How Do You Measure Diversity?

Diversity is a difficult thing to quantify. Measuring racial and ethnic diversity at a college or university is made all the more difficult, as we explained in our last post, by how the U.S. Department of Education categorizes students’ race—and how the department has changed its categorization methods over time.

But that hasn’t stopped researchers, administrators, and even Supreme Court justices from trying to use data to measure diversity in higher education. Keeping in mind that there’s plenty of nonquantifiable aspects of diversity—the “relationships,” as one recent commenter put it—and myriad complications posed by the underlying demographic data, we will explore two ways diversity can be measured and briefly evaluate some of the benefits and flaws of each approach.

Diversity Index

The diversity index is a tool, developed in 1991 by a researcher at the University of North Carolina at Chapel Hill and a USA Today reporter, for measuring the likelihood that two people, selected at random from a group, will be from different racial or ethnic backgrounds.

The diversity index is a useful indicator, especially for revealing changes in racial diversity over time—with the caveat that how race is reported has changed over the years. Nevertheless, an increase in a college’s diversity index suggests that the college has become more diverse; a drop can show that it has become less diverse.

For example, at the University of Florida, the diversity index in 2012 was 61.40, up from 38.10 in 1992.

Implicit in the diversity index is that the highest levels of diversity are achieved when every racial or ethnic group is equally represented. In reality, we live in a country that is almost two-thirds white and about 12 percent African-American. So even at the most diverse colleges, you wouldn’t expect white and African-American students to be equally represented.

And since most colleges primarily attract students from a particular state or region, and the racial makeup of those states may be quite different, the diversity index is a flawed tool for comparing the relative diversity of different colleges. One would expect, for example, that a college in southern Florida will be more diverse than one in North Dakota.

Instead it may be better to compare the racial and ethnic makeup of the University of Florida to that of the State of Florida, to determine how diverse the flagship university is compared with the state’s college-age population.

Comparison to State Demographics

Last year, when the U.S. Supreme Court considered the use of race in admissions at the University of Texas at Austin, The New York Times compared the percentage of black and Hispanic freshmen at several public universities in states with bans on race-conscious admissions to the proportion of college-age residents of those groups in those states.

Justice Sonia M. Sotomayor quoted that analysis—graphs included—in her dissenting opinion in the court’s recent decision to uphold Michigan’s ban on race-conscious admissions. Her point was that state bans on race-conscious admissions policies negatively affect the diversity of public universities in the state.

In 2011, 18 percent of freshmen at the University of Florida were Hispanic, compared with 27 percent of the state’s college-age adults, a nine-percentage-point difference. African-American students made up 10 percent of the freshman class and 24 percent of the state’s college-age population, a 14-point difference.

This method is effective because it takes into account the differences in racial and ethnic diversity around the country. Rather than assuming that diversity means that all racial and ethnic groups should be equally represented at a university, it assumes that enrollment at a public university—a flagship, in particular—should be representative of the racial diversity of the state’s population.

Nevertheless, some data problems make that particular calculation somewhat misleading. The Times used data from the Centers for Disease Control and Prevention on 18- to 22-year-olds to represent the state’s college-age population. But not all 18- to 22-year-olds in a state are in the pool of potential freshmen at the state’s flagship university.

A better indicator of how well the freshman class represents the state’s demographics might be the racial and ethnic breakdown of the state’s high-school graduates. (Or, with regard to admissions policies, the pool of applicants.) Unfortunately, those data are hard to come by. The National Center for Education Statistics collects data on public-high-school graduates by race, but that statistic doesn’t include graduates of private high schools. And the Census Bureau’s American Community Survey has data on high-school graduates by age or race but not both.

A Diverse Range of Views

As we’ve shown, no matter how you look at the numbers, it’s difficult to get a full picture of diversity on campuses. Either method requires making some choices about what exactly diversity means. Is it a measure of equal representation among racial and ethnic groups? Is it a measure of how closely a college’s racial makeup represents society writ large? Or is it something else entirely?

Neither of those diversity measurements captures the relationships or feelings of individuals on campuses. Many institutions, including Cornell University, the University of California system, Texas A&M University, the University of Wisconsin at La Crosse, and others, have used campus-climate surveys and other qualitative research methods to try to measure diversity—or the level of inclusion or exclusion of minority groups and others—on the campus. Some would argue that such an approach is a better way to measure diversity than quantitative measures, like proportions or statistics.

Furthermore, as many people have pointed out, diversity is more than racial or ethnic identity. From socioeconomic background to gender to religious affiliation and more, diversity is tough to measure and challenging to write about. It means different things to different people, and no measure of it is perfect. That doesn’t mean we should stop trying. It just means that we need to understand what we can measure—and what we can’t.

Return to Top