AFFIRMATIVE ACTION….The Supreme Court is hearing the University of Michigan affirmative action cases today, and to prepare for the event the LA Times ran a couple of op-eds on the subject on Monday. After reading them I had an idea for a post I wanted to write, but it would have been a real munge of disconnected ideas on reparations, diversity, the meaning of merit, etc. etc.

However, I was saved from writing it — and you from reading it — by the discovery of a genuinely interesting report from The Century Foundation about alternatives to affirmative action. The authors, who pulled together a ton of longitudinal data and compared it to polling data on public support for various forms of racial preferences, came up with some compelling support for an alternative to the typical race-based affirmative action programs used today in higher education.

The problem is basically this: standard issue affirmative action is based simply on ethnicity. If you’re black or Hispanic, for example, you get extra credit when you apply for university admission. Although this achieves the admirable goal of helping disadvantaged minorities, the problems are multiple: (a) the public doesn’t support it, (b) it violates the principle that we should try to be a race neutral society, and (c) most of the benefit goes to minority students from well off families. Plus, as Sam Heldman says, the Supreme Court is likely to find it unconstitutional.

So what to do? One popular idea is to guarantee university admission to the top 10% of every high school class. The authors investigated this plan, however, and concluded that it has a problem: too many unqualified students are admitted and the dropout rate is high. The point, after all, is not to simply admit students from underrepresented classes, it’s to see them through to graduation. If graduation rates are low, then the policy is a failure.

Instead, they recommend another approach that’s gotten some attention recently: preferences for students from low income families. I’ve been skeptical of this approach in the past simply because it’s inefficient, requiring large interventions in order to obtain minimal amounts of racial diversity. In other words, too little racial bang for the preferential buck. But the Century Foundation report concludes otherwise: giving preference to low income students would produce almost as much racial diversity as we have today. Thus, in order to maintain today’s standards of racial diversity, only a modest additional amount of traditional affirmative action would be required.

And it turns out that this approach is actually fairer than pure merit-based approaches since it’s more likely to find the truly talented kids. Think of it this way: one kid has been to weight training all his life and can bench press 200 pounds. The second kid has never had any training and can bench 180 pounds. If you provide them both with four years of weight training, which one is mostly likely to be able to bench 250 pounds when the program is over? If you’re smart, you’ll put your money on the second kid even though his initial “test score” is lower. Despite the difference in scores, he really is the more naturally talented of the two.

The conclusion of the Century Foundation report is good news on several fronts. First, according to polling data the public supports both the inherent fairness of preferences for low income students as well as the idea of modest amounts of race-based affirmative action. It’s only substantial racial preferences that most people object to.

Second, it promotes economic diversity as well as racial diversity. This by itself is a positive development.

And third, it works. In the computerized simulations run by the authors, students chosen this way have graduation rates that are actually higher than those chosen purely by test scores and high school GPAs. By analogy with the weight training example, this makes considerable sense: after all, a kid who lives in a crappy neighborhood and goes to a crappy school but still scores, say, 1100 on his SATs, is probably every bit as naturally talented as one who attended a top high school and scored 1200. Provide the first kid with a good education, and he’s likely to shine.

I like this idea a lot, and I hope their research holds up. I mostly like it because it moves the debate away from the same stale issues we’ve been discussing for decades now, and suggests a way out that is both effective, fair, and likely to be widely supported. It will be interesting to see how much attention this report gets.