Diversity and meritocracy in legal education: a critical evaluation of Linda F. Wightmans's "the threat to diversity in legal education." (response to Linda F. Wightman, New York University Law Review, vol. 72, p. 1, 1997).

• Author: Thernstrom, Stephan

The debate over racial preferences in admissions to higher education has grown more heated than ever, now that a vote by the Regents of the University of California and the passage of Proposition 209 have compelled the university to admit students on a color-blind basis.(1) Those who have drawn far-reaching policy conclusions from the California numbers have neglected the elemental point that a one year experiment in one state cannot yield definitive results. Thus it is fortunate that the results of a massive investigation of the impact of affirmative action in law schools across the nation have recently become available. Linda F. Wightman's article, "The Threat to Diversity in Legal Education: An Empirical Analysis of the Consequences of Abandoning Race as a Factor in Law School Admissions Decisions," published in the April, 1997 issue of the New York University Law Review, has received considerable publicity. It includes remarkably rich and compelling evidence about the effects of racially preferential policies in law school admissions.(2) Unfortunately, Professor Wightman does not herself draw the most important conclusions that her own data support. Indeed, her interpretations run directly counter to her own evidence at many points. A thorough critical evaluation of her analysis will help to clarify the issues in a vital current dispute about legal education.

For a highly technical piece appearing in a law review, Professor Wightman's study attracted a surprising amount of national attention, doubtless because of its apparent relevance to the current heated debate over racial preferences in higher education. Her study purported to show both that the number of minority law school students would drop drastically in the absence of racial preferences in admissions, and that students admitted under affirmative action double standards were nonetheless just as successful as anyone else after they began their legal studies. Their inferior academic qualifications at the time of admission did not hinder their progress through law school or their entry into the profession.

These conclusions were welcome news to many writers, particularly the highly optimistic findings about the alleged academic success of students admitted as a result of affirmative action preferences. According to Time, the study established that African American and other minority students admitted to law schools due to preferential policies "had graduation and bar-exam pass rates similar to whites.'"(3) Newsweek columnist Ellis Cose declared that Professor Wightman had found "no real difference ... between those minorities who would have been admitted without affirmative action and those admitted because of it."(4) NPR's "All Things Considered" featured an interview with the author that stressed the same conclusions.(5) Professor Wightman, the public was led to believe, had established that affirmative action admissions policies in law schools had no significant costs. Although the beneficiaries of racial and ethnic preferences were admitted with distinctly weaker academic qualifications than other students, their initial handicaps apparently vanished in the course of their graduate education.

Only one criticism of the Wightman study has appeared in print thus far, a brief piece by Gail Heriot in the Weekly Standard.(6) An article that has attracted so much attention but so little critical evaluation clearly calls for more detailed scrutiny.

The research for Professor Wightman's article was conducted while she was Vice President for Testing, Operations, and Research for the Law School Admissions Council. The resources of the council allowed her to compile a formidable data base. She had information about more than 90,000 applicants to ABA-approved law schools for the academic year 1990-1991. A second sample, a subset of the first, consisted of some 27,000 of those applicants who actually enrolled in law school in the Fall of 1991. They represented about 70 percent of the entire entering class that year, These students were followed over time, making it possible to determine graduation rates and pass rates on the bar exams. This treasure trove of information could have done a great deal to advance our understanding of the effects of preferential policies. Unfortunately, the author's analysis is badly flawed. Much of the evidence she gathered contradicts the conclusions she draws from it, and key questions that could and should have been answered were never asked.

I

Professor Wightman's first set of findings concerns who was admitted to law school in the 1990-91 academic year. The central question here is how different the composition of the class of accepted applicants would have been if all admissions decisions had been made on the basis of college grades and LSAT scores alone. Some 6.8 percent of all the students who were accepted to law school during the 1990-91 admissions cycle were African Americans.(7) How much lower would the figure have been had race not been a factor in admissions decisions?

The author employs two different statistical models to estimate the answer. If admissions had been determined by GPA and LSATs alone, she calculates that only one fifth of the 3,435 black students who were accepted at a school to which they applied would have won a place at that school.(8) Four out of five-exactly 80 percent--had grades and test scores that would have put them in the reject pile if they had been white, according to this estimate; only one in five was admissible on straight academic grounds. The proportion who were admissible on the basis of grades and scores alone was lower still among the students who actually enrolled in law school in the Fall of 1991--only 8.9 percent!(90

This estimate does not answer the question of how many would have been able to secure admission to some law school. If racial preferences in admission were to end throughout the land, it is possible that black students would then apply to less selective and less prestigious schools that they could afford to spurn when their race gave them a big edge in the competition.(10) Professor Wightman employs a second model to estimate how much shifting towards less selective schools would have taken place in the absence of preferences. The conclusion she draws from this model is that overall black enrollment in law school would have dropped not by 80 percent but by at least half--that instead of being 6.8 percent of the pool of accepted students, African Americans would have been just 3.4 percent.(11)

Furthermore, Professor Wightman argues plausibly, this estimated 50 percent drop errs on the low side, because many of the lower-echelon law schools we might expect African Americans to attend in the absence of preferences are relatively expensive private institutions that many could not afford to attend.

II

The author glosses over a serious methodological problem with both models. Her calculations equate color-blind admissions with admissions strictly by the numbers--by GPAs and LSAT scores. But her own figures reveal that grades and test scores are far from the only determinant of admissions, even for whites. Almost 11 percent of the whites whose grades and test scores should have been good enough to have won them a spot in a law school class, it turns out, did not in fact receive an acceptance letter.(12) Something else in their records apparently offset their academic strengths. Conversely, fully 15 percent of the whites who were admitted to law school that year did not have strong enough academic records to be accepted on the basis of their grades and test scores alone.(13)

Considerations other than the two readily quantifiable measures of academic achievement thus often influence law school admissions decisions. For example, state institutions typically give preferences to applicants who are residents. Many schools give a plus for applicants from disadvantaged family backgrounds, a practice that a ban on racial and ethnic preferences would not call into question. Some candidates have letters of reference that attest to their character and personality in glowing terms; others have only letters that praise with faint damns at best. A record of unusual community service or having a parent who is an alumnus of the school can make a difference.

Professor Wightman concedes that her models are quite imperfect as predictors of the admission of whites.(14) But the only conclusion she draws from this limitation is that white applicants who are turned down by law schools despite having good grades and scores may be wrong to blame affirmative action preferences given to minorities; it is more probable that they lost out to whites with lower GPAs and LSATs.(15)

This is an interesting point, but there is a more important one that the author fails to notice. Her estimates of precipitous drops in black enrollment under color-blind admissions policies mistakenly assume that eliminating race as a factor amounts to eliminating all other factors except grades and LSAT scores. Surely the additional factors that were responsible for the admission of 15 percent of the successful whites--family economic background, for instance--would also have brought about the admission of some African Americans who would not have made it in on the basis of the two objective measures alone. After grades and scores have been taken into account, that is, all the remaining residual influences that explain admissions decisions cannot be classified as simply "race." That is what Professor Wightman does for the non-whites, though obviously not the whites, in her sample. The estimates of black enrollments under color-blind policies that can be derived from both of Professor Wightman's models thus have a downward bias. They are not true estimates of the effects of color-blind admissions--only of admissions based purely upon grades and LSATs.

Even with ample allowance for this methodological flaw...