Matthew Yglesias comes out in favor of gender and socio-economic profiling:
Now clearly if you rely on stereotypes to predict other people’s behavior, you’re going to wind up with a lot of bad predictions. But on the other hand, the number of situations in which it would be useful to have some idea of how some other person is going to behave are enormous. Not only do these situations arise all the time, but they normally arise in circumstances where the other person in question is someone you know almost nothing about. One of the few things you can accurately and easily assess about the sort of total strangers with whom people interact constantly in modern urban settings is the other person’s gender. Since gender (along with a rougher estimate of the person’s socioeconomic status and ethnicity) is probably all you’re going to have to go on, you’d have to be pretty foolish not to engage in a little stereotyping.
It may not work very well, but it’ll work better than the alternative of doing nothing. Of course, once you get to know someone you can move beyond this sort of crude analysis, but it’s been a long time since most Americans interacted only with people they know reasonably well on a daily basis.
This is, of course, absolutely correct. As I note in his comments section, if you’re going to approach every new situation as a tabula rasa, there’s not much advantage not being an idiot. Or a three-year-old.
And while I’m being mildly facetious with the profiling jab, there is some element of truth in it as well. As a matter of experience, most of us walking down a city street at night react differently to strangers based on their age, gender, race, manner of dress, and body language. While it’s certainly possible that the little old lady is wielding a knife that she intends to jab into me so that she can use the contents of my wallet to support her crack habit, my anxiety level is going to be much lower than it would be upon encountering three teenaged boys wearing bandanas and matching jackets. Strangely, patrolmen who spend far more time wandering the streets of the city at nighttime than I do operate the same way, albeit upon more nuanced and developed instincts.
Our pre-conceived notions are often proven wrong, but one has to start somewhere. Pretending otherwise is just silly.
Update (1745): Matthew has added a postscript:
Nelson Lund’s very convincing “Conservative Case Against Racial Profiling” (via Volokh) which is specifically dedicated to the proposition that this sort of rational stereotyping in one’s personal or business affairs does not justify the adoption of profiling as a government policy.
I don’t read Lund as saying anything particularly novel. Indeed, he says in 15 pages essentially what I said in the comment above. Clearly, a policy of indiscriminately harrassing black or Arab individuals would be incredibly shoddy police work, aside from any moral and fairness objections. But, just as clearly, going to idiotic extremes in the other direction–pulling over little old ladies for the “random” airport search is not a good idea either. Lund’s just saying–I think–that taking race into account as one variable among many (ironically, like post-Bakke affirmative action programs) may be useful.
Lund also cites the DC sniper case, in which officers were looking for white guys in a white van when the perps were (allegedly) two black men in a blue van. I honestly have no idea how to solve this problem. If the eyewitnesses are all independently telling you it’s a white guy in a white van, I’m afraid that’s where I’d be directing my resources. Pulling over everybody driving anything is just too cumbersome.