Florida Polling

 (Geology.com)

Most polls in the last week show Romney ahead in Florida. The last polls in the state show Obama +1, Romney +3, Romney +7, Romney +4, Romney +1, and Romney +1.  Average them together as RCP does and it's Romney up 2.5.

The poll showing Obama up 1 is the oldest so obviously the trend line is moving in Romney's direction since the debate. I don't usually like to get bogged down in the internals of polls, since almost every poll's sample is off a bit and that's why it's best to just average them. But the poll showing Romney up 7, had him winning among Latino voters. Some other recent polls show Florida Latino voters moving toward Romney with Obama's lead being cut from high teens to single digits, so again that is also trending in Romney's direction. Florida's Latino voters tend to be more conservative than Latino voters in the rest of the country so this isn't a huge surprise. But Obama won Florida Latinos by 15 points in 2008, so he hoped to do at least that well in 2012, especially considering that nationally he's leading Romney among Latinos by around 50 points.

Here are some demographic changes to think about, though. According to reports there are 44,000 less white voters registered in Florida in 2012, while there are 55,000 more black voters registered and 195,000 more Latino voters registered.

The caveat here is that in Florida, it depends what type of Latino voter you're talking about. If most of the 195,000 newly registered voters are Cuban, then Romney may end up having a net gain in this demographic.

In 2008 Obama won the state 51%-49% by a margin of 236,450 votes.  White voters made up 71% of the electorate in 2008. Depending on turnout, registrations at least so far show that white voters will make up less than 71% of the electorate in 2012.  Anyway, I did a little estimation that is hardly scientific, but I subtracted 44,000 from the number of white voters who voted in 2008, while adding 55,000 to the number of black voters who voted in 2008 and 195,000 to the number of Latino voters who voted in 2008.

But let's try this and play around with the numbers and give Romney 60% of the white vote (up from 56% in 2008), give Obama 55% of the Latino vote (down from 57% in 2008) and black support remaining at 95%, this election is a toss up, with Romney ahead 4,190,075 to 4,176,404, a lead of less than 14,000 votes. If you tweak the Latino support to give Obama the 57% he got in 2008, he's ahead by about 40,000 votes, 4.2M to 4.16M.

If you play around with the numbers further and give Romney a 2 point lead among Latinos, as the one poll showed, then he'd be winning the state by 180,000 or so votes 52-48, or roughly what some of the polls are finding.

So I think the keys in Florida are: 1) how much, if any, did Romney increase his share of the white vote, 2) how much, if any, did Romney increase his share among the Latino vote, and 3) turnout, turnout, turnout.

Polls are simply predicting and estimating all three of these factors. But no matter what the current polls show, I'd be shocked if either side won this state by more than 100,000 votes this time around.