Currently, national polling assumes a big dropoff from registered voters to likely voters. I don't believe that'll be the case, and we're certainly not seeing it in the early vote—Democratic turnout is up. And the RV models have been more accurate historically.
So I'm going with the registered voter models. Obama will win the national vote by a little over three points.
Colorado has been trending Obama's way hard over the last several days. PPP has it 52-46, but Reuters/Ipsos has it 48-48 among likely voters but ... 50-42 Obama among registered voters. So I'm pretty comfortable about an Obama victory, except for the early voting. Four years ago, Democrats came out of the early voting period with a 38-36 advantage, but this year the GOP holds a 37-35 advantage.
Much of that could be the GOP vote shifting from Election Day to early vote. Remember, Republicans will vote no matter what, while it's our voters who are less likely to turn out. That's why you see those big RV-LV vote differentials in the polling. But I'm going to hedge and fear the worst—that our intensity is down. But the polling shows a clear Obama lead, so I'll shave some of that off.
Florida is a coin toss, so I'll call it for our team on the basis of the early vote. Most of the recent polling I checked had Obama and Romney splitting the independents, so if that happens, and we get our people out (which is happening), we win.
Iowa and Nevada have been consistently moving away from Romney. I'm calling Iowa pretty much based on Ann Selzer's latest poll, since she's the gold standard when it comes to the Hawkeye State. Jon Ralston, who knows Nevada politics better than anyone, says Obama +4 in the Silver State. I'm being dumb by not parroting Ralston, but hey, I'm a partisan optimist.
There's not much chance New Hampshire will matter, but it's being hard fought regardless. The polling consensus looks about right, so I'll stick with Obama +3. There will be far less drama in Wisconsin. It flirted with Romney after the first debate, then decided that yeah, he really is a dick. The polling composite says Obama +5.3, which is a safe bet. But with momentum and the trends headed in Obama's direction, I'll round up to +6.
Ohio is the big one, and it has always been Obama's best firewall. The polling composite gives Obama a comfortable three-point lead, but it's that close just because of the GOP hack pollsters in the field. Take out Rasmussen, and it's a 3.9-point lead. I'm going with five, which really, after all the money and time and attention showered on the state, would be really pathetic for Republicans.
So that leaves North Carolina and Virginia. The polling in the Tarheel State has narrowed, and let's remember, Obama trailed in the polling average in 2008 before winning the state by a smidgeon. The polling has also an assumption that "unlikely" Democrats won't vote, yet we've seen that key Democratic demos have turned out at similar or higher rates than in 2008. Just peruse MattTX's excellent daily series on the early vote.
However, even as Democrats have turned out, Republicans have turned out in even bigger numbers. Again, this could be vote shifting from Election Day to early voting, but Obama's victory margin was already so razor thin that even a slight uptick in intensity for the Republican nominee could erase that advantage. So I'm going to go with a narrow Romney victory, but let's hope I get this one wrong. And in Virginia, the polling has been dramatically improved after the first debate nadir, with the consensus being a 2-3 point Obama lead.
Early voting was looking tough for Obama the first week or so, but then Saturday happened:
VA EARLY VOTE: On Fri, turnout was just 67.6% of '08 absentees in Obama counties, 75.8% in McCain. Now 83.3% in Obama, 86.7% in McCain
— @Redistrict via web
The polling composite is 1.8 point-Obama lead, but take out the crap pollsters like Rasmussen, and the numbers fall in line with my 2.5-point prediction.
Comments are closed on this story.