The diversity defense

The University of Michigan’s excuse for racial discrimination

By Shikha Dalmia

Weekly Standard, March 26, 2001

One of the first acts of Lee Bollinger upon taking office as president of the University of Michigan, Ann Arbor, four years ago was to vow publicly to mount the most vigorous and “comprehensive” legal defense of affirmative action or racial preferences yet. What prompted this show of bravado from a man otherwise known for his soft-spoken — almost meek — demeanor were two lawsuits by white applicants challenging the university’s undergraduate and law school admissions programs. The lawsuits, filed on behalf of the applicants by the Center for Individual Rights, a public-interest legal organization in Washington, D.C., argued that the university had unlawfully rejected them in favor of less qualified minorities.

As the lawsuits wended their way through federal court, reaching trial late last year, Bollinger demonstrated that he was not kidding: The university has already spent $ 4.5 million — a sum that would cover a full year’s tuition for about 700 in-state students — to hire one of the most prestigious law firms in the country, line up a slew of academic heavyweights as witnesses, and commission research to “scientifically” prove the benefits of affirmative action.

In the process, Bollinger has produced a case that is remarkable — both for its candor and its disingenuousness.

The university admits openly that it discriminates by race. Yet, it claims that this violates neither the Constitution nor any civil rights legislation because it serves a “compelling state interest” — diversity.

To make its case, Michigan relies upon the 1978 Supreme Court ruling in University of California Regents v. Bakke. In this ruling, Justice Powell, in his plurality opinion, allowed universities to use race as a tipping or “plus factor,” when everything else is equal, to recruit a diverse student body.

But Michigan has taken this diversity defense further than any other institution. Elite universities in Texas, Georgia, and Washington that employ racial preferences have been sued. They have defended their practices primarily by demonstrating the advantages of preferences for their intended beneficiaries. The Shape of the River, a book co-authored by former presidents of Harvard and Princeton, has been their bible. The book attempts, among other things, to document how crucial preferences have been in strengthening the black middle class.

The University of Michigan, however, uses the diversity argument to make an even more grandiose claim: It argues that racial preferences are good not just for minorities — but for all students.

This defense has secured the university a partial victory in the first round of litigation: Federal district judge Patrick J. Duggan, a conservative Democrat appointed by Reagan, ruled last December that the university’s 1995 undergraduate admissions policy, the one directly under challenge, was illegal because it operated like a quota, something that Bakke explicitly prohibited. But he also ruled that Michigan’s post-1995 policy of awarding under-represented minority applicants an automatic boost of 20 points or one grade point (when a perfect SAT score earns an applicant only 12 points) is not per se illegal. Oddly, he made this ruling despite the fact that the university itself admits that race gets no less weight now than it did in 1995. The university is awaiting judgment on whether similar favoritism by the law school toward minorities constitutes an illegal double standard. Regardless of the outcome, both these cases are likely headed for the Supreme Court.

How plausible is Michigan’s pathbreaking argument, the diversity defense? Suppose one concedes that diversity is educationally valuable because it forces students to “encounter differences rather than one’s mirror image,” as Bollinger puts it. Still, why should skin color be the primary measure of diversity?

Bollinger’s answer is that “race is educationally important for all students because understanding race in America is a powerful metaphor for crossing sensibilities of all kinds.” The assumption behind this claim, as university expert witnesses explained during the trial, is that the primary experience shaping the psyche and intellectual viewpoint of Americans is their position on the “oppressor-oppressed” divide. Few would deny that some link exists between many people’s race and their views. But the university reduces everything to race.

If diversity is as important as Bollinger says it is, why stop at racial diversity? Would not intellectual diversity, which is what truly matters, be enhanced more by using some direct measure of ideological, political, or religious beliefs? And supposing Bollinger were right that Americans cannot transcend the oppressor-oppressed divide, wouldn’t the cause of diversity be best served by recruiting ever more non-Americans — students from other countries? Arguably, a society can learn most about itself by confronting not differences that exist within it but differences it has with others. Nor would the University of Michigan have to dilute its academic standards to offer this kind of diversity in its undergraduate program.

Another factor that is as important as race, if not more so, in shaping one’s worldview is one’s age or stage in life. A mature student who has had time to accumulate interesting experiences is surely as valuable a classroom resource as a privileged minority kid.

This point is relevant, for if Michigan had taken it into account, it would have deemed Barbara Grutter, the plaintiff in the law school case, a tremendously desirable candidate: When she applied in 1996, she was a fortysomething mother of two who had married at 19. She was returning to school 18 years after getting her undergraduate degree from Michigan State University, where she had paid her way by working at clerical jobs. After graduating, she had worked at the managerial level in several companies and then had started a successful consulting business that had allowed her to work from home and attend to her children at the same time. Given Grutter’s proven ability to overcome adversity — a quality that Michigan emphasized at the trial as what makes minority candidates attractive — combined with her solid academic credentials (a 3.8 GPA and an LSAT score in the 81st percentile), her application ought to have escaped the reject pile.

Social psychologists could no doubt list a whole host of factors besides race, country of origin, and age that influence the worldview of individuals. Michigan’s decision to privilege race above all else in its student mix is based at best on a hunch. To make an empirical case for its hunch, the university has produced a scientific study. But the study, though replete with charts, correlation coefficients, and regression analyses, is tall on claims while short on scientific rigor.

Conducted by a psychology and women’s studies professor at the university, Patricia Gurin, the study supposedly shows that students in racially diverse environments demonstrate “the greatest engagement in active thinking processes, growth in intellectual engagement and motivation, and growth in intellectual and academic skills.”

Gurin, by all accounts, is a remarkable woman, both in her professional and personal life. An interim dean of Michigan’s College of Literature, Science and the Arts, she graduated Phi Beta Kappa from Northwestern University in 1954. She was married to a black Ph.D. candidate in social work for five years, something that temporarily estranged her from her father.

The Ann Arbor Observer recently described the Gurin report as “U-M’s Secret Weapon” because of the extent to which Judge Duggan relied on it in his ruling. But Tom Bray, a columnist for the Wall Street Journal and the Detroit News, notes that accepting Gurin’s research on the matter is like “accepting the research of a Firestone scientist on the causes of tread separation.” Indeed, one of her own colleagues describes her as so “fanatical” on the subject of racial preferences that he says he could have predicted her findings even before she had “discovered” them.

It would of course be extremely hard to design a study that accurately captures all the tangled aspects of a subject as complicated as affirmative action. But Gurin’s study, which has yet to be accepted for publication in a peer reviewed journal, suffers from some basic — even shocking — methodological flaws.

The study essentially has two parts. The first analyzes data complied from University of Michigan students. The major problem here is that Gurin does not measure what she says she is measuring. The obvious way of studying, for instance, the effect of classroom diversity on what Gurin calls “learning outcomes” would be to compare the academic performance of students in racially homogeneous classes with those in racially heterogeneous classes.

But this is not what Gurin does, as Howard Schwartz, a professor of organizational psychology at Oakland University, Michigan, notes. She arbitrarily defines “classroom diversity” as participation in an ethnic studies course or workshop regardless of the racial makeup of the class. She then compares the “learning outcomes” of students who have enrolled in such courses with those of students who have not. Obviously, what she is measuring is not the effect of racial diversity but the effect of a multicultural course of study. It is entirely possible that the classes themselves were monoracial.

Nor is the report’s method of measuring “learning outcomes” or “academic benefits” much better. Gurin argues that diversity makes students think “more deeply” and in “more complex ways.” But to measure this she relies not on any objective measure of student performance, such as grades or standardized test scores or, for that matter, any of the proven psychological instruments that measure cognitive complexity. Instead, Gurin developed her own “instrument.”

But this instrument, observes Patricia Hausman, a Virginia-based psychology consultant, does not even “pass the laugh test.” It asks students questions such as whether they “prefer simple rather than complex explanations to phenomena” and whether they “enjoy discussions of causes of people’s behavior” or, instead, “take people’s behavior at face value.” Students who say they prefer complex explanations and enjoy discussions of causes of behavior are supposedly complex thinkers and the others are simple thinkers. Gurin found more complex thinkers among students who had taken the ethnic studies course.

The obvious problem with Gurin’s test is that it measures not whether students can actually think in a complex way, but whether they feel they can. But even her assumption that complex explanations are better than simple ones is questionable. As Albert Himoe, a retired chemist with a doctorate from the University of Chicago, asks: What if a student takes the scientific view of the situation and believes that the simplest explanation consistent with the facts ought to be preferred? “This answer would, according to her scale,” he observes, “label you as an inactive thinker, a failure of the educational system.” One of Gurin’s retired colleagues from the English department, Leo McNamara, suspects that she devised the instrument to obtain the results she wanted. “Without any shred of malice, I have to say, Gurin’s study is a Party piece,” says McNamara, a one-time supporter of affirmative action who has a masters degree in psychology from Harvard University.

As if that weren’t enough, Gurin also commits the cardinal sin in social science: using a biased sample. She should have relied on some random method for picking and assigning students to a control and an “experimental” group. Instead, she queried a group of students who had chosen to take an ethnic studies course and compared them with another group which had chosen not to. This raises real questions, points out Richard Cutler, also a retired Michigan psychology professor, as to whether the gains in “learning outcomes” were actually an effect of the course — or an effect of the pre-existing attitudes that had prompted the students to take the course in the first place.

The second part of Gurin’s study, in which she analyzes national longitudinal data, is equally problematic. The data are compiled by the Cooperative Institutional Research Program (CIRP) and the UCLA Higher Education Research Institute on 9,316 students attending nearly 200 colleges. This is one of the most comprehensive databases on college students, offering rich possibilities for research. Gurin’s simplistic study, however, is an opportunity missed. She claims that, just as in the Michigan study, she found in the CIRP data a high correlation between institutional diversity and academic outcomes. But once again, Gurin chooses to study the educational benefit not of campus racial diversity, but of participation in an ethnic studies course. Indeed, Alexander W. Astin, another researcher, conducted a study on the same data set in 1993, using a more direct measure of campus diversity, and found “weak to no effect on academic outcomes.”

It turns out, in fact, that the Gurin report is contradicted not just by the Astin study (which Gurin at least cites, albeit misleadingly), but also by many others that she fails even to mention — an odd departure from standard academic practice. For instance, researcher Ernest Pascarella, some years ago, conducted a series of studies comparing the performance of black students in historically black colleges with that of black students at predominantly white colleges. He found no difference in the academic ability of the two groups on any standardized measure of reading comprehension, mathematics, or critical thinking — a finding that Gurin conveniently omits from her work.

More significantly, a year after the Gurin report was released, economists Harry Holzer and David Newmark concluded, in a fall 2000 article published in the Journal of Economic Literature: “There is no evidence of the positive (or negative) effects of a diverse student body on educational quality.” Their conclusion is notable because it was based on one of the most comprehensive reviews of the literature on affirmative action. Moreover, most of Holzer and Newmark’s previous work has tended to support affirmative action, especially in the workplace.

The university claims — unconvincingly — that racial diversity produces great educational benefits. But it also claims that, to produce these benefits, only modest use of racial preferences is required. This claim too is disingenuous, something eloquently demonstrated in court by Kinley Larntz, a statistics professor at the University of Minnesota.

A longtime supporter of affirmative action, Larntz analyzed the university’s law school admission data from 1995 to 2000, the latest year for which information is available. His calculations show that, averaged over these six years, the estimated relative odds of acceptance for black candidates were 234.5 times better than those for white applicants with the same GPA and LSAT scores. Relative odds are a method of statistical comparison typically used in medicine. Odds of the magnitude observed at Michigan, according to Larntz, are virtually unheard of. The relative estimated odds of smokers’ getting lung cancer, for instance, are 14 times those of non-smokers, a figure that is regarded as huge. The difference in Michigan’s odds ratios is the difference between a presumptive admission for all but the worst black, Hispanic, or American Indian candidates and a presumptive rejection for all but the best white and Asian-American applicants.

Despite such overwhelming evidence, the university denies that it uses “quotas or the functional equivalent of quotas,” a “two-track” admission system, or a “double standard” in its admissions, things that Justice Powell in Bakke explicitly prohibited, even in the quest for diversity.

But given the questionable educational benefits of diversity and the inevitability that in its pursuit, racial preferences would slide into quotas, why did Powell open the door to this argument? After all, no Supreme Court ruling before Bakke — or since — has approved the use of race on anything but the narrowest of grounds, namely, the correction of identifiable instances of past discrimination by a specific entity. Indeed, in a string of rulings, the high court has rejected racial preferences for what appear to be much more “compelling” reasons, such as the rectification of past social discrimination, creation of “role-models” for minorities, and training of professionals to serve in under-served areas.

The reason the high court has consistently limited the use of preferences is that the Constitution demands that individuals be treated equally by the law without regard to race or other accident of birth. The appeal of the diversity argument is that it seems to offer a way around this hurdle. While the Constitution bars illegitimate discrimination, it does not regard all discrimination as illegitimate. Sports teams may discriminate on the basis of athletic ability, choirs on the basis of ability to sing, and modeling agencies on the basis of beauty. If universities can show diversity to be necessary to achieve their central mission — education — they may secure the right to discriminate by race.

Given that Michigan and other universities have demonstrated very little concern for any form of diversity other than racial, one may doubt that they truly believe the diversity argument. They have embraced it, it seems, because it allows them to promote a certain conception of social justice.

But there are real dangers in allowing universities, under the cover of diversity, to define racial discrimination as necessary to their mission. Should the Supreme Court endorse the diversity argument, affirmative action, adopted three decades ago as a temporary measure to remedy historic oppression, will become permanently enshrined in the American educational landscape.

The University of Michigan’s Gurin report cannot withstand close examination. But given the stakes in the diversity debate, such politicized science is likely to proliferate. Furthermore, there is always the theoretical possibility that more respectable studies somewhere, someday, might show real educational benefits of diversity. Even so, the crucial question would remain whether these benefits rose to the level of a “compelling state interest” as required by law.

To settle this, courts should consider: What if some future research were to demonstrate convincingly that students learn best not in a diverse but in a homogeneous environment? Would the educational benefits then justify discrimination against minorities? The answer clearly is no. But if such discrimination is wrong, how can that practiced by the University of Michigan be right?