After reading the link that Chris_ posted, there's good news. One, they undersampled R's by at least 9 points for the way turnout will actually happen. Two, assuming that all the R's and D's voted along party lines, independents are breaking 80% Romney to 20% for O.