Perry pierde apoyos pero Romney no los gana

Nate Silver analiza los últimos sondeos nacionales:

Mr. Perry led in the CNN poll, the first one released after the debate, but his numbers slumped in the Fox News and YouGov polls. He averages about 22 percent of the vote across the three surveys, down materially from 28 percent in polls conducted between his first debate on Sept. 7, and Sept. 19.

However, any gains for Mr. Romney — in an absolute rather than relative sense — are hard to perceive. He averages 22 percent in the surveys, up only nominally from 21 percent before.

Instead, the candidate making the largest gains is Herman Cain, the Georgia entrepreneur who won a straw poll in Florida over the weekend. He’s up to 13 percent in the surveys, more than doubling the 6 percent that he had before.

The other candidate to make gains is Newt Gingrich — up to 10 percent from 7 percent.

The numbers for Ron Paul and Michele Bachmann have slipped somewhat, however. Mrs. Bachmann in particular, who was in the double-digits in many polls at her peak, is now down to about 5 percent of the vote. And a separate poll of Iowa voters shows her having fallen behind Mr. Romney there. (Although I wouldn’t put as much emphasis on that one because the pollster in question, American Research Group, has a dubious track record.)

Certainly, one way that Mr. Romney could win is if the support among the more conservative portion of the Republican electorate were divided among several different candidates. With the decline in Mr. Perry’s numbers, the landscape has become more favorable for him.

Mr. Romney’s odds of winning the nomination would improve, moreover, if his main opposition came from a candidate like Mr. Cain rather than Mr. Perry, since there is a much larger gap between the two in areas like fund-raising and establishment support. (Although I would not discount Mr. Cain’s chances entirely.)

At the same time, you have a candidate in Mr. Romney who has run a very good campaign, who has performed well in the debates, and who leads in fund-raising and endorsements — but who is still barely above 20 percent in surveys, and has made only marginal gains as a number of his rivals have stumbled.

Mr. Romney has emerged — or re-emerged — as the favorite; I’d give him roughly even odds of winning the nomination. But it’s unlikely to be a smooth and linear path, and the alternate hypothesis that Republican voters are determined to pick someone more conservative than him has some support in this data.

That’s not to paper over the problems of Mr. Perry, who entered the race in a strong strategic position and has failed to make much of it. It’s possible, moreover, that the fallout of the Sept. 22 debate is not yet fully realized in the surveys; Mr. Perry performed somewhat worse in the Fox News and YouGov polls than in the CNN poll, which postdated it by a couple of days.

In general, however, I’d caution against using terms like “momentum” when discussing the nomination race (or polling results under most other circumstances). We’ll be publishing a separate article on this shortly, but there’s not much evidence of serial correlation in polling data: candidates who decline from one period to the next are just as likely to rebound as to see their numbers continue falling.

Note, for instance, that the candidates to have made the largest gains recently, Mr. Cain and Mr. Gingrich, had numbers that were going in the other direction before that. And at this point four years ago, there were plenty of negative stories circulating about the polling trajectory of Barack Obama and John McCain.