REVERSE BRADLEY EFFECT IN SOUTH CAROLINA?

Mark Sanford beat Elizabeth Colbert Busch yesterday in a special House election in South Carolina, and it wasn't even close: Sanford beat Colbert Busch 54%-45%, despite a survey from Public Policy Polling showing Sanford with just a 1-point lead.

PPP seems to have been the only pollster seriously following this race, and previous PPP polls showed Colbert Busch with a lead. But I think PPP's failure to call this race is understandable. My guess is that poll respondents -- even respondents to PPP's robo-poll, who never spoke to an actual human being -- were reluctant to admit that they were going to vote for a philanderer. I think it's a reverse Bradley effect -- Republican-leaning voters in this district lied to PPP's poll robot just the way white voters lied (or used to lie) to pollsters about their intent to vote for black candidates, as in the 1982 California governor's race, in which pollsters overestimated the strength of African-American candidate Tom Bradley.

Pollsters have learned how to add questions to surveys that will tease out discrepancies between a respondent's stated preference for a candidate of another race and real likelihood of voting for that candidate. In hindsight, it seems that PPP (or another pollster) should have found a way to tease out a "Sanford effect." Then again, who knew? We naively thought these South Carolina Republicans really believe what they say about traditional values.

I'm predicting this won't be a problem in the New York mayoral race if Anthony Weiner decides to run. One way or the other, I think my fellow New Yorkers and I will tell you the truth about Weiner. I really don't think we're hypocrites or self-deceivers. That sort of phony sexual moralizing isn't likely to be a factor here. But we'll see.