Going into the Republican vote in Florida on Tuesday, the polls couldn't be tighter but you can't deny, John McCain has had a heck of a couple days, seeming to shut off Romney's momentum and gaining some of his own. From what I'm hearing, prognosticators would have given it to Romney a couple of days ago without question; now, they're not so certain -- it's a matter of whether the McCain surge is real or just a temporary bump. Whereas three polls taken entirely on Sunday had McCain and Romney tied, the most recent poll out of Florida, Survey USA (979 LVs, 1/27-28, MOW +/- 3.2%), shows that John McCain had a good Monday, although not so much so that it would lead me to predict the result with any confidence.

In other words, by just adding Monday to the polling time range, McCain nets 1 point over Romney. Not exactly earth shattering but worth noting all the same, especially when we dig a bit deeper into the poll and find where McCain's Monday gain came from: early voters.

As of Sunday, the LVs who told both SUSA and Insider Advantage pollsters that they had already voted were evenly split between Romney and McCain. That shifted dramatically on Monday as you can see from the SUSA numbers of those who've already voted (disclaimer, these voters make up just 25% of the entire voter sample, so the MOE is very high):

A ten point net gain in voters who say they already voted for McCain just adding Monday's results. It seems almost unbelievable, it's such a dramatic shift, so I don't see it as determinative without further confirmation, but it certainly would indicate, if true, that Romney's going to have to have a very good day on Tuesday with election day voters if he intends to pull this out.

In addition to the early vote, another variable that will contribute to tonight's result is what Giuliani and Huckabee supporters decide to do once they enter the voting booth.

As Josh Marshall observes:

As the battle has become more and more clearly a McCain-Romney race and one quite likely to settle the GOP nomination, I think there's a good chance that Huckabee and Giuliani significantly underperform their margins. Because people want to actually have a say in who becomes the nominee rather than throwing their vote away on Rudy or Huck. That potentiallly throws a lot of other votes into the mix that could push either McCain or Romney past the finish line.

On this very topic, both Craig Crawford and Pat Buchanan said on MSNBC last night that they've heard there's word going out to evangelicals that "a vote for Huckabee is a vote for McCain," implying that it's being recommended to the Christian community that they vote strategically for Mitt Romney in order to stop McCain, whom they find unacceptable.

Add to that Rudy Giuliani's signals on Monday that his campaign may not be long for this world (saying the winner of Florida will be the nominee and leaving autographed baseballs - going away presents perhaps? -- for the press on his plane...) one has to wonder if he was in essence giving his blessing to his supporters to jump to another candidate. But to which candidate? Conventional wisdom would dictate that his supporters would be far more likely to shift to McCain, but again, we're left with the unknowns of how much lower these guys' actual support is today than their pre-election polling had indicated.

As I've written this, two new polls incorporating Monday polling have come to my attention, which I'm sorry to say, do nothing to clear things up as it's an even split, although still within the MOE. Below find the results of the new Braynard Methodology poll (500 LVs, Jan. 28, MOE +/- 4%) and the Reuters/CSPAN/Zogby tracking poll (941 LVs, Jan. 27-28, MOE +/- 3.3%.)

The Braynard poll would appear to push undecideds, which, by this model, go to Romney, but there's some dire news in the Zogby poll for Romney if true:

"Senator McCain seems to have consolidated support among conservatives, where he now leads by 7 points, perhaps as a result of the strong endorsements he received over the weekend from conservative Gov. Charlie Crist and Republican Sen. Mel Martinez. He is leading handsomely among moderates, and appears to be picking up support from likely voters who might otherwise have gone to Rudy Giuliani."

Then again, the poll that got closest to predicting Obama's SC margin of victory, Democratic polling outfit Public Policy Polling, does have Romney up by 7.

Despite the uncertainty of the result, I'm going to go with Romney as my prediction, partially out of wishful thinking but also because the one poll that was taken entirely on Monday has him slightly ahead, and looking at the SUSA internals, McCain's support appears much softer than Romney's, which tells me the bump that Crist gave McCain will disappear as quickly as it materialized. I'm also counting on conservatives (evangelical and otherwise) breaking for Romney. Final prediction: Mitt by 3. Care to lay one down?