ELECTIONATE OCTOBER 28, 2012
The need for additional polls was waning over the last few days. The state of the race and pollster house effects have been stable for so long that the results of the next poll can usually be predicted just by knowing the state and the firm. But last night, two less-frequent yet strong pollsters surveyed their home turf and added a different take on two crucial states.
Plenty of surveys have found a one or two point race in Ohio, but with one exception since the first presidential debate, live interview surveys contacting cell phones have shown Obama leading by 3 to 5 points in Ohio. This pattern added credibility to the case that Obama held a larger lead in Ohio than a simple polling average might suggest. But last night's well-regarded Ohio Newspaper Poll (conducted with the University of Cincinnati) found a tied race at 49 a piece, and although it's just one poll, it's the best survey to support the possibility of a dead-heat in Ohio. The small number of undecided voters should also raise eyebrows. Most Ohio polls show 5 or 6 percent undecided, but the Ohio poll found few, if any. If that's the result of pushing undecided voters harder than other pollsters, then there's a chance this is a harbinger of late movement in the Buckeye State.
On the other hand, the Ohio Poll was conducted over the same period as many other surveys last week, and all showed a larger Obama advantage. That doesn't mean that the Ohio Poll is wrong, but it means that the tighter race can't be attributed to the possibility that Romney has made recent gains, since similar polls conducted over the same period show Obama ahead by a few points. For that same reason, the Ohio poll isn't sufficient to change the assessment of Ohio in the absence of additional data.
Further east and closer to Hurricane Sandy, the Washington Post found Obama leading by 4 in Virginia, 51-47. Like the Ohio Poll, the Washington Post poll diverged a bit from the mean, although this time the poll was to the president's benefit. Most Virginia polls have shown a true dead heat and the Washington Post poll is the first live interview survey since CBS/NYT/Quinnipiac to show Obama with a modest advantage. A recent Fox News survey, for instance, showed Romney ahead by 2 points. For what it's worth, much of the difference between the two surveys appears to lay with non-white voters, with the Washington Post poll showing Obama performing near '08 levels and with '08 turnout, while Fox News does not. Again like the Ohio Poll, the Washington Post poll has some support from automated surveys like PPP, but it's not yet sufficient to revise the view of the race.
There were also two polls in the intriguing state of Minnesota, where demographics and history suggest that the race should be quite close. The two surveys showed Obama ahead by 7 and 3 points, roughly in-line with the other October surveys showing Obama ahead by 4 to 10 points. Realistically, Minnesota is close and should have been more aggressively targeted by the Romney campaign, but the balance of evidence suggests that Obama maintains a modest lead.