Gratz, et al. v. Bollinger, et al., No. 97-75321 (E.D. Mich.)
Grutter, et al. v. Bollinger, et al., No. 97-75928 (E.D. Mich.)


Study Methods

To determine how learning and democratic sentiments are related to structural, informal interactional, and classroom diversity, as our theoretical review suggests that they should be, I reviewed the literature (see Appendix B) and undertook three new sets of analyses developed specifically for this litigation. These systematic analyses were designed to provide scientific insight into the processes by which students are changed by their college experiences. I use research data specifically collected from students at the University of Michigan, as well as data collected from students attending colleges and universities across the country.

Before reviewing the conclusions based on this research evidence, it is important to convey a general sense of the approach that was used in these investigations (Appendix C provides a complete technical description of the analyses). The approach was based on standard, generally accepted methods for analyzing data that were collected by ongoing programs of research on college students. As developed through decades of research on college students, the approach has two characteristics, each of which is an essential aspect of the quality and soundness of the results:

Data over time Growth and development among college students obviously takes place over time. As a result, the most effective research approaches use data collected from the same individuals at more than one time point. This so-called "longitudinal" approach, in which researchers collect information from students on two or more occasions, allows for a systematic analysis of how students grow and develop by comparing data collected from individuals at one time to data collected from these same individuals at later points in time. Moreover, by comparing patterns of growth with the educational conditions and activities that students experience between the collection of data, it becomes possible to understand how different experiences promote growth and development among college students.

Taking choices
and consequences
into account
In studying students over time it becomes apparent that individuals do not make choices randomly, nor do they leave their previous attitudes and experiences at the front doors when they enter their colleges. As a result, the choices that students make (and the consequences that these choices have) need to be taken into account in order to make sound judgments about how campus experiences affect students.

For example, we are likely to find that students majoring in mathematics and science have growing interest in science, as compared to those majoring in the humanities. While this may seem to prove that growth in scientific interest is caused by majoring in science, it is important to recognize that those who were drawn into science majors are likely to have been more interested in science when they entered college. In order to make a fair judgment about whether majoring in science or the humanities is differentially related to growth in science interest, we need first to take into account the initial differences in interest between these two groups.

Similarly, to study the growth and development of learning and democracy outcomes as related to diversity experiences, it is important to take into account (or control for) differences across individuals in terms of their initial position on learning and democracy outcomes, as well as their likelihood to be drawn to more intensive diversity-related experiences. I accomplished this through either statistical approaches or through matching students who did or did not have a diversity experience, as in the study of the Intergroup Relations, Conflict, and Community Program.

The results I present here provide a conservative estimate of diversity's effects, in that the analyses consistently allow other variables in the analysis (i.e., characteristics of colleges and entering characteristics of students) a greater opportunity to account for, and possibly explain away, the influence of campus diversity on college students. Whereas in baseball a tie always goes to the runner, in these analyses a "tie" always goes against the diversity explanation. Despite the fact that this approach tends to diminish the likelihood of demonstrating effects related to diversity, it is important to take these relationships into account in order unambiguously to demonstrate change related to diversity. In sum, this approach ensures that where I report significant effects related to diversity, they are truly diversity effects, as opposed to being a consequence of the characteristics, choices, and preferences that students bring with them to college.

The data bases used for the analyses span a broad range of approaches typically used to study college student development issues. For example, I analyzed data provided by the Cooperative Institutional Research Program (CIRP) and the UCLA Higher Education Research Institute that were collected from 9,316 students attending nearly two hundred colleges and universities. In addition to the national perspective provided by the CIRP data, I also analyzed data from the Michigan Student Study (MSS) provided by 1,321 students on the educational dynamics of diversity on the Michigan campus. The data came from a series of extensive questionnaires given to all undergraduate students of color and a large, representative sample of white students at the time they entered the University of Michigan in 1990, and again at the end of their first, second, and senior years. A more specific study, focused on the Intergroup Relations, Community, and Conflict (IGRCC) Program, demonstrates these dynamics related to a specific diversity initiative at the University of Michigan.

Primary Variables in the Studies

Figure 1 graphically shows the elements of the research approach used in the three sets of analyses developed for this litigation. Variables identified in the box in the upper left corner of Figure 1 (student background characteristics) represent control variables across the studies, and while these are not of primary substantive interest, they are important considerations in the analyses because they represent the previous choices, preferences, and experiences of students that, unless taken into account, could have influenced the outcomes and caused me to overestimate the effects of diversity.

The primary variables of interest are those related to campus diversity in its many forms (represented in the center of Figure 1). I was interested in understanding how these variables affect (or predict) different student outcomes. Therefore, each analysis contains a variable representing a student's level of contact with classroom diversity and a variable representing a student's informal interactional diversity. Structural diversity is also directly represented in the analyses that are based on data from the national study of many institutions, as these institutions vary in the degree to which they attract and enroll a diverse student body.

As detailed below (as well as in Appendix C), not all of the elements shown in Figure 1 were available in each of the three sets of studies. Although the studies were designed to be as parallel as possible, differences in questions asked and in research design made identical analyses impossible. The most obvious example of this is the omission of the information on institutional characteristics -- especially structural diversity -- from the analyses of data on the single institution, the University of Michigan. This is obvious given that while institutional characteristics vary across institutions, they do not vary for a single institution except over time.

I examined classroom diversity in all studies. It was measured in the CIRP study by students' enrollment in ethnic studies courses in college. In the Michigan Student Study, it was measured by the extent to which students were exposed to and influenced by classes that dealt with issues of race, ethnicity, and interracial relationships.

I also examined informal interactional diversity in all three studies. In the CIRP and Michigan Student Study, the measures covered a broad range of ways in which informal interactions occur on campus. In both studies, distinctions were made between the diversity of a student’s closest friendships and more general interracial interactions on campus. Within the latter, both studies also distinguished between the amount of interracial socializing and the extent to which these interactions involved discussions about racial issues and attempts to deal with those issues. In addition, the Michigan Student Study included questions on the quality of these campus interracial interactions, whether they were positively personal and honest, or negatively cautious, guarded and somewhat hostile.

The intensive study of the University of Michigan's Program on Intergroup Relations, Conflict, and Community provided the opportunity to examine the combined effect of classroom and informal interactional diversity. This Program integrates a classroom experience with explicit interaction with diverse peers, using dialogue groups that were built into the formal class on intergroup relations.

Major Outcomes and Their Relationship to My
Theoretical Arguments for the Impact of Diversity

Since I was able to conduct analyses to understand how diversity influences student learning and democracy outcomes at the national level, the institutional level (focusing on the Michigan context), and at the level of a classroom in which interaction with diverse peers was fully integrated with curricular content, I was able to take an increasingly close look at the impact of diversity. Together, these analyses are akin to macro- and microscopic looks at how diversity works at various levels. Although the studies were not originally designed to have parallel measures, they did include similar concepts, which can be grouped into long-term learning and democracy outcomes.

The outcomes I examined conform to the learning and democracy consequences that I discussed above in my theoretical statement. I argued that a more diverse university environment stimulates a more active engagement in the learning process and results in the development of less automatic and more complex thinking about issues and causality, as well as in the greater learning that comes from this engagement. The major categories of learning outcomes, therefore, refer to measures of:
  • growth in active thinking processes that reflect a more complex, less automatic mode of thought (in the MSS and IRGCC studies),
  • engagement and motivation (included in both the CIRP and MSS),
  • learning of a broad range of intellectual and academic skills (in the CIRP study),
  • and value placed on these skills in the post-college years (in the CIRP study).

With respect to democracy outcomes, I argued that students educated in diverse institutions are more motivated and better able to participate in an increasingly heterogeneous and complex society. I reasoned that to participate effectively, students have to (1) learn to understand and consider the multiple perspectives that are inherent in a diverse environment; (2) deal with the conflicts that different perspectives sometimes entail; and (3) appreciate the common values and integrative forces that incorporate these differences in the pursuit of the broader common good. The major categories of democracy outcomes refer to:
  • citizenship engagement (in all three studies),
  • racial/cultural engagement (CIRP and MSS),
  • and compatibility of differences (in MSS and IRGCC).

"Citizenship engagement measures motivation to participate in activities that affect society and the political structure, as well as actual participation in community service in the five years after leaving college. It also includes a measure of understanding how others think about issues, what (as described earlier) is commonly called perspective-taking in cognitive psychology. "Racial/cultural engagement" measures cultural knowledge and awareness, and motivation to participate in activities that promote racial understanding. "Compatibility of differences" includes belief that basic values are common across racial and ethnic groups, understanding of the potential constructive aspects of group conflict, and belief that differences are not inevitably divisive to the social fabric of society.

In addition to these learning and democracy outcomes, the nine-year CIRP study has enabled me to study behaviors and perspectives, which I will call living and working in a diverse society. Attending a diverse college and participating in its educational and peer environments that utilized diversity should help break the pattern of perpetual segregation that so many social scientists have documented. To test this, I analyzed post-college interracial interaction patterns in friendships, neighborhoods, and work settings, and obtained graduates' views of how the college years had prepared them for graduate school and for jobs after college.

Next "Gurin" Section  |   "Gurin" Table of Contents  |   "Compelling Need" Table of Contents

Questions? Comments? Please send e-mail to
Site last updated: September 5, 2012.   Copyright © 1997–2013 Regents of the University of Michigan.