Election forecasting is a fascinating cottage industry in political science. Our own Michael Alvarez once demonstrated how you could predict presidential election outcomes using the results of the previous year’s baseball World Series.

The paper citation is: Can Bush Hit a Home Run?” With Brian Loynd. The Political Methodologist, Spring-
Summer 1994, vol. 5, no. 2, 2-4, and you can actually read it here. Ain’t the Internet wonderful?

But forecasting, which started out as something of a model fitting exercise for data geeks (like Mike) has turned into a serious part of the profession. There are whole volumes dedicated to election forecasting, and many authors argue that our ability to forecast an election provides us insights into the fundamentals that underlie American elections and electoral behavior.

So what about this year? The models–all produced six months ago–are looking pretty good. You can read all of the results at the American Political Science Association webpage: October 17, 2006: Political Scientists’ Models Predict Democratic Takeover of House of Representatives. Most predict a 22-29 seat gain for Democrats in the House, and 2-3 seats in the Senate.