The answer is “yes,” if you believe a number of the polls currently in the Real Clear Politics poll average… but that’s a sizable if. The NYT/CBS poll shows Pres. Obama winning by 3 points among likely voters… but losing independents to Mitt Romney 44/50. The WaPo/ABC News poll shows Obama up by a point, while losing independents by 11 points, 43/54. The CNN/ORC poll (.pdf) shows Obama winning by six overall, but losing independents by a whopping 14 points, 40/54. Ironically, given people’s perceptions of poll bias, the Fox poll has Obama winning overall and winning indies by a similar 5 point margin.
Historically, a candidate cannot win the presidency while losing independent votes. However, it has happened thrice in my lifetime: 1976, 2000 and 2004. These exceptions are instructive.
In 1976, Carter won 50/48, while losing indies 48/52. However, the D/R/I breakdown that year was 37/22/41. The remarkably high number of indies and remarkably low number for Republicans suggests that post-Watergate, a lot of those indies were closet Republicans. In 2000, Dubya won indies 48/46, while narrowly losing the popular vote in a 39/35/26 D/R/I breakdown. In 2004, Dubya won 51/48, but lost indies 48/50 in a 37/37/26 D/R/I breakdown.
None of these exceptions provide much comfort to Team Obama. First, they highlight an overall Democrat ceiling of 37-39%. This was also true in 2008, the worst presidential environment for the GOP since 1976, resulting in a D/R/I break down of 39/32/29. Indeed, the increasing sorting of voters since 1976 suggests 2008’s 32% is a likely floor; this year’s GOP-identifying turnout is very likely to be higher. The 2000 and 2004 results, showing a narrow split in indies, suggest that Obama cannot afford to lose indies by six points, let alone 11 or 14 points.
Does this mean that the polls showing Obama winning overall while losing the independent vote are necessarily wrong, or biased? Not necessarily. Every poll has to be judged on its own strengths and weaknesses in total. Moreover, it is likely that these polls merely prove that Obama’s standing from roughly September 8-12 really did reflect a convention bounce already fading and likely to fade more, but which produced more people willing to identify as Democrats in these polls. I would expect that a week from now, polls may well show Obama with worse numbers overall, but perhaps slightly improved numbers with independents, as the weakest Dems and hardcore liberals shift back to an independent identification.