Lerner & Nagai: Evidence for the Educational Benefits of Diversity in Higher Education:
An Addendum

By Patricia Gurin
Professor of Psychology and Chair
Department of Psychology
University of Michigan

This short paper responds briefly to a report entitled A Critique of the Expert Report of Patricia Gurin in Gratz v. Bollinger authored by Robert Lerner, Ph.D. and Althea Nagai, Ph.D. This report was referenced in an amicus brief recently filed in the U.S. Court of Appeals for the Sixth Circuit in the Gratz case by the National Capital Association of Scholars, a chapter of the National Association of Scholars ("NAS") for the District of Columbia, Virginia, and Maryland.

On May 30, 2001, I responded to a report by Thomas Wood and Malcolm Sherman that criticized my expert witness reports. The Wood and Sherman report had been attached to an amicus brief filed by the main arm of the NAS. My response is published at http://www.umich.edu/~urel/admissions/new/gurin.html.

The majority of the claims by Lerner and Nagai repeat those in the Wood and Sherman report. In particular, they assert that I fail to show that percentage of minority students on university campuses has direct effects on educational outcomes, and that evidence showing the direct impact of percentage of minority students is the only evidence that is relevant to Justice Powell's argument in Bakke for the constitutionality of using race as one of many factors to achieve diversity. I have already shown in my response to Wood and Sherman that my emphasis upon the educational outcomes of actual interaction with diverse peers in classrooms and the informal campus environment is consonant with Justice Powell's argument -- and, perhaps more importantly from my perspective, with substantial learning in the field of social psychology.

I will not repeat my responses to the Wood and Sherman report here, nor will I engage in a detailed point-by-point refutation of the Lerner and Nagai report. The principal criticisms of all four authors were answered by an amicus brief supporting my methodology and findings filed by the Stanford Center for Higher Education Research. I do, however, want to respond to three particular points raised by the Lerner and Nagai report.

1. Affirmative Action and Intergroup Relations

Lerner and Nagai claim that the use of race as one of many factors in university admissions is thus likely to foster intergroup hostility because affirmative action increases the salience of groups. This claim misunderstands the social psychological research on group categorization, and is contradicted by results of my analyses.

Salience of Groups and Outgroup Hostility

Lerner and Nagai are correct that research on group categorization demonstrates that even minimal groups that are created in the laboratory produce in-group favoritism, expressed in greater liking and more positive evaluations of members of one's own group than for members of the other group. However, they are wrong in concluding that group categorization usually results in outgroup hostility and discrimination. A widely accepted view in social psychology, countering the expectation that Lerner and Nagai hold, is recently summarized by Gaertner, et al., (1999): "whereas social categorization can initiate intergroup biases, the type of bias it initiates represents primarily a pro-ingroup orientation rather than an anti-outgroup orientation, usually associated with hostility or aggression" (emphasis added) (p. 178). This conclusion was in fact reached by Stephan and Stephan in the intergroup relations chapter in the 1985 Handbook of Social Psychology. The motive that is established when groups are made salient, they said, is "to create positive social comparisons, a result which can be achieved by evaluating the ingroup positively without it being necessary to negatively evaluate the outgroup" (p. 615).

Lerner and Nagai are also correct that one model for achieving positive intergroup relationships (though they incorrectly suggest it is the only social psychological model) is to decrease the salience of group boundaries. Sometimes this is done by helping ingroup members individuate and personalize outgroup members (Brewer and Kramer, 1985). Sometimes it is done by altering group composition to achieve nearly balanced numbers of members of two groups (Mullen, 1991; Smith, 1992 — see my supplemental expert report of July 2000). Sometimes it is done in other ways to de-emphasize group membership (Mullen, 1991; Mullen, Brown, and Smith, 1992). Unfortunately, Lerner and Nagai imply that these strategies — and by implication getting rid of affirmative action, cultural awareness workshops, and group-based campus activities — are the only, or certainly the best, way to foster intergroup harmony.

There are other models that also promote intergroup harmony and that do not depend on minimizing the fact that groups exist. One focuses on the efforts of groups to work together and get along despite their differences. An example is a cooperative learning group in which students of different backgrounds are deliberately put together to learn and to solve problems. Slavin (1996) reviews the extensive literature on cooperative learning strategies and concludes that students from diverse backgrounds are more likely to have friends outside their own racial groups in cooperative-learning classrooms than in traditional classrooms (p 631).

Another model for achieving positive intergroup relations is to find common ground (Gaertner et al. 1989). This involves creating a common group, one that does not require subgroups within the superordinate group to forsake their group identities — as African Americans, Latinos (as), women, whites, etc. Gaertner and his colleagues (1999) contend that this approach is just as effective as individuating strategies.

The Predominantly Positive Quality of Interracial
Interaction at the University of Michigan

What actually happens at the University of Michigan? Have affirmative action and multicultural educational activities within and outside of the classroom produced the hostile effects that Lerner and Nagai predict based on their reading of social psychological research on intergroup contact?

I provided three sets of evidence in my expert report that counter the expectation of inevitable hostility when groups are made salient through such policies as affirmative action in admissions and diversity activities.

One concerns the extent to which students at the University of Michigan actually have negative or hostile relationships with students of different racial/ethnic backgrounds. I summarized results specifically about intergroup hostility from the Michigan Student Study in Appendix E to my expert report. No more than 7% of the white students reported having had "tense, somewhat hostile interactions" or "guarded, cautious interactions" quite a bit or a great deal with the students of color with whom they had interacted with most at Michigan. The relationships that Latino(a) and Asian American students had were also almost completely lacking in hostility and guardedness. Although African American students reported somewhat more negative interactions, only 15% spoke about hostile interactions.

The second set of evidence on interracial interaction was in the body of my expert report where I showed, again from the Michigan Student Study, that the students who had the most experience with diverse peers — through classes, informal relationships, multicultural events, and intergroup dialogues — expressed a stronger sense of commonality with students of diverse backgrounds. They also scored higher than students with less experience with diverse peers on a frequently used measure of perspective taking (Davis, 1983). The students with the greatest experience with diverse peers were more, not less, committed to understanding the points of view of other students. Moreover, since the Michigan Student Study had asked entering students about commonality with groups other than their own and had administered this same measure of perspective taking when the students entered Michigan, I was able to control for the possibility that these orientations might have predisposed students to seek out diverse peers. These results show that experience with diverse peers had increased sense of commonality and perspective taking — results completely contrary to Lerner/Nagai expectations.

The third set of evidence, also in the body of my expert report, showed that participation in a diversity program — the Intergroup Relations, Community, and Conflict Program — promoted a sense of commonality with diverse others, perspective taking, and belief that diversity can be congenial with democracy. This program combines the various models of intergroup relations that are delineated in social psychology. It provides equal status and close personal interactions across race, and in this way fosters individuation. It also makes groups salient by requiring students to participate in structured dialogues with an equal number of students from a different identity group. It gives the two groups a common goal of finding some issue on which they could agree to form a coalition or common action plan. Participants in this program were compared with non-participants, matched on gender, race/ethnicity, in/out of state residency, and Michigan residence hall. I was also able to control for participant/non-participant scores on sense of commonality and perspective taking when they entered the university. The matching procedures and control for entering scores increase my confidence in the positive intergroup impact of this program — again results contrary to Lerner/Nagai expectations.

Lerner and Nagai had access to this evidence. However, because they (like Wood and Sherman) apparently and wrongly believe that the Michigan Student Study is irrelevant to these issues, they ignored evidence that directly challenges the conclusions they draw from the intergroup literature. Their reading of this literature exaggerates what it says about outgroup hostility and discrimination. Then they fail to consider actual evidence from the very university they believe could not provide equal status relationships between students of different racial/ethnic backgrounds.

Lerner and Nagai assert that the average SAT differences between African American and white students at the University of Michigan make equality in relationships impossible. This assertion ignores the very sizeable SAT differences within these groups. Moreover, the very premise is questionable: there is no reason to conclude that equal status relationships are theoretically prohibited by SAT differences, and there is every reason to pay attention to actual evidence about intergroup relationships from the Michigan Student Study. That evidence shows interracial interaction is predominantly positive.

2. Critique of the Democracy Measures in My Expert Report

Lerner and Nagai charge that the measures I use in the national study of democracy outcomes of diversity are biased in the liberal direction. They selectively mention measures of citizen engagement from the CIRP study that they consider liberal, such as "helping others in difficulty" or "participating in a community action program." I do not agree that conservatives would not endorse these items. Many conservatives do want to help others in difficulty and do participate in community action. There is nothing in the community action item that implies it is action for a liberal program.

One can categorize the measures that I used from the CIRP study into those with the least political content, those that opponents of affirmative action or critics of multiculturalism might argue have a political content with which they disagree, and those that explicitly mention race or culture. The results show that diversity experiences were as consistently related to the measures with the least political content as to the other two categories.

Lerner and Nagai do not mention that the study of the Intergroup Relations, Community and Conflict Program at Michigan includes a five-item measure of interest in politics that presumes neither a conservative or liberal political stance, for example, discussing political events and keeping up with current events. This study also includes measures of actual participation in campus political activities, community service, student government.

I also include measures in the Michigan studies of democracy outcomes that directly test the argument made by Lerner and Nagai, that diversity and multicultural educational programs increase divisiveness between groups. I measured students' beliefs about diversity and divisiveness (e.g. that the University has put "too much" focus on diversity, that diversity fosters more intergroup division than understanding, that the emphasis on diversity results in inability to talk honestly about ethnic, racial, and gender issues, and makes it hard for the student to be himself/herself). All of these items are phrased in terms of a conservative position on diversity. The results show that the intensive classroom program did not foster divisiveness. In fact the participants, as compared to non-participants, perceived less divisiveness among groups and more compatibility with others.

Lerner and Nagai have provided a highly selective criticism of the evidence I presented on the effects of diversity on democracy outcomes.

3. Data Quality

Lerner and Nagai also raise technical concerns about the quality of the CIRP database, and specifically question the scope of the sample I used and the quality of the variables I selected for study. As I detail below, none of these issues undercuts the findings I presented in my original report. In fact, these criticisms are logically flawed or based on factually incorrect information, and simply serve to misrepresent or confuse the methodological details of my work.

The CIRP database is of high quality.

One of the three databases that I analyzed for my original report was a multi-campus database collected by the Cooperative Institutional Research Program (CIRP) which is located at the Higher Education Research Institute at the University of California, Los Angeles. The CIRP was established at the American Council on Education in 1966, and is the oldest and largest on-going empirical study of American higher education. A recent study of the higher education research literature (Budd, 1990) has shown that research using CIRP data are the most frequently cited by other scholars, which is a clear indication of the quality and importance of these data.

Lerner and Nagai raise questions about the reliability of measurement of the questions I analyzed in my study. A careful analysis of the test-retest reliability for the CIRP study was undertaken when the survey program was being developed, which led the researchers to conclude that "[w]ith rare exceptions, the reliability data obtained in this study indicate that all of the variables considered are sufficiently reliable for analytic use in statistical research" (Boruch & Creager, 1972, page 29). Among questions of the type I analyzed, this research showed that the reliability coefficients ranged from .70 to .99. It is also important to understand the effect of measurement error on the analyses: A well-known outcome of less-than-perfect measurement is known as attenuation of effects, where the strength and significance of findings would be underestimated (Rossi & Freeman, 1993). In any event, to the extent to which Lerner and Nagai's critique is correct, my results provide a conservative picture of diversity experiences and their effects.

The sample is appropriate to the task to which it is put.

As I have noted previously, my intent in database selection was simple: To select the most appropriate databases relevant to the legal cases brought against the University of Michigan. The campus databases I analyzed are by definition directly relevant to the lawsuits brought against the University of Michigan in that they document important aspects of the educational processes occurring on our campus. The multi-campus CIRP database used in the report was appropriately restricted to types of institutions that could provide information relevant to the cases at hand.

Lerner and Nagai raise several questions about the sample and design of the CIRP, but provide factually incorrect information in doing so. One aspect of this relates to the level of analysis issue raised in their critique, in which Lerner and Nagai call for an analysis that appears as Table 1 in my original report. To be sure, the structure of the CIRP data may appear to be complex, but the basic logic and approach is not to the careful reader.

It is important to note that erroneous information is provided by Lerner and Nagai in reporting participation rates in the CIRP. The CIRP is an ongoing research effort, and annually conducts surveys. Lerner and Nagai choose to cite institutional participation rates for a different freshman survey cohort than the one I analyzed (they apparently provide data on 1990 entrants, rather than the 1985 cohort). This is an error that is particularly difficult to understand since information on the database I used is also available in the publication they cite for the 1990 survey (Dey, Astin, & Korn, 1991). The correct total institutional participation rate for the cohort I studied and for the types of institutions I selected is more than double the incorrect rate provided by Lerner and Nagai (31 percent instead of the 14 percent cited on Page 34).

Lerner and Nagai also provide incorrect response rate information by confusing the database I analyzed with that used by Astin for his book What matters in college. As is made clear in my report, my analysis of CIRP database is drawn from students who completed surveys in 1985, 1989, and 1994; Astin analyzed data from only the first two survey administrations as the 1994 data were not available at the time of his writing. To be fair, Lerner and Nagai may have another study in mind, as they cite an Astin publication titled "College and Beyond" (page 31) but I am unfamiliar with any such publication.

Brewer, Marilynn B. and Kramer, Roderick M. (1985). The psychology of inter-group attitudes and behavior. Annual Review of Psychology, 36, 219-43.

Budd, J.M. (1990). Higher education literature: Characteristics of citation patterns. Journal of Higher Education, 61(1), 84-97.

Davis, Mark H. (1983). Measuring individual differences in empathy: Evidence for a Multidimensional Approach. Journal of Personality and Social Psychology, 44, 113-126.

Dey, E.L., Astin, A.W., & Korn, W.S. (1991). The American freshman: Twenty-five year trends. Los Angeles: Higher Education Research Institute, UCLA.

Gaertner, Samuel L., Dovidio, John F., Nier, Jason A., Ward, Christine M., and Banker, Brenda S. (1999). Across Cultural Divides: The value of a superordinate identity. In Deborah A. Prentice and Dale T. Miller (eds.), Cultural divides: Understanding and overcoming group conflict (pp. 173-212). NY: Russell Sage Foundation.

Gaertner, Samuel L., Mann, Jeffrey, Murrell, Audrey, and Dovidio, John F. (1989). Reducing intergroup bias: The benefits of recategorization. Journal of Personality and Social Psychology, 57, 239-49.

Mullen, Brian. (1991). Group composition, salience, and cognitive representations: The phenomenology of being in a group. Journal of Experimental Social Psychology, 22, 297-323.

Mullen, Brian, Brown, Rupert, and Smith, Colleen. (1992). Ingroup bias as a function of salience, relevance, and status: An integration. European Journal of Social Psychology, 22, 103-22.

Rossi, P.H. & Freeman, H.E. (1993). Evaluation: A systematic approach (5th Edition). Newbury Park, CA: Sage Publications.

Slavin, Robert E. (1996). Cooperative learning and intergroup relations. In James A. Banks and Cherry A. McGee Banks (eds.), Handbook of research on multicultural education (pp. 628-34). NY: Simon & Schuster-Macmillan.