In 1978, in Bakke, the first decision by the Supreme Court that focused on preferences in college and university admissions, four justices believed that race-based admissions preferences were justified – at least on an interim basis – by the need to redress past societal discrimination.  A different group of four believed that any consideration of race in admissions decisions was simply racial discrimination, and thus both illegal and unconstitutional.  Justice Lewis Powell famously charted a middle course, holding that the university could use racial preferences (but not racial quotas) to foster educational diversity and thus a better educational environment.  This position was eventually adopted by a full majority of the Court in 2003 in Grutter, which held that diversity in the pursuit of a university’s educational mission was a compelling interest and, if carefully implemented, constitutional.

There is a common sense element to the idea a diverse student body will benefit the educational process.  As Justice Powell pointed out, elite colleges like Harvard have long considered and sought many types of diversity, including geographic origin and diversity of interests.  Being exposed to students from different backgrounds and perspectives – other things being equal – would seem to be obviously beneficial to the process of learning about one’s society and becoming equipped to live and work successfully in a diverse world.

As we have noted elsewhere, numerous advocates have championed this approach and a number of scholars have conducted studies that document various positive educational outcomes associated with a diverse student body (see []). Others both oppose group-identity preferences as a normative matter and have published studies that dispute both the achievement of the claimed outcomes and/or the link between such matters and diversity per se (see social sciences tab and subtabs). It is not the goal or purpose of this website to claim to resolve this dispute or take a position on one side of the debate or the other. Rather, we have tried to collect the materials and give the readers what they need to understand the issues and reach their own conclusions.

That said, it is important to note three important considerations that must be taken into account.

The first is an important contextual reality. When racial diversity is achieved using large preferences, “other things” are not equal. The size of the preference inevitably change the dynamics.  For example, as Richard Sander’s survey of the available data has shown, the preferences used by the typical American law school are so large that there is a vast GPA gap between white and black law students.  The median white student at elite law schools has a GPA at the 53rd percentile of her class, while the median black student has a GPA at the 10th percentile. (racial grade gaps at elite undergraduate colleges are smaller, but still large). In such an environment, there are some obvious risks.  Students and even professors, unaware of the extent of admissions preferences, may stereotype minority students as less capable.  Minority students themselves who notice the correlation of race and performance may conclude that somehow, the university is discriminating against them.  And there can be unanticipated and potentially concerning social developments.

There is a good deal of anecdotal evidence of both of these phenomena, but (perhaps due to its political sensitivity) little hard research. One careful study did explore the experiences of Duke undergraduates, in particular the friendship networks they developed over the course of their college lives.  Arcidiacono et al., 2013. The authors found that at the outset of college, freshmen had relatively diverse friendship networks.  Over time, however, friendships tended to form along lines of academic performance:  high-performing students had more friends who were also high-performing, and low-performing students had more friends who were also struggling academically.  These effects occurred independent of the race of the student (and the race of the student’s friends), but because of Duke’s use of large preferences, performance itself was strongly correlated with race.  The consequence was a sort of racial stratification of friendships between the freshman and senior year. 

A different study, Boisjoly et al., 2006, reached somewhat similar conclusions, finding that assigning roommates on the basis of race had a positive impact on certain initial attitudes, such as support for diversity, but “had little or no effect on harder-to-change behavior . . . and long-term goals.” P. 1902. Tellingly, one of those was a factor identified by the Court in Grutter, “assigning greater importance to the imperative ‘helping promote racial understanding.'” Id.

The Duke study had the virtue of being longitudinal.  That is, it focused on how student attitudes and relationships changed, if at all, during the course of their time at Duke.  It also took into account key structural realities, in particular Duke’s strong commitment to affirmative action and diversify.  As one of the site Editors stressed in one of the very first such studies, “in order to understand the [actual] impact of an educational environment on students’ attitudes [their] pre-enrollment characteristics and experiences and entering attitudes must be considered as influencing factors.”  Killenbeck, p. 105.  They must also account for specific institutional characteristics that have an important role in shaping how students live and learn, including its “structural diversity, faculty diversity emphasis, institutional diversity emphasis, and student diversity experiences.”  Id.  And they must avoid focusing on diversity itself at institutions with a strong pro-diversity culture, given the reality that many student survey responses to matters touching on race will seek to conform to values advanced as “official” campus policy.

A second potentially important flaw in the theory that racial diversity serves a vital educational interest is the low value that many schools place on socioeconomic diversity.  As we note in the Race Versus Class subtab in the Controversies section, the argument is that if diversity of backgrounds among students is an essential ingredient for educational quality, then surely socioeconomic (“SES”) diversity is at least as important as racial diversity.  After all, one’s childhood SES predicts one’s future life course more strongly than does race, and many political beliefs vary more across SES lines than across racial lines.  Yet many schools that give heavy weight to racial preferences give no weight at all to socioeconomic diversity.  Elite law schools, to give a particularly compelling example, almost never give meaningful weight to SES diversity, and are thus overwhelmingly composed of (racially diverse) students from very affluent backgrounds.  [Statistics]. The argument, then, is that schools do not in fact have a commitment to educational diversity itself – they simply want to satisfy political demands for racial representation, even if that racial representation is achieved in ways that avoids, rather than promotes, real diversity.

A final consideration is that the actual evidence proffered to show the educational benefits of diversity is limited and, at least arguably, rather soft.   For example, a number of supposedly key studies consists of analyses showing that students who are in more racially diverse classes demonstrate a greater commitment to particular types of civic values.  But these findings might occur because (a) students who start out with a commitment to certain values are more likely to enroll in classes that have more racial diversity, or (b) the curricular content of the courses with racially diverse student enrollments may be different from the content of other courses.  So, for example, Patricia Gurin’s work, which played a major role in the University of Michigan litigation, does not persuasively separate out the diversity “treatment” so that its effect can be accurately measured.

Once again, our goal is not to take one side or the other in this debate. It is, rather, to identity potentially significant issues that must be taken into account.

List of Citations

Peter Arcidiacono, Esteban Aucejo, Andrew Hussey & Kenneth Spenner, Racial Segregation Patterns In Selective Universities. 2013: Journal of Law and Economics, Vol. 56, No. 4, pp. 1039-60.

Johanne Boisjoly, Greg T. Duncan, Michael Kremer, Dan M. Levy & Jacque Eccles, Empathy or Antipathy? The Impact of Diversity. 2006: The American Economic Review, Vol. 96, No. 5, pp. 1890-1905.

Patricia Gurin, Expert Report of Patricia Gurin. 1999: Michigan Journal of Race and Lw, Vol. 5, No. 1, pp. 363-426.

Ann M. Killenbeck, Racial Diversity in Legal Education: Do Racially Diverse Educational Environments Affect Selected Attitudes of White First-Year Law Students? 2000: Unpublished Doctoral Dissertation, University of Michigan.

Richard Sander, Why Strict Scrutiny Requires Transparency: The Practical Effects of Bakke, Gratz, and Grutter, pp. [], in Kevin T. McGuire (Ed.), New Directions in Judicial Politics in Judicial Politics. 2012: Routledge.

Additional Readings