The Bradley Effect

September 24, 2008

There have been a few recent news items about white supremacists campaigning against Obama.  Obama’s racial identity will of course have an effect on the election, but it is unclear how much negative impact his race will have on the white vote.  This could be determinative in an election in which there is a large majority of white voters.  It was the strong preference for Bush among whites (58% to 41% for Kerry) that accounted for Bush’s second term.

While a surprising 20% of the people admit that race will play a part in their vote, most people do not admit racial bias of any sort, many I am sure are unaware of it.

In 1982 Tom Bradley, a popular Los Angeles mayor, ran for governor of California and led in all the polls.  When he lost a variety of explanations followed. There was a gap of seven percent between how people said they were going to vote and how they actually voted.  One explanation was that the polls could not accurately predict the voting habits of whites with respect to a black candidate.   This unexpected gap of seven percent between what all the polls predicted for Bradley and the number of votes that he actually got is sometimes called the “Bradley Effect” by those that attribute the gap to racial bias that is not uncovered by political opinion polls.

If there is a Bradley Effect, then Obama is not even with McCain until the polls show a lead of 7% (assuming standard deviation with the one 26 years ago).  My point is that this is not a time form Obama supporters to feel smug.

The primaries may suggest that the Bradley Effect is still strong facotr with many voting results not matching exit polls and uniformly favoring Clinton over Obama.