Apropos of my last post about polling problems, I just had a strange thought. Is it possible the systematic pro-Democrat bias in 2014 polling was a vestigial remnant of the Bradley effect?
A quick reminder, the Bradley effect refers to California voters telling pollsters they'd vote for L.A. mayor Tom Bradley - an African-American - for governor when they knew they would vote for George Deukmejian instead. Bradley led in the polls but lost.
It is possible some few individuals intended to vote Republican as a way of expressing disapproval of our African-American president, but didn't want to admit it as they feared it made them appear vaguely racist. It wouldn't take more than a few percent to throw off the polls.
If that were the case, there may have been nothing wrong with the pollsters' models, referenced below. I wonder if any pollsters build a Bradley effect fudge factor into their models?
Just sayin' ....
