The Black Student Mismatch Myth in Legal Education: The Systemic Flaws in Richard Sander's Affirmative Action StudyCheryl I. Harris and William C. Kidder
Editor's Note: Richard Sander's recent paper contends that racial preferences in admissions to law school will produce fewer rather than more practicing African-American attorneys. Sander also leaves the impression that black students at our major law schools are failing in ordinate numbers. Here two scholars dissect Sander's numbers and come up with entirely different results.
AS WE CONCLUDE the celebration of the fiftieth anniversary of Brown v. Board of Education, it is clear that the work of desegregation in American education begun in that landmark case is far from complete. Despite the historic 2003 ruling in Grutter v. Bollinger, in which the Supreme Court approved the use of race as a legitimate means of achieving the critical state interest in diversity, actual racial diversity in higher education is proving to be elusive and fragile. This year's freshman class at the University of Michigan included the fewest number of African Americans in 15 years; the story is even worse at selective universities prohibited from using affirmative action. At the University of California at Los Angeles, only 2.3 percent of this year's freshman admission offers went to African Americans. At the University of California at Berkeley, only 2.5 percent of admitted students are black and at the San Diego campus of the University of California, only 1.8 percent of all students accepted for admission are black.
Given these distressing trends, one would hope that we could turn our collective energy to addressing the appalling inequities in educational access. Instead, recent media attention has focused on whether to further restrict or even eliminate affirmative action in America's law schools. This is the question posed in a lengthy study by Richard H. Sander, a law professor at UCLA, in the November 2004 issue of the Stanford Law Review. Sander argues that affirmative action "mismatches" black students at law schools where their lower entering LSAT scores and undergraduate grades place them at an academic disadvantage producing lower grades and graduation and bar passage rates for African-American students. Most controversial is Sander's attempt to prove with empirical evidence that if affirmative action ended, African Americans would "cascade" to lower-ranked law schools where they would perform dramatically better, causing a net increase of 8 percent in the number of new black lawyers.
These arguments are not new. Many will recall earlier versions of the "mismatch" thesis advanced by conservatives including Stephan and Abigail Thernstrom, Ward Connerly, Walter Williams, Gail Heriot, and Thomas Sowell with respect to affirmative action and college graduation rates. Time and time again the basic premises of these claims have been refuted. The Thernstroms, the reader may recall, predicted that among the benefits of California's Proposition 209 was the "redistribution" of African Americans from UCLA and UC Berkeley to less selective campuses like UC Riverside that will, in the end, increase the number of blacks graduating from the UC system overall. Yet there is little if any indication that these benefits have materialized or ever will. What is clear is that as black enrollments decrease at the University of California's most selective campuses, the educational pipeline to law school will show the racial divide produced by "cascading." For instance, last year UCLA and Berkeley produced the most and third-most applicants to law school in the country. The University of California at Riverside, the 114th-ranked feeder institution, sent only one seventh as many applicants to law school as UCLA.
Since Sander's study replicates many of the claims that have been made before, it is difficult to understand the level of interest in his analysis. Perhaps Sander's article is attracting attention because, unlike many vocal affirmative action critics, he is a self-described liberal. Ultimately, however, what matters is not biography but whether or not Sander's calculations and conclusions are sound. Regrettably, Sander significantly underestimates the harms of ending affirmative action, and seriously overestimates the benefits of ending affirmative action. Even his own data do not support the mismatch hypothesis.
Law School Admissions (and Omissions)
Does Sander's claim about the end of affirmative action causing an 8 percent increase in the number of black lawyers sound implausible? That's because it is. In fact, a decline of 25 percent to 30 percent in the number of African-American attorneys is a much more realistic outcome. Sander's estimate is an accumulation of several significant errors, oversights, and implausible assumptions.
For example, using 2001 data, Sander claims that nationwide, ending affirmative action would only curtail African-American law school enrollments by 14 percent. However, the 2001 admissions cycle upon which Sander relies had an atypically high ratio of black to white applicants; the soft job market for college graduates has since driven many more whites to apply to law school. Thus, calculating the same "grid model" estimate that Sander relies upon, but with 2003 and 2004 national data, reveals that ending affirmative action would slash African-American law school admissions by 24 percent and 33 percent, respectively.
Moreover, several unrealistic assumptions in Sander's model render his 14 percent estimate far too optimistic. He assumes that applications from African Americans would not decline were affirmative action to end. In reality, applications from black students consistently dropped when the public law schools in California, Texas, and Washington banned affirmative action in the late 1990s.
Likewise, if without affirmative action all the highly competitive law schools had only 1 to 2 percent African Americans (a point Sander concedes), it is unrealistic to assume that all of the accomplished African Americans now at law schools like Michigan and Northwestern would go to law school at all if their best option would be to attend a lower-ranked law school like Arizona State or Marquette, and to do so, more likely than not, in an environment of social isolation. It is far more likely that many of the African Americans now at elite law schools with a "critical mass" (African Americans are typically over 10 percent of the student body at schools like Penn, USC, and Georgetown; at top-30 schools the average is 7 to 8 percent) would understandably prefer graduate school or the job market if law schools suddenly ended affirmative action.
Sander's model also combines admission decisions from 185 law schools into a single pool. The underlying assumption is that law schools are fungible in terms of attractiveness to black applicants. Yet dozens and dozens of the lower-ranked law schools where Sander assumes African Americans would "cascade" are in regions of the country with small African-American populations and law school enrollments, including the Great Plains, Southwest, Pacific Northwest, Rocky Mountains, and rural New England. There is no reason to believe that an African-American candidate from New Jersey would attend the University of Montana Law School when she never applied there in real life. This is because black law students, like their white peers, prefer schools where they will make contacts with peers and alums in the area they intend to practice. While the most competitive schools generate a nationally recognized credential, this is simply not true of lower-ranked law schools.
Sander's model also fails to consider other local policies and preferences that would cut against the presumption of a seamless cascade. For example, in 2004 African Americans were only 2.7 percent of the applicants to the University of New Mexico School of Law, a school with a solid commitment to affirmative action. Even if there were a surprising surge in black applicants to New Mexico as students "cascaded" from more prestigious schools, potential enrollment gains for blacks would be wiped away by an admission policy requiring 85 percent New Mexico residents (not uncommon at public law schools). In short, Sander's "cascade" metaphor is inapt because assuming that nearly all of the African Americans now enrolled in law school will gently cascade to lower-ranked schools without affirmative action is like assuming that water will naturally flow uphill.
LSAT Is Not Destiny
A key premise underlying Sander's mismatch theory is that entering credentials (LSAT scores and college grades) are powerful determinants of bar passage. Sander argues at great length that LSAT and GPA explain "well over 35 percent" of the variation in bar performance. Yet his claim does not come from the Law School Admission Council's Bar Passage Study, of 27,000 students in 163 law schools a dataset he relies on throughout the article. Rather, Sander shifts to a study of the California bar exam one which is arguably unrepresentative because California has the most difficult bar exam in the country, and California also allows graduates of unaccredited schools to sit for the exam.
Tellingly, published reports of the Bar Passage Study indicate that LSAT and GPA only explain about a tenth of the variation in bar pass/fail status. Sander's assertion that there is a strong correlation between entry credentials and bar passage rates is simply not supported by the data.
Sander's estimates also rely on the unsound assumption that eliminating affirmative action will erase all LSAT and college grade differences between blacks and whites attending the same law schools. Yet a significant black-white LSAT gap will inevitably carry over from the applicant pool to the entering class under "race-neutral" admission practices given the distribution of scores. Indeed, as Bowen and Bok observe in The Shape of the River, "The only way to create a class in which black and white students had the same average [test] scores would be to discriminate against black candidates." Ironically then, the more seriously one takes Sander's claim about the importance of the LSAT, the more skeptical one must regard his optimistic estimates of the number of African Americans who would become lawyers without affirmative action.
Legal Education: Inside the Black Box
At the heart of Sander's argument is the claim that affirmative action causes the disparities in law school GPA, attrition, and bar passage because under affirmative action, blacks are mismatched to institutions where they cannot compete. Sander views the performance of black students as the direct result of the deficiencies they bring. Thus, he asserts that the performance difference between black and white students in law school "seems to be simply a function of disparate entering credentialsÉ" Once the disparity in entry credentials is eliminated by ending affirmative action, Sander concludes that group differences in academic performance within each law school would then disappear.
The evidence cited in support of Sander's claim is not the more robust Bar Passage Study, but his own study of 20 law schools that shows that black and white students with the same entry credentials earn the same grades in the first semester of law school. Sander's finding, however, is contradicted by the Bar Passage Study as well as the consistent findings of researchers in the field showing that black students with the same entry credentials as their white classmates within the same law school tier earn lower grades. This leads to an inference that directly contradicts Sander's thesis: If black and white students with the same entry credentials do not earn the same grades, then something else besides credentials must explain the outcome. This phenomenon, which researchers call "underperformance," has been the subject of study by Stanford psychologist Claude Steele, who has shown in experiments involving highly qualified and motivated black students that the fear of confirming negative racial stereotypes actually suppresses academic performance. When the threat is removed, academic performance rises. Differences in academic performance are not fully explained by "stereotype threat," but it is clear that differences in entry credentials do not fully explain what the weight of the research shows.
In fact, the institutional environment of law school has a critical impact on law students and their relative performance in school. An article just published in Law & Social Inquiry (the American Bar Foundation's faculty-refereed journal) utilizing the same Bar Passage Study data as Sander demonstrates that the educational process at American law schools actually exacerbates the entering educational gaps of minority and other atypical law students, including not only African Americans, but also Latino students, Asian American students, and those who begin law school at 30 or older. These students get lower grades than their entry numbers would predict. Hostile institutional climates, students' sense of belonging or isolation, as well as teachers' expectations have a significant effect on academic performance. These results are not surprising, least of all to the scores of law students of color at UCLA, Boalt, Davis, and Hastings who filed a brief in Grutter describing the added academic pressures and social stigmatization they encountered after Proposition 209. However, Sander remains wedded to a paradigm that treats law school as a black box.
Affirmative Action Does Not
In reality, performance differences in law school are the product of a complex set of factors that can hardly be reduced to LSAT scores. That is why across all the accredited law schools studied by the Law School Admission Council the LSAT typically explains only 16 percent of the variation in first-year law school grades. Moreover, even the combination of LSAT scores with all of the other factors in Sander's dataset still only explains 19 percent of differences in grades.
The danger is that Sander's analysis, by treating statistical significance levels as if they indicated how strongly factors correlate with each other, may mislead the statistically unsophisticated. At each stage, the variables that capture Sander's attention only predict a small fraction of the variation in the outcomes of interest. Other variables are ignored or minimized. Ultimately, he aggregates each tenuously supported conclusion into the next set of claims, which are, in turn, added to the next set of even more weakly supported claims. Thus, at the end of the day there is quite a gulf separating his data and the conclusions he draws.
This gulf is evident from the fact that Sander's mismatch theory is largely refuted by the most straightforward test of the Bar Passage Study data. A forthcoming critique of Sander in the May 2005 issue of the Stanford Law Review reports that African-American students who attended various tiers of schools but had the same LSAT scores and college grades. In 11 of 13 comparisons, attending law schools in the higher-ranked tier was associated with higher African-American graduation rates. And in eight of 13 comparisons, attending higher-ranked schools was associated with higher African-American bar pass rates.
Resegregating Our Nation's Leadership
The leadership echelons of the legal profession are dominated by graduates of the top 20 or so elite law schools. Ending affirmative action at elite law schools would have profound long-range consequences for America. Take, for example, federal appellate and Supreme Court law clerks, often the professors, judges, and law firm partners of tomorrow. Over half of these prestigious clerkships in 2002 went to graduates of the top 13 schools, less than half went to graduates of the other 170 ABA law schools. Likewise, a majority of all current African-American law professors graduated from the top 10 or so law schools. David Wilkins and Mitu Gulati also found that 77 percent of African-American partners in major corporate law firms went to elite schools, 47 percent from Yale and Harvard alone.
In Grutter, the Supreme Court was wisely persuaded by similar statistics when it declared, "In order to cultivate a set of leaders with legitimacy in the eyes of the citizenry, it is necessary that the path to leadership be visibly open to talented and qualified individuals of every race and ethnicity."
According to U.S. Census Bureau statistics, in 1970 there were only 3,845 African-American lawyers in the U.S. By 2000 there were over 39,000 black lawyers and judges in the U.S. a tenfold increase and a sea change in American society that simply would not have been possible without affirmative action.
Cheryl I. Harris is a professor at the UCLA School of Law. William C. Kidder is a researcher at the Equal Justice Society. A draft of his Sander critique for the Stanford Law Review is available at: http://www.equaljusticesociety.org/research.html.