Election Forecasting is Fun, but….

The models are predicting a rather substantial Bush win. While accurate in many past elections, they blew the 2000 election–why? My argument is that all elections before 2000 were fundamentally different focussing on the economic cleavages and not the dominant cleavage we see now–social issues.

More to the point, the models are essentially fun toys for most political scientists. Many of the better known also do the same thing with the baseball rankings every year. One particular political scientist was not spoken to for several months after he used most of the department’s computer time (back in the mainframe day) to calculate his predictions.

3 thoughts on “Election Forecasting is Fun, but….”
  1. I remember this model from 2000, and I think Arch hit the nail on the head by pointing out that it ignores the growing importance of social/cultural issues.

    I think the model works well in a manufacturing-based society where party loyalty to a large extent is dictated by income. We don’t live in that world anymore — party loyalty now seems to be determined by a variety of non-economic factors. David Brooks in the NYT has been writing some good stuff about this recently.

Leave a Reply

Your email address will not be published. Required fields are marked *