The Curious Sociology Of Electoral Predictions

by Jonathan Chait | November 2, 2010

Nate Silver notices something about electoral predictions that makes no sense:

it doesn’t make sense to say, for instance — and a lot of people are saying things like this — that Republicans should gain between 50 and 60 seats, and the number could be higher, but it shouldn’t be lower. If that’s what you think, you should project their gains to between 60 and 70 seats (or whatever) instead.

There is a mathematical distribution here.  Whatever your median guess is, the probability of the result ending up above it or beneath it should be equal. There's no reason to project some lopsided curve with a huge right tail.

Yet lots of pundits are saying things like Republicans House gains will be "in the mid-60s or (maybe much) higher." Why are they saying nonsensical things?

I think there's a strange sociology at work here. Obviously, there has been a strong Republican undertow in Washington over the last year or more. The primary effect of this undertow has of course been to improve the GOP's prospects. But the secondary effect has been to make conservatives giddy and liberals depressed. On the right, a can-you-top-this sensibility is at work in predicting the November bounty. And while there are many conservatives making outlandishly hopeful electoral predictions, and very few liberals doing the same. Nate Silver is, properly, the lodestar of liberal electoral analysis, and his prediction is admirably down-the-line. Not even partisans like Markos Moulitsas are challenging his median prediction of a 53-seat House loss.

And so, in a lopsided environment like this, it seems like all the social pressure is pushing pundits toward Republican-friendly positions. That's why you see them making nonsensical guesses where they throw out a reasonable number (like 60-some seats in the House) along with a caveat that the number could be vastly higher. What does this show? It shows they're afraid of guessing low but not afraid of guessing high. Nobody wants to be on the wrong side of this wave. Nobody wants to be the mainstream reporter who is ridiculed by Republicans for not getting it. This sociology dynamic might explain why Charlie Cook and Stu Rothernberg are predicting higher aggregate GOP House gains than their own race-by-race forecasts would suggest.

I am not saying Democrats will exceed expectations. I'm not even saying Republicans won't have a massive 70+ seat wave. I do think, like Silver, that a 70+ seat GOP wave is about as likely as Democrats holding onto the House. There's a lot of uncertainty in both directions. But the curious sociology of Washington at the moment is making everybody focus on one end of the probability tail while ignoring the other.

Source URL: http://www.newrepublic.com//blog/jonathan-chait/78836/the-curious-sociology-electoral-predictions