Last week “Andy mentioned”:https://themonkeycage.org/2010/09/a_calibrated_cook_gives_dems_t.html that my colleague Sandy Gordon’s new methodology for predicting House races based on using the past accuracy of forecasters (in this case, Charlie Cook and Stuart Rothenberg) to measure likely outcomes based on their forecasts this year. His method puts “the likelihood of a Republican takeover somewhere between 14-37%”:http://politics.as.nyu.edu/docs/IO/2602/short_forecast.htm.
In contrast, Nate Silver, relying on lots and lots of polling data, currently has the odds of a “Republican take-over of the House at about 67%”:http://fivethirtyeight.blogs.nytimes.com/.
Since I have the advantage of having Sandy down the hall from me, I went and asked him about his methodology, and in particular about Andy’s point about incorporating national trends into the seat by seat estimates. In the course of our conversation, he raised two potentially interesting points. First, he noted that – in contrast in 1994 – all the prognosticators now *know* that we are in a wave election. Thus their predictions ought to reflect national trends. Thus a “likely Republican” rating today _ought to already incorporate that national trend_. Thus to say “every likely Republican take over seat should go Republican in this kind of a year” isn’t actually bringing in any useful new information – it is just exposing crappy coding. If every likely Republican take-over should go Republican, those seats should be coded as definite Republican, not likely Republican!
Second, Sandy notes that 2010 elections bear the hallmarks of a bubble – there are lots of people saying the Republicans are going to take over Congress in part because lots of people are saying Republicans are going to take over Congress. We know that the wisdom of crowds can be deadly even when people have lots of their own money at stake (read: sub-prime mortgages); we shouldn’t necessarily expect it to be any better just because people’s reputations are on the line, should we?
Based on this logic, I’d also like to suggest that at this point, there are probably fewer potential Republican “surprises” out there than Democratic surprises, given that everyone has probably already scoured just about every possible Republican take-over and written about it somewhere. If a Republican has a prayer of winning a seat this year, I’m sure someone has called it a “potential pick up” or “how you’ll know the night is going badly for the Democrats”. Thus ironically, we’re probably going to hear about more Democratic “surprises” on election night. “What? Russ Feingold wins in Wisconsin! That’s crazy – must mean the Republican tidal wave has crested and the Democrats are on the comeback trail!” So I’d say there is a somewhat decent chance that the narrative on Nov. 2nd, a night when the Republicans are picking up large numbers of seats in both the House and the Senate, is going to be about the “wasted Republican opportunity”. Who says the Democrats aren’t being strategic this campaign season!