As of this writing, Barack Obama is not polling consistently above 50 percent in a number of electoral-vote-rich swing states, including Ohio and Florida. He should be worried. ~Bill Greener

Yes, he is, and no, he shouldn’t.  There are two problems with Greener’s argument.  First, it doesn’t matter whether Obama polls above 50% in Florida and Ohio, because he has been consistently polling above 50% in Iowa, Colorado and New Mexico.  With these states and all of the Kerry states, where he leads comfortably, Obama can and will win even if he loses all the other toss-ups (and it is unlikely that he’s going to lose Virginia or Nevada).  McCain’s grand prize of Pennsylvania seems to be out of reach, as Obama’s average lead in the state is 11 points.  At present, we are seeing Florida and Ohio trend towards Obama, but even if they do not flip to the Democrats Obama can win.  As it is, Obama’s RCP average in Ohio is at 49.9, and in Florida it is 47.8.  It’s conceivable that McCain can rally to hold these two states, but that will almost certainly have no effect on the ultimate outcome.  At this point the GOP is in the position of an outmatched team that is just playing for pride against a far superior opponent that has the game wrapped up; the best that they can hope for is to narrow the margin of victory and put in a respectable final effort. 

The other, related problem with Greener’s argument is that the undecided voters don’t matter, either, because Obama is far enough ahead in enough states that all undecided voters could break for McCain (who also technically represents the incumbent in this scenario, not the challenger) and still not put him over the top.  The problems with Greener’s argument are plain to me because I was arguing much the same thing for much of this year, and it has become clear that I did not take into account the possibility that Obama would poach enough voters from the Bush coalition to make what probably are McCain-friendly undecided voters more or less irrelevant to the outcome.  Even if you accept that undecided voters are more likely to break for McCain, which sounds plausible but may not be correct, there is no realistic scenario given the information we have right now in which they will propel McCain to victory.

Greener then goes on to make somewhat strained claims about 2006 elections that purportedly show the “undecided” voter’s bias against black candidates.  It seems to me that there is a combination of factors at work in these cases.  Take the Ohio governor’s race.  Strickland outperformed his polling, and Blackwell underperformed his, or more precisely polls that did not force respondents to declare which way they were leaning underestimated Strickland’s support and overestimated Blackwell’s.  But what Greener actually manages to show is that undecideds in Ohio and Maryland broke for the candidate representing the challenger party and against representatives of the incumbent party: late-deciding voters went against the known quantity and chose the new face, which would seem to undermine Greener’s larger argument. 

It is instructive to compare the 2006 Ohio Senate race’s polling to the final result to see how much greater Sherrod Brown’s final result was than pre-election polling.  What I think you will find is that Sherrod Brown outperformed his polling as much as, if not more than, Strickland.  One reason for this is probably stronger Democratic and Democratic-voting independent turnout in the ’06 elections.  DeWine had some advantage as an incumbent Senator despite low approval ratings, while Blackwell was running in an open election as a member of a deeply unpopular state administration.  So Blackwell was part of and therefore associated with the Taft administration, even though he was personally not implicated in any of the corruption tainting the governor, and he was running against a popular economic populist who was able to make inroads in southeastern Ohio because of his roots in that part of the state.