None So Blind
In a Fox News editorial, professor Susan Estrich writes:
The problem with this analysis is that it assumes exactly what it purports to prove: that any difference between exit polls and actual ballots is due to secret anti-black animus which emerges only in the privacy of the ballot box.
That was, according to the pollsters, the problem: about 10 percent of the electorate claimed that they were going to vote for [Tom Bradley, who is black], and in many cases even told pollsters that they did, but they lied.
Shocking. Racism in America. Who’d a thunk it?
Doug Wilder, who wasn’t elected to the Senate from Virginia, faced the same problem. We who are Democrats would like to believe that race is not a factor in the polling of our party members, but maybe we’re wrong.
No one doubts, or at least no one who is honest does, that both racism and sexism come into play as people decide between Clinton and Obama, but could it be that people are more willing to admit that they won’t vote for the woman than that they won’t vote for the black?
It is equally possible that the difference arises in precisely the opposite fashion: that voters are eager to tell pollsters how they're going to vote for the black. Perhaps they are not hiding truly anti-black feelings, but are pretending to pro-black feelings they do not actually possess. Steve Sailer calls this "status seeking", which seems partly true.
Ms. Estrich does not appear to have any plan for disambiguating these two complementary effects. She thinks she has found racism: but she has seen only her own expectations. We should all get used to this cheap substitute for analysis, because we are going to see a whole lot more of it.
[HT: Glenn Reynolds. Cross-posted to Chequer-Board.]