The shift away from Newt Gingrich in the Florida polls has been as significant as the one toward him in South Carolina. The first two polls released after South Carolina showed Mr. Gingrich with a lead of 8 and 9 points, respectively. But the news has gotten worse for him almost every day, and he trails Mitt Romney by nearly 15 points in the final FiveThirtyEight forecast of the Florida primary.
There is a relatively large amount of disagreement in the Florida polls — more than in the first three early-voting states. Some polls seem to imply a modest rebound toward Mr. Gingrich over the final 72 hours of the campaign, while others show continued momentum for Mr. Romney.
The disagreement may reflect, among other factors, the different ways in which pollsters handle early voting — perhaps one-third of votes in the Republican primary have already been cast by way of early or absentee ballots. And pollsters have different techniques for measuring the intentions of Hispanic voters, an increasing problem in survey research in Florida and other states.
Mr. Gingrich, however, has not held the lead in any Florida poll in roughly a week, and his odds of pulling off an upset are extremely slim. The FiveThirtyEight forecast model gives him just a 3 percent chance of victory.
Although polls of primary elections are subject to much larger errors and anomalies than those of general elections, Mr. Gingrich would need to overcome roughly twice the polling deficit that Hillary Clinton did in her upset win over Barack Obama in the 2008 New Hampshire primary. And he would need to do so despite Mr. Romney having accumulated what is almost certainly a substantial head start in early voting.
Other than a more-or-less spontaneous revolt by Florida voters against the expectation that they will elect Mr. Romney, it is hard to see how that 3 percent scenario materializes.
Still, if a win by Mr. Gingrich is exceedingly unlikely, there is some uncertainty about Mr. Romney’s potential margin of victory. Given that the shift toward Mr. Romney in the polls has probably been priced into the news media’s expectations, the margin may make some difference in the way the outcome is interpreted.
As we have done for past early-voting states, we will consider the implications for each of the remaining candidates in light of their range of possible outcomes. These ranges are meant to represent the 90 percent confidence interval on the forecasts, meaning that there is some chance that the candidates’ actual vote totals will fall outside the range. So far, however, the forecasts made by FiveThirtyEight and other polling aggregators have been fairly accurate, which suggests that the forecasts may be underconfident and may err on the side of overestimating the true degree of uncertainty.
Mitt Romney
FiveThirtyEight forecast (most likely outcome): 44 percent
High end of forecast range: 51 percent
Low end of forecast range: 33 percent
There are three relatively substantive thresholds that Mr. Romney could surpass that could make his win in Florida seem more impressive.
First, if Mr. Romney appears on track to win by 10 or more points based on initial exit polls and the first few precincts that report results, some news networks will probably “call” Florida for him almost immediately after polls close at 8 p.m. Other news organizations, including The New York Times, take a more cautious view toward declaring winners in elections. Nevertheless, such an early call would ensure that Mr. Romney gets a favorable night of news coverage.
Meanwhile, it is likely that some precincts will report early-voting and absentee ballot results before disclosing their Election Day returns. Since these ballots are thought to favor Mr. Romney, that could create the perception of a commanding victory even if his margin erodes some over the course of the evening.
The second threshold would occur if Mr. Romney surpassed the combined total of Mr. Gingrich and Rick Santorum — an outcome our model now considers about an even-money proposition. Currently, we project Mr. Romney to receive 44.0 percent of the vote, versus 43.2 percent for Mr. Gingrich and Mr. Santorum together. Such an outcome would give Mr. Romney a good talking point against the notion that Mr. Gingrich or Mr. Santorum could overtake him if one of the candidates were to drop out.
Finally, there is some chance that Mr. Romney will receive an outright majority of the vote — his forecast range runs up to 51 percent. In addition to having some favorable implications for the news media coverage about the race, this could also have some substantive impact.
Florida, in defiance of Republican Party rules for states that jumped in line to hold early primaries, plans to award all its delegates to the winner of the primary, regardless of the candidate’s vote share or margin of victory. It is possible that this preference will be challenged. However, other early-voting states award all of their delegates to the winner should that candidate reach 50 percent of the vote. If Mr. Romney meets that threshold, it may make it easier for him to retain all of Florida’s delegates even in the event of a rules challenge.
Newt Gingrich
FiveThirtyEight forecast (most likely outcome): 29 percent
High end of forecast range: 38 percent
Low end of forecast range: 20 percent
Absent a miracle victory in Florida, Mr. Gingrich’s goals are essentially the mirror opposite of Mr. Romney’s.
If the results are close enough that it takes some time to declare a winner in Florida, Mr. Gingrich might be able to declare a moral victory of sorts, chalking up the result to an uncharacteristically poor performance in the debates and to Mr. Romney’s substantial advantage in advertising dollars. These excuses are not necessarily convincing ones, but they are liable to be given more credence by the news media the longer it takes to call the state.
Barring a win or a close call, Mr. Gingrich’s ability to spin the outcome might depend on the extent to which he is able to point toward any signs of life in the exit polls. One reason that Mr. Gingrich’s victory in South Carolina seemed so persuasive was that he beat Mr. Romney among almost every demographic cohort. If Mr. Romney’s victory instead appeared to result from groups like Cuban Americans that have more presence in Florida than in other states, Mr. Gingrich might make a credible claim toward being poised to rebound in subsequent contests.
At the other end of the spectrum, a particularly decisive loss could make it harder for Mr. Gingrich to proclaim his continued viability in the race.
Rick Santorum
FiveThirtyEight forecast (most likely outcome): 14 percent
High end of forecast range: 21 percent
Low end of forecast range: 8 percent
Rarely has a candidate had a better reason to depart the campaign trail in a state early than Mr. Santorum. He took what was supposed to be a short journey away from Florida and then extended it because his 3-year-old daughter fell ill.
Mr. Santorum performed strongly in the Thursday debate and his standing has improved modestly in the polls. There is a remote chance that he could beat Mr. Gingrich and take second place — the margin separating Mr. Gingrich and Mr. Santorum is roughly the same as that separating Mr. Gingrich and Mr. Romney.
More likely, however, Mr. Santorum will simply be looking to beat expectations by accumulating as many votes as possible. He might have a rooting interest in seeing Mr. Gingrich do as poorly as possible, which could allow Mr. Santorum to cast himself as the most viable alternative to Mr. Romney.
Ron Paul
FiveThirtyEight forecast (most likely outcome): 11 percent
High end of forecast range: 18 percent
Low end of forecast range: 5 percent
Mr. Paul, like Mr. Santorum, has largely abandoned the Florida campaign trail in the late-going, choosing instead to campaign in states like Maine and Colorado that will soon hold caucuses.
Mr. Paul had a strong ground game in Iowa and New Hampshire — but that will produce a smaller marginal benefit in a state like Florida that has a large population and expensive advertising markets.
Whether Mr. Paul finishes toward the higher or lower end of his forecast range might provide some indication of how robust his support is in the absence of his campaign operation. But otherwise it will make relatively little difference, as he is already focused on the caucus states that will vote in February.