Race Matters: Obama and the 'Bradley Effect'
Let's talk about race.
Actually, let's not. I'm white and Jewish -- two groups not exactly very oppressed these days. So what do I know? But we can talk about the impact race might be having right now on the presidential election, and how it could all play out in November.
The dirty little secret no one will talk about (well, almost no one -- Mickey Kaus is unafraid to go there) is the Bradley Effect. Put simply, people lie to pollsters when the candidate in question is black. How many of them are actually racist? Probably not too many -- but no one these days wants to even be thought of as a racist. So they lie. In the weird world of American race relations, that's progress.
How big is the Bradley Effect? Nobody knows. But Kaus theorizes that it -- along with McCain sucking up a bunch of the independent voters -- might have been enough to give New Hampshire to Hillary Clinton back in January. Heck, as soon as Obama looked like a lock on the nomination, even people who really did intend to vote for him seemed to switch back to Clinton, giving her lots of wins at the end. Too little, too late for Hill -- but maybe enough to give us a feel for the size of the Effect.
Call it ... two points? Three? I've seen estimates as high as five percent, but let's be conservative here.
The next question is: Is the Bradley Effect evenly distributed, or does it vary by state or region?
My guess -- again, just a guess -- is that it varies, and maybe pretty widely. Counterintuitively, perhaps, I'd say there's less of an effect in the South. People there deal with race (not -- ahem -- always very successfully) almost every day, and in ways the rest of the country never had to. And frankly I'd expect a southerner to be more willing on average to tell a pollster exactly what he or she really thinks.
In liberal bastions like California and New York, the Bradley Effect might be greater than in other places. There's more pressure to conform to the prevailing liberal orthodoxy, and fewer conservatives to lean on, too.
Midwestern voters sure are nice, even to poll takers. I'd expect a pretty solid BE from Ohio through Kansas.