The president’s modest lead in the critical battleground states and tenuous advantage in the national polls is small by historic standards, especially for an incumbent president running for reelection. Nonetheless, the president’s lead is clear enough and consistent enough to deem him an obvious favorite heading into Election Night.
Obama leads by at least 3 points with 49 percent of the vote in the states won twice by Kerry and Gore, plus New Mexico, Nevada, and Ohio. These states are worth 272 electoral votes, and with the exception of a stray poll in Michigan, Romney doesn’t lead in a single non-partisan survey in any of those states. Despite a close contest and Romney’s brief national advantage following the first presidential debate, Obama has never trailed in a polling average in these states.The polls also show the president with a narrow lead ranging from 1.3 to 2.5 points in four additional states that would bring the president to 303 electoral votes: Virginia, Colorado, New Hampshire, and Iowa. Florida is perhaps the closest battleground state, where Romney leads by .4 points and the two candidates have split the polls almost evenly, with Obama ahead in 7, Romney ahead in 8, and three tied.
The only battleground state where the polls show Romney with a clear edge is North Carolina. Here, Romney leads by 1.8 points with 48.1 percent of the vote. But the polls show the president better positioned in Ohio than Romney is in North Carolina, where Romney holds a smaller share of the vote, and a smaller and less consistent lead. Although this doesn’t suggest that the president is a favorite in North Carolina, it clarifies Romney’s challenge—especially if one has unwisely dismissed the president’s chances in the Tar Heel State.
The national polls have often provided better news to Romney than the state polls, but the president now holds a 1.3 point edge in national polls conducted since October 25, with Romney only leading in three of the final 21 national surveys--Gallup, Rasmussen, and Zogby’s JZ Analytics. To the extent that momentum is relevant, the president has steadily gained in national polls since mid-October and seized a 3 point lead in many of the better surveys adhering to the industry’s best practices conducted over the campaign’s final weekend. The president's new advantage narrows the gap between national and state pollsters, reducing the uncertainty about the prediction of either group of polls.
If the polls are right, Romney has a difficult task: sweep the five states where a non-partisan poll shows both candidates ahead, and then carry one of the states where Romney doesn’t lead in a single non-partisan poll. If the polls are about as accurate as they usually are, this would be tough to pull off. Usually, the averages correctly predict the outcome of all but one or two states, and Obama’s advantage appears broad enough to have a very good chance of withstanding typical polling errors. Instead, much of Romney’s chances depend on the possibility of a broader, systemic polling failure where surveys overestimate the president’s standing across the board. Such an error is possible, but it is unlikely. It might be worth observing that the polls showed Obama with a narrow but clear lead in Nevada, and most conclude that the president wrapped up the state during early voting.
There was nothing in yesterday's polls that undermined this view of the race. Gallup returned from its post-Sandy slumber and showed Romney's lead among likely voters shrinking to just 1 point. Similarly, ABC/Washington Post and Democracy Corps also saw a multi-point shift in Obama's direction. A well-regarded Ohio poll showed Romney trailing by just 1 point, but it was canceled by Obama claiming a 5 point lead in a new SurveyUSA poll.
Many have attempted to peer into the arcane details of the polls to detect signs of an impending polling catastrophe, but, for the most part, these efforts are unscientific and supported more by instinct than evidence. This isn’t to say that there aren’t reasons to believe that the polls are likelier to err this year than they have in the past: response rates have plummeted, the remarkable consistency of this year’s polling house effects suggests that methodological decisions could easily change the outcome, and many believe that low-frequency Democrats are likely to stay home in 2012. But there are also strong critiques of the polls founded in a methodological argument suggesting that traditional surveys might underestimate Democratic support, perhaps because they don’t call cell phones, let alone enough of them, or conduct bilingual interviews with Spanish-speaking voters.
The importance of turnout is often overstated, but Romney's chances depend on low Democratic turnout. A majority of registered voters support the reelecting the president, making it hard to envision how Romney could prevail in a full turnout election. But Obama's coalition is extremely dependent on demographic groups with a history of low turnout. Demographic changes can help compensate if turnout rates decline to '04 levels, but the possibility that turnout falls further than '04 can't be completely dismissed. Even so, likely voter models have a surprisingly good history of projecting the composition of the electorate, and although they show that Democrats are more likely to stay home than Republicans, they also find enough Democratic turnout to reelect the president.