Thursday, October 30, 2008

Bradley Effect


Have you heard of the Bradley Effect? The short description is that when polled, Caucasians will claim to be voting for an African-American candidate but when the same Caucasians are in the voting booth with guaranteed anonymity they will actually vote for a Caucasian candidate. The polls are skewed by the "white-guilt" of the responders.

Obama is ahead in the polls. How much of his lead is due to the Bradley Effect?

I also wonder about a form of the Dewey Effect (if you will recall, Dewey though he had won the election due to polls, but the polls were skewed due to the number of people without phones). Since increasing numbers of people are opting to not have a land-line telephone and to only have a cell phone, those individuals are under represented in the polls.

Do the Bradley Effect and the Dewey Effect offset?

1 comment:

Mark said...

In the case of Bradley, the polls went down slightly before the election day. I do believe this will slightly be in effect at least.

I wasn't aware of the Dewey effect, but I have wondered about weekend polls. Most families are either out at football games on Friday nights, or at church on Sundays, so the people at home to answer polls are a certain segment of society.

This will be very interesting. There are people who definitely won't vote for Obama because he's black, and there are other people who will vote for him no matter how bad he is for the country, simply because he's black.

I'm really nervous about this election.