Over the next few days, I’m going to be summarizing key findings from my article on law school racial preferences, and discussing the findings with readers. Today is:
Entry One: The General Operation of Racial Preferences
When law schools talk about race-based admissions preferences – something they generally discuss as little as possible — they make three claims: (a) the preferences are small and not automatic, (b) race is one of a myriad of factors taken into account to create a diverse class, and (c) everyone admitted is fully qualified to do well at the school. These were the central messages advanced by the University of Michigan Law School in their defense of affirmative action before the Supreme Court. I found in my research that all three claims were substantially untrue, both for Michigan and for law schools generally. More interestingly, I found that each law school follows such a similar pattern that a powerful “cascade” effect sets in, creating interesting collective action problems for the system as a whole and for any school that wants to approach admissions less mechanically.
The problem every university faces is the gap in test scores and grades between whites and Asians on the one hand, and blacks and Hispanics on the other. On a 1000-point academic scale I use throughout my paper, the median black/white gap among law school applicants was about 170 points in the early 1990s and is about 135 points today. As you may recall, the Supreme Court issued two opinions in last year’s affirmative action case. In Gratz, the Court found that undergraduate college at the University of Michigan violated the constitution by awarding a fixed number of points to black applications. The Law School’s admissions passed muster, according to Justice O’Connor’s decisive opinion, because race there was part of a complex individualized assessment of each applicant – the opposite of a mechanical award of points.
Many of O’Connor’s colleagues hinted that she was elevating form over substance. With good reason. Using logistic regression and other techniques, I estimated the weight given to race and to academic numbers by the College and the Law School. The Law School’s admissions were more dominated by numbers, and the implicit “boost” given black applicants was larger (and as mechanical) as the College’s system. The only substantive difference between the two schools is that the College took more account of factors like socioeconomic background, writing samples, and extracurricular activities – differences that should have made it more constitutional in O’Connor’s eyes, not less. What apparently saved the law school was the way they talked about their admissions system, and perhaps the plaintiff’s failure to adequately demonstrate its actual workings.
The Michigan law school admissions cycles litigated in Grutter (mostly the 1995 through 1999 cycles) are highly representative of practices at law schools nationally. I gathered data from seven other public law schools through FOIA requests, and analyzed another database that has data on 27,000 law students from the Class of 1994. At every law school, at least 80% of admissions decisions (and usually more like 90%) could be predicted by knowing the LSAT, undergraduate GPA, and race of the applicant. Nearly every law school used the same metric for white and black students, but either added points to eliminate the black-white gap in credentials or simply segregated admissions files by race. Nearly every school admitted black and white applicants at rates that were statistically indistinguishable. And in the 1990s, a virtually identical 170-point gap could be found between the credentials of the median white matriculant and the median black – a gap that reached from the most elite schools to the smallest regional schools (ironically, only the historically black law schools were exempt from this pattern).
Herein lies the collective action problem. The preferences awarded by the top tier law schools absorb all the black applicants that would be admitted, in a race-blind system, to second-tier schools. These schools must therefore choose between having essentially no black students or duplicating the types of preferences pursued in the top-tier. Nearly all the second-tier schools choose the latter course, thus putting third-tier schools into the same bind, and so on. The net effect of this system is to move nearly all blacks up a tier (or two) in the law school hierarchy, thus placing nearly all blacks at an enormous academic disadvantage in the schools they attend. The only net addition of blacks to the system comes in the lowest-tier schools, and the black students they admit have such marginal academic credentials that they face long odds against every becoming attorneys.
Tomorrow: The effects of preferences on grades, graduation and the bar
Comments are closed.