Thursday, January 10, 2008

Why Did Hillary Beat Barack in New Hampshire?

Stuart Rothenberg gives an interesting analysis. As far as I can tell, he is the only one to notice that Clinton's gains did not come at the expense of Obama. He argues that Clinton's victory was at the expense of Edwards.

The data appear to bear out portions of this analysis. According to the Real Clear Politics average, Obama was on target to receive around 38% of the vote. had Obama at 37% of the vote. CNN reports that he received 37% of the Democratic vote on New Hampshire primary night--just one percent less than the share predicted by the RCP polls and the same share as expected by

Rothenberg is, of course, right that Edwards performed much more poorly in New Hampshire, a less fertile ground for populism, than Iowa. However, Edwards did not perform much worse than predicted by the pre-primary New Hampshire polls. The RCP and pre-primary averages of 18% for Edwards was just one percent higher than the 17% he received on primary night--the same difference as for Obama.

Clinton received 39% of the vote--a full 9% above her pre-primary RCP average or expected by Where did her votes come from? Two possible sources include undecided voters (roughly 4-5% in the polls) and one other source not much mentioned in the analyses I've seen: Biden. RCP doesn't report data for Biden but's models predicted he would receive 2.5% (and says his last-five polls average was 2.6%)--he received just 0.2% of the primary vote.

Clinton outperformed the pre-primary polls but Obama did not underperform. There isn't a lot of evidence to support claims that white voters were lying to pollsters to hide their preference for a white candidate over a black candidate as suggested in today's New York Times. Quick claims of bias are unhealthy and perhaps overly common in Democratic contests. Let's not rush to the wrong judgment here, especially regarding a black candidate who has clearly been embraced fervently by so many white voters.