I’m up today at U.S News (see here) trying to make sense of the variation in the more quantitatively-driven Senate forecasts that are popping up. At first glance they seem all over the map, running the gamut from the Princeton Election Consortium’s Sam Wang’s optimistic 75% probability that Democrats will retain a majority on November 4th to the more pessimistic mid-30% probabilities projected by the New York Times and FiveThirtyEight. Here are the projections at six of the sites as of last Friday – the percentage is their assessment of the probability that Democrats will retain control of the Senate:
Five Thirty Eight 36%
Huffington Post 55%
New York Times 38%
Princeton Election Consortium 75%
Washington Post 47%
As I argue in my US News post, however, there is an underlying pattern to this seemingly disparate set of projections. The projections that are purely poll driven (the bold-faced forecasters above) are much more optimistic regarding the Democrats’ chances of retaining the Senate than are those forecasts that include some structural component into their model, such as Obama’s approval ratings, or a measure of economic performance or a generic party preference. It’s too early to say which type of forecast model will prove more accurate – my guess is that they will converge as we get closer to election day. See my US News post for a more detailed discussion.
As I noted to those of you who are on my email distribution list, with classes starting today, the frequency of my posting is likely to go down a bit. Moreover, I’ll be depending a bit more on you to flag interesting items about which you will want me to post. So keep those suggestions coming!