenn in the Chronicle of Higher Education notes that six out of seven quantitative models presented at the APSA yesterday predict that Bush will win the popular vote. Note that these models aren’t trying to do the same thing as Erik and Paul’s paper – they’re predictive (sort of) rather than analytic. Note also that modelling has had a mixed success in predicting outcomes in the past. Still, all caveats aside, it’s an interesting datum.
Updated thanks to comments.