Modelling Congressional Elections

Introducing my ‘Congressional Psuedo-swingometer’

As a PhD student who has to complete his thesis by June I haven’t got the time to handicap hundreds of Congressional races, or to produce a complex swingometer. However, I’ve produced a simple model to explain how national votes translate into Congressional Seats.

GOP Seats = 195.446 + 4.225x

Where x = GOP% of national vote – DEM% of national (not the two party) vote.

Although this model does not take gerrymandering, uncontested seats or incumbency into account it does a pretty good job of modelling how changes in the national vote translates into Congressional seats. For the statisticians among you the model has a correlation of 0.902 and an r2 of 0.81. It correctly predicts control of the House 21 out of 24 times since 1960 with the worst errors being 2006 (GOP getting 28 more seats than expected) and 1976 (GOP getting 18 less seats than expected). Given that portions of the most controversial boundary changes of the last eight years have been unwound and large numbers of Republicans are retiring, a scenario with the Republicans doing better in terms of the national vote than they did in 2006 but losing a significant number of seats is not impossible.