Give it up. No one can predict elections with accuracy and, most important, with consistency. Howard Kurtz surveys the errors. Not even the real experts can do it. I recall an article in The New York Times Magazine in 2000 in which a scholar from the rational choice school made the prediction that Al Gore would win the presidency–and not just win it, slam dunk it. On the cover was a graphic of his elaborate formula (rational choicers love those formulas because they posit the possibility that political science can/should be a “real” science). Some of their models are pretty good. In fact, I’d say the rational choice method comes as close to real prognostication as we’ll ever get. And they were wrong about Gore.
UPDATE (11:00 a.m.): Brill’s Content used to keep a score card of Sunday pundit predictions. And many commentators have wondered over the years why it is that none of these talking heads pay a price for error. Why do they still have jobs? Why does anyone continue to listen to them? These questions only make sense if one assumes that prediction is a legitimate journalistic function. Political prognostication is a sport, an entertainment. No one pays a price because such antics have so little political value. It’s simply fun to watch.