It's time to just say this clearly: A straightforward read of the polls suggests we're likely to see Mitt Romney win the popular vote and Barack Obama win the electoral college -- and, thus, the presidency. But most pollsters don't think that will happen.
The national polling averages are clear: Romney is slightly ahead, and has been for weeks. As of this writing, he leads by 0.9 percent in the Real Clear Politics average and by 0.4 percent in the Pollster.com average.
But the state polling averages are also clear: Obama leads in Ohio, Wisconsin, New Hampshire, Iowa, Nevada, Michigan and Pennsylvania. Colorado and Virginia are essentially tied. Florida and North Carolina favor Romney. Even if Obama loses both Colorado and Virginia, he still wins with more than 290 electoral votes.
Friday morning, I asked a range of pollsters why I shouldn't, at this point, view a popular vote/electoral college split as a very likely outcome.
Tom Jensen, of Public Policy Polling, agreed. "Over the last couple weeks we’ve done polls finding Obama doing more than 10 points worse than 2008 in states like Connecticut, Massachusetts, Illinois, and Washington and those big shifts make a Romney popular vote win more likely but of course have no impact on the electoral college," he wrote in an e-mail. "And Obama’s drop in key swing states like Ohio, Virginia, Florida, and North Carolina relative to 2008 hasn’t been as small as the national drop because he’s run such a strong campaign in those places and in the latter 3 because they’re so racially diverse so his big decline with white voters doesn’t have as big an impact."
Scott Rasmussen, of Rasmussen Reports, was more circumspect. "For that to happen, Romney would probably have to win the popular vote by less than a point, and Obama would have to win Ohio and Wisconsin," he said. "It could happen, but there's plenty of time for that scenario to change. A small shift could put Obama ahead in popular vote. Or, Romney could win the popular vote by 2 or more points. If so, he probably wins electoral College."
A number of pollsters and poll-watchers said they simply didn't trust the polls right now: They figured that either the national polls or the state polls were slightly off. "I still wouldn’t bet on [a split]," said Alan Abramowitz, a political scientist at Emory University. "Right now, the gap looks too large to me to be real. I can’t see Obama leading by the margins he appears to have in some of the swing states and still be trailing by 2-3 points nationally. And I think the most likely explanation is that the national polls right now are a little off in part due to the influence of Gallup and Rasmussen."
Another argument I heard was that with the polls this close, the election really comes down to turnout. "The number of variables in terms of turnout that still exist are major," e-mailed David Winston, director of the Winston Group. "Party ID, minority composition, younger voters [all] make it more complicated to assess.'
In other words, in an election this closely divided and with so much effort focused on historically low-turnout groups like Latinos and young voters, the polls have to do a lot of guesswork to figure out the shape of the electorate, and they may be guessing wrong. If Obama's ground game outperforms Romney's ground game in a big way, Obama could easily win the popular vote. But if Obama's coalition stays home, then Romney could easily win both the popular vote and the electoral college.
Most of the pollsters I spoke to seemed cautiously confident that we weren't going to see a split between the electoral college and the popular vote. Such splits, they said, are very rare in American history, and they tend to be the result of extraordinary circumstances or electoral irregularities (like the "butterfly ballots" in Florida). Nevertheless, a split is, at the moment, what most of their polls are showing.