Two media polls out in normally safe Democratic states show the presidential race in considerable flux — and pundits with few touchstones for their final calls. First we go to Pennsylvania, where Mitt Romney has made a last-minute personal push for a state that hasn’t gone Republican in 24 years — and the new poll from the Pittsburgh Tribune-Review suggests he has a real shot at winning:
President Obama and Republican Mitt Romney entered the final days of the presidential race tied in a state that the campaigns only recently began contesting, a Tribune-Review poll shows.
The poll showed the race for Pennsylvania’s 20 electoral votes locked up at 47 percent in its final week. Romney was scheduled to campaign in the Philadelphia area on Sunday, and former President Bill Clinton planned to stump for Obama on Monday. The campaigns have begun to saturate the airwaves with millions of dollars in presidential advertising.
“They’re both in here because of exactly what you’re seeing” in this poll, said Jim Lee, president of Susquehanna Polling & Research, which surveyed 800 likely voters Oct. 29-31. Most of the interviews occurred after Hurricane Sandy inundated Eastern and Central Pennsylvania. The poll’s error margin is 3.46 percentage points.
The PTR didn’t release the poll internals, but Lee told me this week that the poll is D+7, 44/37/19. SPR wanted to stick with one turnout model, and that closely resembles 2008’s electorate in PA. This result matches very closely to yesterday’s SPR poll results on the Senate race in PA, also a virtual dead heat.
In Michigan, perhaps the results come as an even bigger surprise, where Obama figured to ride the auto bailouts to an easy victory. A new poll from Foster McCollum White Baydoun of over 1900 likely voters has Romney ahead of Obama by less than one percent, 46.86%/46.24%. The same pollster had Obama with nearly a 4-point lead before the first debate. Also, this same poll shows Democratic incumbent Debbie Stabenow beating Pete Hoekstra rather easily, 50/43 — so it doesn’t appear to have a significant Republican tilt.
In fact, the sample is reasonably Democratic, although perhaps underrepresenting independents. The D/R/I is 44/35/21; in 2008, 41/29/29, and no exit polling exists for 2010. The poll did reweight in several categories to make up for deficiencies in demographics, as almost all pollsters do, but there are a couple of interesting points about the weighting, too. After their initial sample only included 8% African-Americans, the pollster reweighted them to 17.5%. In 2008, though, this bloc only comprised 12% of the MI electorate. At the same time, though, they appear to have overweighted seniors at the expense of middle-aged voters, but the two together look similar to 2008.
It looks like a surprise could be brewing in two Democratic states.