A Newsweek blog had an interesting article last week about the Bradley Effect and Reverse Bradley Effect involving how honest people are regarding race when doing phone polls.

It seems the article is right about the effect and wrong about the cause. If Obama was winning by larger margins than the polls it wasn’t because people were dishonest about wanting to vote for a black man it is because of a mistake in the poll.

Polls depend on estimates of likely voters based on past elections. If white evangelicals typically have 80% turnout then pollsters will count up the white evangelical population and discount their presence in the poll by their turnout percentage.

This is typically a good strategy but it causes issues when you have large departures from past voting habits. This election there has been a surge in registration and participation by blacks and youth. Both groups will obviously break for Obama. But these groups have poor turnout in the past and thus are underrepresented in polls. Plus young people who are more likely to use a cell as their primary (or only) phone are less likely to be reached by pollsters. The article mentioned this effect, but I think this is a far bigger factor than the reasons and studies listed by Newsweek.

I think this accounted for the boost Obama got in the primaries that was unaccounted for in the polls. It stands to reason that it’ll happen again in November.

This entry was posted on Friday, October 17th, 2008 at 2:50 pm and is filed under Politics. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.