The New York Times’ Nate Silver regards President Obama as a heavy favorite to win re-election, prompting conservatives to question his methodology and the NYTimes Paul Krugman to write a histrionic indictment of those conservatives.
“This is really scary,” Krugman wrote of a National Review column daring to challenge Silver’s model. “It means that if these people triumph, science — or any kind of scholarship — will become impossible. Everything must pass a political test; if it isn’t what the right wants to hear, the messenger is subjected to a smear campaign.”
But there’s nothing crazy about suggesting that Silver’s model can err in picking favorites in a horse-race — he did it himself in 2010, in a postmortem of his blog’s treatment of the U.S. Senate race in Nevada.
(A note for Krugman’s benefit: This post is not an attack on Nate Silver. I can’t make the numbers talk good. Silver appears to be a math wizard.)
On October 27, 2010, Silver wrote that Angle had “better than a three-in-four chance to win her race against Harry Reid, according to the FiveThirtyEight model. Ms. Angle’s odds improved today on the strength of a CNN poll showing her 4 points ahead of Mr. Reid.”
Senate Majority Leader Harry Reid, D-Nev., remains in power because he won that election by six points. Silver acknowledged this possibility in his model — which gave Reid a 25 percent chance of winning — but he still offered reasons for why Angle had appeared to be a favorite.
“My guess is that the error instead might have been that the polls, in essence, overestimated the enthusiasm gap,” Silver wrote on election night. “Mr. Reid is a candidate for whom one votes grudgingly — because his opponent is unacceptable to you, or because Nevada makes it easy to vote early on the way home from your shopping trip . . . He’s not someone you’ll necessary be happy about voting for. He’s certainly not someone you’ll be excited to vote for.”
Is it difficult to imagine a similar dynamic working in this election, which features a disappointing incumbent and a Republican challenger who has — almost intentionally — run a campaign devoid of razzle-dazzle? The analogy to Reid’s campaign might bode well for President Obama, but maybe it plays to Romney’s favor as well, as conservatives and some independents might understate to pollsters their willingness to vote for Romney when Obama is his opponent.
The day after the election, Silver fielded a theory that the polls had underestimated the Latino vote in the election — in other words, that they overestimated the Republican percentage of the electorate. In 2012, with some pollsters relying on sample sizes that would require Obama to outperform his 2008 victory, some conservatives (and the Romney campaign) think that polls favoring Obama overestimate the Democratic percentage of the electorate.
As Silver’s Angle-Reid commentary makes clear, his model is only as accurate as the polling data that he considers. If it’s right, he’s got a pretty good read on the election. If it’s wrong, Democrats who know that Silver currently gives Obama an 86.3 percent chance of winning tomorrow are in for a shock.