The tricky thing about measuring a polling bounce, which can occur after a candidate performs well in Iowa or another early-voting state, is that it can begin to deflate just as soon as you detect it.
In 2008, Barack Obama surged in the New Hampshire polls after his win in the Iowa caucuses, and polls conducted on the weekend before the state’s primary showed him beating Hillary Rodham Clinton. But Mr. Obama lost to Mrs. Clinton by three points when the state voted the following Tuesday.
A more dramatic example came in 1980, when George H.W. Bush, then a relatively obscure candidate, beat Ronald Reagan in Iowa. Mr. Bush, who had once trailed Mr. Reagan by 30 points in the New Hampshire polls, surged to the lead in some surveys there. But there was more than a month separating Iowa and New Hampshire that year, and by the time New Hampshire voted, Mr. Bush was back where he started, losing to Mr. Reagan by 27 points.
It looks as though Rick Santorum, who initially got a fairly large bounce after his near-win in Iowa, may be the latest example of a candidate with somewhat ephemeral momentum.
In a Suffolk University tracking poll of New Hampshire, Mr. Santorum’s highest standing came in the version of the poll released on Friday. That poll was based on interviews conducted on Wednesday and Thursday — the two days just after the Iowa caucuses — and had Mr. Santorum at 11 percent.
However, in the newest edition of the poll, released Sunday morning and based on interviews conducted on Friday and Saturday, Mr. Santorum’s support is back down to 8 percent.
An American Research Group poll, also based on Friday and Saturday interviews, had Mr. Santorum with slightly more of the vote, at 12 percent. However, Mr. Santorum was in a clear fourth place in the survey, whereas most polls conducted immediately after Iowa had Mr. Santorum in third.
FiveThirtyEight’s latest forecast in New Hampshire shows Jon M. Huntsman Jr., who had somewhat favorable results in the Suffolk and American Research Group polls, moving ahead of Mr. Santorum and into third place. And the dynamics that helped Mr. Santorum in Iowa — a strong retail operation and a character that is a cultural fit for the state — are just the opposite in New Hampshire, both favoring Mr. Huntsman.
Meanwhile, the latest poll in South Carolina, from Public Policy Polling, has Mr. Santorum with 19 percent of the vote there, and in third place behind Mitt Romney and Newt Gingrich. That’s much better than Mr. Santorum was polling before Iowa, but slightly weaker than he was polling in surveys conducted on Wednesday and Thursday, which had him with 22 percent support on average.
This might seem to be making a big deal of relatively small shifts in polls. However, 538’s experience in looking at thousands of primary polls from past years when building forecast models suggests that the greater error is usually in detecting a momentum change one day too late, rather than anticipating it one day too soon. Voter preferences change extremely quickly in primaries.
The larger point is that because Mr. Santorum was so far behind in New Hampshire and South Carolina, he needs continued, substantially positive momentum to perform well in those states and become a serious threat to Mitt Romney for the Republican nomination.
There is abundant talk in news accounts about Mr. Santorum’s being a contender for second place in New Hampshire, something that is certainly not impossible but that now appears less likely than it did 48 hours ago. Because Mr. Santorum was the big story out of Iowa, the news media might be surprised by a finish like the one that FiveThirtyEight forecasts now project in New Hampshire, even though Mr. Santorum has never been a good fit for the state.
Since the volume and the tone of media coverage is as much a function of the media’s expectations as objective measures of performance, that could contribute to somewhat negative coverage for Mr. Santorum in South Carolina when he instead needs to make up ground there.