Different Bradley, Same Effectiveness

October 6, 2008

When folks at the National Review weigh in on race, it’s always a safe bet that most people not at the National Review will get offended. So you knew there was going to be trouble when Jonah Goldberg wrote this:

Because I’m so cool, I’m watching a tivo’d “This Week” and Stephanopolous said that one of the ways the polls may be “undercounting” is that they aren’t counting all of the people who won’t vote for Obama because he’s black. Nonsense. The people who aren’t voting for Obama “because he’s black” have undoubtedly already told pollsters they’re not voting for Obama. If there’s much undercounting because of race, it’s overwhelmingly because non-racist people are afraid to tell pollsters they’re not voting for the black guy for fear of people like Stephanopolous calling them racist.

And, if there are racists who might be costing Obama the election, they’re Democrats.

This went over with the rest of the blogosphere about as well as one might expect. The odd thing about this dispute, though, is that Jonah is on the right side of it. This is not to say that his argument makes any sense. If someone – or, in this case, almost everyone – is predicting elections based on a story he made up that strikes him as kinda plausible, the correct response is not to make up your own, conflicting story that strikes you as kinda plausible, but rather to demand they provide some concrete reason to believe in their story or else shut-up about it.

Polling and projecting is a tricky business. It would certainly be a mistake to take every poll at face value, as systematic inaccuracies like the Bradley Effect can and do plague a lot of polling data. But vague hand-waving isn’t the solution, and, as has already been covered here, actual analysis of Obama’s results in the primaries doesn’t support the widespread intuition. There is certainly a history of black candidates underperforming their poll numbers in this country, but attitudes about race have shifted dramatically over the course of that history, and there have been exactly zero black candidates for president on major party tickets, so there simply isn’t much relevant data.

There may well bea phenomenon that roughly corresponds to what people think of as the Bradley Effect, at least in some parts of the country. But there are much stronger cases to be made for other polling problems, such as the cell-phone only effect, which are talked about less, because the fairy tale behind them is less fun to talk about.

Advertisements

3 Responses to “Different Bradley, Same Effectiveness”


  1. […] my friend Akhbar set what I take to be a reasonable standard for sorting out the reasonable thoughts from the […]


  2. […] on the Bradley Effect October 7, 2008 A response to my dismissal of the Bradley Effect, drawing on Nicholas Kristof’s article ‘Racism without […]


  3. […] I questioned blogger Akhbar’s application of evidential principles to Jonah Goldberg’s argument about […]


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: