Monday, May 4, 2009

Are admissions officials actively anti-Asian?

Yglesias cites "expert" (in this case, propagandist) Kevin Carey saying that "It’s clear that, given the opportunity, elite American universities are prone to implement discriminatory admissions policies that artificially limit the number of American students of Chinese, Korean, and Japanese descent." MY continues: "But the quota system appears to have merely re-emerged with Asian-Americans as the victims. And nobody talks about it, though the discriminatory mechanisms are only slightly more subtle."

There are two issues here that need to be kept distinct. The first is whether Asians and middle-class whites are losing out at the expense of minorities, legacies, broken-family cases, etc. The second is whether Asians are discriminated against relative to whites of similar socioeconomic status.

I continue to be somewhat skeptical that the second kind of discrimination is rampant. The data from SAT scores is not well-controlled for socio-economic effects, nor is it controlled for whether students were legacies or athletes. If there is evidence that it's appreciably harder for middle-class Asian students to get into college than for middle-class, non-legacy, non-athletic, whites, I haven't seen it. From anecdotal experience -- the people I knew at college, for instance -- I would consider the claim fairly implausible.

A universal characteristic of quota systems is that they lead to a large disparity in abilities between the limited group -- Asians, in this case -- and the non-limited group, especially at second-tier institutions. If Harvard got the very best of everybody, you wouldn't see this disparity there even if they had a quota system. However, as you moved down the pyramid, you'd see the quality of Asian students exceeding that of white students; the distributions would completely separate at some point, as they do for international students at second-tier grad schools. I would have expected Amherst (roughly 9-12 on most people's lists) to show signs of this disparity, but I never saw any. There were no bimodal distributions with an "Asian" peak; neither the summa list nor the PBK list was abnormally heavy on Asian names; etc. There were, on the other hand, a disproportionate number of international students -- a group that was openly discriminated against in the admissions process -- graduating summa.

It might be that one has to move further down the US News list to see this effect clearly, or it might be that the apparent bias against Asians is an artefact of Asians belonging disproportionately to categories that gain nothing from affirmative action (relatively few athletes, legacies, etc.). If the former, this provides an upper bound on how serious the effect is (one shouldn't lose sleep over Asians who should've gotten into Dartmouth ending up at Cornell); if the latter, that brings us to the general question of whether affirmative action is justifiable. I think it is, in principle, though most institutions do a mediocre job of implementing it.

(Should note, btw, that this is all about Asian-Americans. Admissions for foreign students are an entirely different story, given that colleges are unfamiliar with most of the high schools, grading systems, etc. involved.)


Westcott said...

I imagine if they did this kind of study at Amherst, my data point would have to be thrown out as an outlier. My incompetence clearly transcended my race.

- do

Zed said...

you could also be thrown out (cf. last paragraph) for not being an american.

Not-a- Generic-User-ID said...

I suspect that a lot of the discrimination against Asians is concentration related. Colleges won't take more than N kids from the same high school or X kids from the same neighborhood.

So someone who went to an Asian heavy high school will see that a lot of their friends are rejected from schools where they are solidly more qualified than the top quartile of accepted students. This might lead them to think there is a quota against asians.

This hypothesis depends on Asians being concentrated rather than distributed evenly throughout the population, but I think that is the case.

Its also worth noting that Amherst may be bad data point to look for disparity at since in high school I observed asian applications skewed towards universities over colleges.

Zed said...

I agree about the geography. (Wonder if Asians who went to say Scarsdale feel discriminated against relative to whites from Scarsdale...) As for Amherst being a bad data point -- maybe, but I think the number of premeds and the strength of the premed program would counteract that effect.