The 2015-16 academic year was one of numerous student protests demanding increases in the admission of minority students and the hiring of minority faculty members — not to mention numerous other measures to promote inclusivity on campus.

But what exactly do students want? While some would say that the various lists of demands of campus protests provide much of the information, two Dartmouth College professors disagree. On Tuesday they proposed on the political science blog The Monkey Cage a new way of measuring student interest in different forms of diversity. And they tested their system on students at Dartmouth, an admittedly nontypical student body given that the college is highly competitive in admissions.

The results at Dartmouth (which they hope to compare to results at other institutions) suggest that students are indeed committed to diversity in admissions and faculty hiring. But students (and the results are generally true across student racial and ethnic groups) are more concerned about enrolling students and hiring faculty members from some groups than others. For both admissions and faculty hiring, Asians are not as valued as other minority groups, the survey found. And when it comes to the diversity of faculty hires, black students and female students are more concerned about the issue than are other students.

The two professors who did the research are John M. Carey, the Wentworth Professor in the Social Sciences and professor of government, and Yusaku Horiuchi, the Mitsui Professor of Japanese Studies and professor of government.

They argue that simply asking students if they are pro-diversity or favor certain measures to promote diversity may not yield accurate results. “Conventional surveys on sensitive issues are susceptible to various biases,” they write. “If those who choose not to answer the survey have different opinions from those who do, the results won’t accurately reflect group opinion. Even those who do participate may adjust their responses to harmonize with what they think the researchers want to hear — a phenomenon known as ‘social desirability bias.'”

So Carey and Horiuchi argue that a better way to discern student sentiment is to use “conjoint analysis,” in which respondents are given choices between two options who are candidates for admission or for a faculty position. The candidates have a range of characteristics, including race and gender, but also academic record, disciplinary interests and a range of other characteristics — all randomly rotated. In this way, researchers can find out what students really want in admitted students and in faculty hires. They note that this approach means they are giving students questions that test their interest in certain kinds of diversity without ever saying that the survey is about diversity.

With regard to admissions, this study found that Dartmouth students highly value applicants with strong high school grades and SAT scores. Further, they favor the admission of recruited varsity athletes. And (with responses the same for students from all demographic groups) the students said that they valued diversity. Given candidates who were equal in other ways, the Dartmouth students:

• Preferred a black or Native American applicant by 15 percentage points over a white applicant, and a Latino applicant over a white applicant by seven percentage points.
• Women were slightly preferred over men.
• Students “strongly preferred” disadvantaged students over “the affluent.”

The Monkey Cage piece does not say how the students reacted to Asian applicants. Via email, Carey said that students were given scenarios involving Asian-American, Chinese and Saudi applicants. For Asian-American and Chinese applicants, there was no statistical significance in preferring them over white applicants. But there was a small preference for Saudi applicants over white applicants.

With regard to faculty diversity, traditional academic credentials were highly valued, but students also cared about faculty diversity. Comparing otherwise equal candidates, students picked African-American candidates by 12 percentage points over white candidates, American Indians or Latinos over whites by nine percentage points, and Asians over whites by five percentage points. (Minority students and scholars at Dartmouth have complained in recent years about what they see as a pattern of tenure rejections of minority scholars, most recently of an Asian-American English professor.)

Carey acknowledged the limitations of the questions asked. Much of the debate over affirmative action does not focus on candidates who are equal in every way beyond race and ethnicity. Proponents of affirmative action argue that many minority and disadvantaged students with great ability may not have test scores or high school transcripts that are as impressive as those from wealthy high schools.

“An even thornier set of issues confronting affirmative action programs is around trade-offs between demographics and other attributes — for example, how much to give up in terms of percentile rank (on standardized tests, class rank, etc.),” Carey said. “We will be able to produce estimates along those lines, but our first cut estimates a set of ‘all else equal’ quantities — specifically, differences in probabilities of selecting a candidate/applicant if we shift a given attribute (say, race) from the baseline level (white) to some other level (say, black), all else equal.”