Should Harvard start admitting kids at random?

November 28, 2012

(This article was originally published at Statistical Modeling, Causal Inference, and Social Science, and syndicated at StatsBlogs.)

Tyler Cowen links to an article where Ron Unz provides evidence that Jews are way overrepresented at Ivy League colleges, with Asians-Americans and non-Jewish whites correspondingly underrepresented. Unz attributes this to bias and pressure in the admissions office and recommends that, instead, top colleges should switch to a system based purely academic credentials (he never clearly defines these, but I assume he’s talking about high school grades, SAT scores, and prizes in recognized academic competitions). He recommends that Harvard, for example, get rid of preferences for athletes, musicians, and rich people, and instead reserve one-fifth of their slots based on pure academic merit and with the remaining four-fifth “being randomly selected from the 30,000 or so American applicants considered able to reasonably perform at the school’s required academic level and thereby benefit from a Harvard education.”

A lot would depend on where that lower threshold is set. As a teacher at Harvard, I have a selfish desire to teach the very best students (or that subset of these who would be interested in taking a statistics class). On the other hand, if the threshold is lower, there’s less of an attraction to teaching at Harvard; maybe that’s part of the point. Unz discusses the conflicting goals of giving an elite education to the elite, as compared to mixing students from diverse backgrounds. He argues that the current system is so screwed up that it would possible to improve on both dimensions (although at a cost to the colleges’ sports teams).

One thing that Unz does not discuss is the possibility of a college changing its composition midstream, by kicking out some students who can’t hack it and then admitting an equivalent number of transfer students to take their place. I’ve always liked this idea, and it seems like more of a possibility if you start admitting students at random.

It’s hard to imagine just about any part of Unz’s plan happening (shutting out the children of rich alumni is a non-starter; reducing the number of Jewish admits by 75% would make a lot of people upset; even something as seemingly innocuous as freeing up some spaces by reducing the preference to athletes always raises controversy when it comes up). In short, the status quo wins. Ivies are doing pretty well right now (maybe not in 10 years, who knows, but they’re doing fine at the moment) so who’s gonna want to poke the hornet’s nest? You might imagine that some non-elite college could try Unz’s system, but the trouble there is that such colleges are trying so hard to get OK enough applicants who can pay the bills, that it’s not clear they have much room to maneuver. The more equal representation that Unz might be easier to achieve at the back end, by having various institutions refuse to hire so many Ivy graduates or to fund Ivy projects. If it were harder for Harvard students to get a job—just by virtue of being Harvard graduates—that would change things pretty fast.

Some of Unz’s arguments work against each other. For example, he brings up William F. Buckley’s quote “that he would rather entrust the government of the United States to the first 400 names listed in the Boston telephone directory than to the faculty of Harvard,” but Buckley first came to fame by promoting the idea that alumni should take more control of Yale. In which case I don’t think these alumni would be so happy to give up their kids’ slots and submit to a lottery.

But maybe I’m missing something here. In any case, this is all interesting to think about as a comparison point.

Please comment on the article here: Statistical Modeling, Causal Inference, and Social Science

Tags: ,