Just before I unblogged myself for a week (if you’re curious, we installed our dear daughter at college; be still my heart) I summarized the Senate race rankings of the usual Washington-based suspects, which continue to show that Republicans will make big gains but have a hard time winning control of the Senate in November.
On the House side, Repubs chances of gaining a majority are generally rated much better (in part because all 435 are up). Specifically, three out of five political scientists who try to forecast such things predict that Nancy Pelosi’s days as speaker are numbered.
The American Political Science Association convened in Washington last week and, as usual in election years, put on a panel featuring political scientists who have developed statistical models that try to predict House elections – not on a race by race basis as Charlie Cook and Stu Rothenberg do, but on a statistical indicators designed to pick up the general direction of the electorate.
Mark Blumenthal covered the panel for the Huffington Post. Five individual or teams of professors revealed the results of their models, which rely on factors including economic data and national poll results . All of the models show the Repubs making very significant gains in November. Three predict that the gains will be enough to give Repubs control of the House and the other two say no, Dems will lose ground by retain control of the House next year.
- Alan Abramowitz of Emory University uses a model that relies mostly on the “generic ballot” poll question. (The ones that ask a national sample of voters “if the election was held today would you vote for a Democrat or a Republican). Current generic ballot polls show Repubs with roughly a five-percentage-point lead. Abromowitz’ model says this will translate into a 49-seat pickup. Repubs need 39 to take control.
- Dartmouth College professor Joe Bafumi forecast a 50-seat Republican gain, which would lead to a new House comprised of 229 Repubs and 206 Dems. Bafumi works with two other professors and will continue re-running their analysis until two weeks before the election. Based on the possibility of change between now and election day, Bafumi’s group says their model is about 79 percent reliable. They spell out the model and its past successes here.
- The third forecaster who thinks the Repubs will win a House majority is James Campbell of SUNY Buffalo. His model, which is based on the number of seats currently rated as “in peril” by Charlie Cook, plus Pres. Obama’s recent approval ratings, predicts a gain of 50 to 52 seats for the Repubs.
- Alfred Cuzan of the University of South Florida, whose model, Blumenthal writes, “relies mostly on measures of economic growth and inflation rather than voter preference polling” forecast a Republican gain of 27 to 30 seats, which would leave the Dems still in control.
- The most optimistic outlook for Dems came from Michael Lewis Beck of the University of Iowa, whose model is based on the assumption that the election is a referendum on the economy. Beck and his colleague, Charles Tien, project a Republican net pickup of 22 seats, which would leave the Dems still in fairly solild control.