The Bradley Effect is a myth

Say it after me: there is no such thing as the Bradley effect.

The Bradley Effect -- the theory that voters tells pollsters they're going to vote one way, but in the polling booth are overcome by their own racism -- is minor at best, and possibly entirely apocryphal. It may be that it used to exist: the effect was named more than 25 years ago, and the country has come a long way since then.

In the 2006 races it was not detectable at all. In fact, in 2006 black candidates outdid their polling figures. Nate Silver of the excellent FiveThirtyEight in fact speculates that there may be an "elephant effect" where people say they're going to vote McCain because they have to socially, but vote in their economic self-interest in the polling both, thus boosting Obama.

Furthermore, the largest percentage difference the Bradley effect was ever alleged to have created was a 3% gap. To swing this election back to McCain, there would have to be a nationwide shift of more than 6.5% at the last minute. Even if the Bradley effect is twice as bad in the general election in 2008 as it may or may not have been in a gubernatorial race in 1983, that still would not be enough to save McCain.

So please stop asking me about the Bradley effect. Nervousness is understandable this close to election day. It's hard to believe that the long national nightmare of Bush is over, that we really will have a democrat in the White House again, especially one as inspiring and well-intentioned and (whisper it) liberal as Barack Obama. The way to make yourself believe is to stop worrying and to vote -- right now, if your state allows it, but as early as you can next Tuesday if not.