John Derbyshire, Trayvon Martin, and the ignorance of racial profiling.

John Derbyshire, Trayvon Martin, and the Ignorance of Racial Profiling

John Derbyshire, Trayvon Martin, and the Ignorance of Racial Profiling

Science, technology, and life.
April 10 2012 9:40 PM

John Derbyshire’s Error

The ignorance of racial profiling.

(Continued from Page 1)

In a slam at black government workers, Derbyshire asserts, “ ‘The DMV lady’ is a statistical truth, not a myth.” But what exactly is a statistical truth? It isn’t a truth about your DMV lady.  It’s a probability estimate you might fall back on if you know nothing about your DMV lady. It’s an ignorant person’s weak substitute for knowledge.

The starting point for Derbyshire’s talk is street crime, a situation in which you might have little time or opportunity to learn about the person you’re approaching. All you might see at first glance is color. But once Derbyshire opens that loophole, he extends it to other situations such as the DMV, where you have more of an opportunity to observe and converse instead of guessing. What’s striking in Derbyshire’s advice is his constant emphasis on not learning anything. “Avoid concentrations of blacks not all known to you personally.” “Stay out of heavily black neighborhoods.” “Do not attend events likely to draw a lot of blacks.” Those are great ways to avoid getting to know black people. “If accosted by a strange black in the street, smile and say something polite but keep moving.” Don’t even stop to help someone “in apparent distress,” since you might get drawn into a conversation.

The only context in which Derbyshire recommends investigation is politics: “Before voting for a black politician, scrutinize his/her character much more carefully than you would a white.” But this, too, comes across as a license for laziness: It’s OK to relax your vigilance, as long as the candidate has the right pigment.


To me, the most telling passage in Derbyshire’s talk is this weird observation: “There are, for example, no black Fields Medal winners.” Derbyshire calls this fact “civilizationally consequential.”

Really? If you follow Derbyshire’s link, you’ll find that the Fields Medal is awarded for “outstanding discoveries in mathematics." It’s been given to 10 people in this century and to another dozen or so since the first cohort of post-desegregation students reached the age at which top-level math achievement could be assessed. That’s an absurdly small sample on which to base any claim about the mathematical ability of a minority population. (If every Fields Medal were awarded to an American, on a proportional basis you’d expect one black recipient in this century rather than zero.) The list of Fields Medal winners tells you nothing about blacks. But it tells you a lot about Derbyshire. It tells you he’s a math nerd who substitutes statistical intelligence for social intelligence. He recommends group calculations instead of taking the trouble to learn about the person standing in front of you.

Racial assumptions based on aggregate impressions or data can be overcome in particular cases. Look at the Washington Post survey that came out yesterday. In 2004, Americans considered George W. Bush more likable than John Kerry by a margin of 44 to 36 percent. Today, Americans consider Barack Obama more likable than Mitt Romney by a margin of 64 to 26 percent. When whites are asked which candidate “seems like the more friendly and likable person,” 30 percent choose Romney. Sixty percent choose Obama. That’s what can happen when you get to know someone, even at a distance.

None of this depends on refuting Derbyshire’s data. You can believe in group differences in performance (by race, sex, religion, or any other category) on any measure, including intelligence. You can argue that such differences are partially heritable, as long as you’re clear that heritability patterns are ultimately genetic, not racial or ethnic. I’ve defended such arguments before. Egalitarian fundamentalism—the idea that the right to be treated as an individual depends on the strict equality of group averages—is a dangerous mistake.

But if you’re going to present evidence for aggregate differences, you have to tell the rest of the story. You have to acknowledge socioeconomic status, stereotype threat, and other factors that can affect performance. And you have to remind people that drawing inferences about anyone based on race, sex, religion, or any other crude category is a lousy substitute for inspecting or interacting with that individual. If you tell people to protect themselves by avoiding interaction with the person they’re judging, you’re not just rationalizing racism. You’re perpetuating it.

“If you are white or Asian and have kids, you owe it to them to give them some version of the talk,” Derbyshire writes at the end of his piece. “It will save them a lot of time and trouble spent figuring things out for themselves.” And that, in a nutshell, is his error. If your kids don’t take the trouble to learn more about the people they meet, they’ll never outgrow your prejudices.