Three NYT/CBS/Quinnipiac polls grabbed the headlines by showing Obama with double-digit leads in Ohio, Pennsylvania, and Florida. Importantly, Obama hit 53 percent in all three states, giving him more than enough of the vote to withstand substantial losses among undecided voters.
Obama’s lead in the battleground states is getting large enough that the margin might not accurately represent his advantage. The remaining undecided voters are probably latent Romney supporters—voters who tend to vote for Republican candidates, disapprove of the president’s performance, but dislike Romney. Should they turnout and cast a ballot in the presidential race on Election Day, they’re probably not going to vote to reelect the president.
Obama might not be quite at his ceiling nationally, but Romney still appears to have more growth potential. Most polls find Romney down into the mid-to-lower forties, a tally even beneath McCain's 46 percent from 2008. The polls tell us that Obama's disapproval rating is somewhere around 47 percent nationally, so it's fair to infer that many of the undecided voters disapprove of the president's performance. Most of these voters will eventually flock to Romney's side, perhaps as soon as after the first presidential debate.
Similarly, Quinnipiac found Obama's approval rating near 50-47 in Ohio and Florida. The voters who are ambivalent about the president's performance but currently support him appear to doubt Romney’s fitness for the presidency: they’re not sure about whether he can restore the economy, and they don't even think he cares about people like them. This problem even extends to voters who disapprove of the president's performance, but who just aren’t convinced by Romney's pitch.
At the same time, Obama’s 50-47 approval/disapproval isn’t far off of the national numbers, which might suggest that the poll is more accurate than the big margin suggets. The apparent difference between the state and national numbers could be that Team Obama has succeeded in disqualifying the Republican challenger in the battleground states. Reading a bit into correlations, you might think that Romney's big problem is the perception that he doesn't understand the problems of voters. About 57 percent of voters didn’t believe that Romney cared about people like them, a number that happens to match-up pretty well with Romney’s 43 percent of the vote.
I'm about to oversimplify things (these categories aren't divided cleanly, as I've depicted below), but you can envision the electorate breaking down accordingly:
-43 percent disapproves of Obama, thinks Romney understands them, and are voting for Romney;
-4 percent disapproves of Obama, doesn’t think Romney cares about them and are consequently are undecided;
-3 percent don’t think Romney understands them, are ambivalent Obama’s performance, and like the president personally. They’re currently voting for Obama;
-50 percent approve of Obama’s performance, don’t think Romney understands them, and they’re also voting for Obama.
My suspicion is that Romney is going to win over the 4 percent who disapprove of Obama's performance by Election Day, especially since history tells us that a major party Republican will struggle to fall far beneath 47 percent in any of these battleground states. So in a sense, these polls don't suggest that Obama's going to win Florida by 9 points or Ohio by 10--53 percent is probably near his ceiling in these states. But they do suggest he is very well positioned heading into the heart of the campaign.
Besides relatively reasonable approval ratings, voter registration numbers in Florida should also influence our reading of these polls. Quinnipiac took an unusual but very useful approach: they asked voters the traditional question about whether they identified as Republicans, Democrats, or Independents, and then they asked them a second question in Florida and Pennsylvania: are they registered as Democrats or Republicans.
Given recent debates, you won’t be surprised to learn that the result of the normal party-ID question was again the subject of controversy, since the Quinnipiac poll showed Democrats holding a 9 point edge in Florida, compared to 3 points in the high-water election of 2008. But self-reported voter registration aligned comparably better with the official data from the Florida Secretary of State; In Florida, 40 percent are registered as Democrats compared to 36 percent of Republicans, while the Quinnipiac poll found 43 percent identifying as Democratic compared to 36 as Republicans.
That's still a Democratic lean of a net-four points, and perhaps a little more after recalling that this was a likely voter survey. But it's still closer to expectations than the traditional party-ID metrics, which indicates that the "skewed" party-ID tallies are underpinned by a real phenomenon: more Republicans calling themselves independents. More on this later today.