Finally, we have real data showing a shift to Romney. Both Quinnipiac and Rasmussen show a clear trend in the data (comparing polls from the same pollsters over time – apples to apples). Prior claims of a shift, as I noted previously, were not correct – just lucky.

So it looks like Newt is losing some ground. But how much?

Impossible to say because all these polls are all based on unreliable turn out models for “Republican likely voters in Florida”. How you define that class of poll responders depends on how accurate your poll is.

And how you typically determine the voter pool for any primary election is using historical turn out data. Which will not work this year.

Right now it is safe to assume Romney could win, but I would not lay more than a dime on the line for it. Because this election is not typical of past Florida primaries on many fronts.

First off, we still have the 2010 insurgent voter out there. This can be seen in the fact that the current GOP voters in Florida are not the same ones from 4 years ago:

Republicans have narrowed the Democrats’ registration edge in Florida since November 2008, when Barack Obama carried the state. And with the Jan. 31 primary still nearly two weeks away, more than 446,000 Florida Republicans have requested absentee ballots — far exceeding the 307,744 absentee requests for the 2008 GOP primary.

Figures released by the Division of Elections today show Florida has 11.2 million voters, with 40.5 percent registered as Democrats and about 36.2 percent as Republicans. The gap of 4.3 percentage points between Democratic and Republican registrations compares to a 5.8-point gap that favored Democrats heading into the 2008 presidential election.

“The gap is closing due to the enthusiasm people have to oust Obama,” says Republican Party of Florida spokesman Brian Hughes. “People are hurting. The economy is turning around in Florida, but slowly, and they see at the national stage there’s not enough momentum and they’re ready for a change in leadership.”

This means the voter models used by pollsters are likely not accurate to what the turn out will be Tuesday. How many of these new registrants are insurgent voters who have no history of voting in Florida Primaries but took the time to register since Obama took office? My guess is this new GOP voter is likely to fail the ‘likely voter’ screens. They have no history of voting.

This is what caused many to underestimate the 2010 backlash wave. The mood of the electorate was so energized it defied all historical trends. This is the Achilles’ Heel of polls – they rely on stability in the voter pool to bring confidence to their turnout models.

The second big change for Florida is their role as kingmaker. Florida moved their primary date up again this cycle (cutting their delegates from 99 to 50, with no super delegates) to move off of one of the Super Tuesdays. One thing is true, if voters don’t feel their vote counts, they don’t take the time to vote. Florida has never been in this position, where their vote will make a huge difference in who takes on Obama. So voters are going to come out in historic numbers (like they did in SC).

This ALSO destroys turn out models. Primaries are the most volatile of elections to gauge in terms of turn out, since for Presidents seeking a 2nd term their party primary is a fate compli. This means half the historic record is inaccurate from the beginning.

Given the completely unique political environment we have today, a surge in GOP voter ranks in Florida and a primary that actually matters forget the turn out models in these polls. If their statistical Margin of Error is 3-4%, their actual margin of error due to turn out model uncertainties could range from 5-10%.

Be prepared to be surprised.

Update: Hot Air poses a question with some answers:

And with another day we get another fresh round of polling showing one of two things: either the good citizens of Florida are prone to fits of multiple personality disorder or the pollsters are just having some fun with us.

Or, the Florida race is unmeasurable by pollsters for the reasons I gave above.

Addendum: Watch for one other phenomena that might arise. The Santorum and Paul voters may realize their only hope of stopping Romney (and RomneyCare) is to get with Newt. If the vote is to stop the establishment at all costs, this could easily happen. In some polls 30+% are open to changing their minds. In SC that late deciders broke to Newt in a wave. Paul has been signalling he has not interest in the White House any more, so his supporters should shift to someone more viable given the meaning of this race.

Tags: Florida Primary, Gingrich, Paul, Polls, Santorum