what bradley effect?

In the weeks leading up to the election, when Obama was shown to be consistently ahead in the polls, once burned twice shy Democrats bit their nails and hemmed and hawed, spouting out reason after reason why the polls should not be trusted: pollsters were over-estimating Obama supporter turnout (all those newly registered voters who poured into polling places to vote in the primaries would suddenly disappear), Republicans would manage to invalidate the voting rights of the hundreds of thousands of newly registered voters if they even bothered to turn up in the first place, there would be so many voting machine malfunctions that long lines would prevent most people from voting, and, of course, all those white voters who claimed to be voting Obama were actually going to cast their ballot for McCain.

This became a favorite theme for the GOP as well as Election Day drew near. Don't worry, Republican colleagues introducing McCain would wink, all those voters who tell pollsters they're voting for Obama will do something quite different once they actually get inside the privacy of the voting booth. And yet, as the dust settles from a decisive Obama victory, the Bradley effect is nowhere in sight. The Democratic Senator from Illinoins captured 53% of the vote against the GOP candidate's 46%, giving him a 7% margin of victory (as compared to the final RCP poll of polls which showed Obama gaining an average of 52.1% with McCain at 44.5%, a 7.6 margin of victory). While these numbers may indicate that undecideds (who appear to be at 3.4% in the final RCP calculation) broke slightly for McCain, it shows no evidence that those who said they were supporting Obama secretly slipped a McCain vote into the ballot box.

Among the swing states, Obama frequently did slightly better than he did in the polls (in Indiana, Pennsylvania, and Colorado, Obama won by a margin 2-3 points greater than the ones shown in the polls; in North Carolina, Florida, and Virginia, Obama won by a margin of .1 to .8 larger than those predicted on RCP; McCain's margin of victory in Montana was 1.3 points closer than the polls indicated). McCain slightly outdid his poll predictions in New Hampshire (he narrowed Obama's victory by .6 points) and Georgia (where he won by a margin 1.1 points greater than projected). The most striking differences (and probably the only legitimately significant ones) were in New Mexico and Nevada where Obama won by nearly twice the predicted margins. These differences, however, are probably due to the unusually high youth voter turnout, especially in New Mexico where 70% of registered voters 18-29 came out (as opposed to 49% in 2004), rather than some sort of reverse Bradley effect.

So, in the first presidential election involving a major African-American candidate, the Bradley effect appears to have had no major consequences. Does that mean that it will not affect future elections? No. Does that mean that racism is dead? No. Does it mean that all the racists would never vote for a Democrat in the first place? No. But at the very least it means that when the country is facing the greatest economic crisis in a generation, two extraordinarily costly wars and rampant foreign animosity, an intelligent, charismatic, eloquent, personable, handsome black man with no skeletons in his closet who promises a new kind of politics in Washington can attract enough white voters to win the election. Which may not be much, but given the history of racial issues in America, it really is something.

Are all those "undecideds" really closeted McCainites?

We were all wondering who they were (God knows, I was), those undecided voters, and Bill Greener gives us an interesting, if infuriating, answer in his article today in Salon. He suggests that they are all going to vote for McCain. Further, he declares, unless Obama is attracting at least 50% of the vote in key swing states, he is likely to lose them because all the undecideds are white racists who are unwilling to admit that they're actually voting for McCain (I think he phrases it more along the lines of how undecideds "trend" toward the white candidate, but be that as it may); he cites two gubernatorial and two senatorial races where two non-incumbents (one black, one white) were running against each other in 2006. The problem, of course, with his argument is that much of it isn't exactly true. First of all, Greener cites a single poll in two of the cases (he refers only to a single Survey USA taken the day before the election when discussing the Ken Blackwell-Ted Strickland gubernatorial race and again with the Harold Ford vs. Bob Corker senatorial race) as opposed to referring to any sort of general trend in the polls. Furthermore, Greener's claim that "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" is misleading at best. Four of the last six polls in Ohio show Obama receiving between 50 and 53% of the vote (the remaining two show him at 49%). The polls in Florida are much tighter and less conclusive, but also potentially irrelevant (as are those in Ohio). Obama has been doing so well in Colorado, New Mexico and Virginia, where all but one of the last seven polls show Obama receiving between 51 and 54% of the vote that he doesn't actually have to win either Ohio or Florida to carry the election. Nate Silver also gives an excellent rebuke of Greener's argument here.

Undoubtedly the Democrats could still blow this election, but I doubt that it will be because of flocks of undecided voters secretly fall asleep musing dreamily about the prospect of McCain and Palin in the White House.