Harvard Admissions Officials Are Grilled on How They Use Race in Admissions

Updated (10/24/2018, 9:16 p.m.) with the afternoon’s testimony.

How exactly do Harvard College admissions officials think about race when they are evaluating hundreds of applications? Three admissions officers had to answer detailed questions about their process in court on Wednesday as part of a trial challenging Harvard’s admissions policies.

Students for Fair Admissions, an organization that opposes affirmative action, has accused the college, Harvard’s undergraduate division, of discriminating against Asian-American applicants. If all goes according to plan, Wednesday marked the halfway point in the three-week trial in a federal court here. Harvard has denied the plaintiff’s claims, saying race is never considered a negative factor in admissions.

Lawyers for the plaintiff tried to make the case that Harvard admissions officials aren’t trained in how to use race when making admissions decisions. They also tried to show that Harvard’s dean of admissions, William R. Fitzsimmons, has target figures in mind for the racial makeup of an admitted class. The admissions officials pushed back hard against those allegations. They each said they had never witnessed anyone in the admissions office display bias against any racial or ethnic group.

Fitzsimmons shares overall statistics about the admitted students when the 40-member admissions committee meets, the admissions officials testified. Charlene Kim, an admissions officer, said that the statistics concern more than the students’ race. They also include the gender breakdown, what parts of the country the students come from, and how many qualify for financial aid, among other things.

Kim was also asked about a process at the end of the admissions cycle called “lopping,” in which admissions officials reject or wait-list students who have thus far been admitted only tentatively. The officials know the students’ race, among other characteristics, during the “lopping” process, Kim said.

A document displayed later in the day showed that admissions officials consider whether applicants are the children of alumni, are athletes, and would receive financial aid, in addition to their ethnicity. Another admissions official said even those four factors are not the only ones considered in the “lopping” process.

A Legal Landscape

Kim also testified that she had received no written instructions on how to use race when assigning numerical ratings to applicants on four metrics: academic credentials, extracurricular activities, athletic ability, and personality. Admissions officials repeatedly said they do not consider race when assigning those ratings.

But she added that she and her colleagues talked regularly about race. She attended talks on the legal landscape of race in higher education, she said, and her colleagues had given presentations to their co-workers on different racial groups in Harvard’s applicant pool.

Tia Ray, another admissions official, was one of those colleagues. She gave one presentation on Asian-American and African-American students and another on Native American and Latino and Latina students. The presentations, she said, were designed to give admissions officers more context about students of color.

In one presentation, she noted that only a little over 2,000 African-American students in the United States score above 700 on the SAT, while more than 50,000 white and Asian-American students earn such a score.

Harvard on Trial

Detailed background on the lawsuit over the university’s race-conscious admissions policy, the case’s implications for selective colleges, and coverage of the trial as it unfolded, in a federal court in Boston.

• Harvard Doesn’t Discriminate Against Asian American Applicants, U.S. Appeals Court Rules
• 3 Takeaways From the Appeal of the Harvard Admissions Lawsuit
• A Judge Advised Harvard to Give Its Admissions Officers Training to Stop Bias. Will That Help?

“You were training your fellow admissions officers to understand the relative rarity of these applicants?” asked Katherine L.I. Hacker, a lawyer representing Students for Fair Admissions.

No, Ray responded. The presentation was meant to give her colleagues information about what might influence a student’s ability to prepare for and take the college-entrance exam.

Roger Banks, senior associate director of admissions at Harvard, was questioned about his involvement in the Association of Black Admissions and Financial Aid Officers of the Ivy League and Sister Schools. He said he had attended meetings at which admissions officials from each college in the group shared the numbers of applicants they’d accepted by race. Occasionally he would relay those numbers to Fitzsimmons.

But Banks said that the numbers had never been used to make admissions decisions at Harvard and that the association had no role in Harvard’s admissions process. The purpose of the organization, he said, is “to become more expert in this area, which is a very complicated part of administration in higher education. I think the case law proves that.”

Typos and Tea Leaves

Earlier in the day, a former official in Harvard’s Office of Institutional Research tried to play down the significance of some charts that he had made to analyze how much various factors, such as race and ethnicity, contributed to applicants’ chances of being admitted.

The former official, Mark Hansen, was questioned by a lawyer for Students for Fair Admissions who argued that the charts show that being an Asian-American applicant was a disadvantage in Harvard’s admissions process.

A lawyer for Harvard tried to show that Hansen was not a statistician but someone with a master’s degree in education who was just starting to learn how to create logistic regression models. Some of the charts shown in court on Wednesday contained typographical errors and incorrect dates, which Hansen agreed was evidence that the documents were more like “working notes” that he had created to familiarize himself with the data.

But one of the charts was eventually shown to William R. Fitzsimmons, Harvard’s dean of admissions. It included four models analyzing the pool of applicants for admission to the classes of 2007 to 2016. In the chart, Hansen tried to show who would be admitted if officials considered various factors. In the first model, he looked at who had a chance of admission if only academic credentials were considered. He then added preferences given to recruited athletes and the children of alumni in the second model. In the third model he added the scores that admissions officers give applicants to rate their personal qualities and extracurricular activities, and in the fourth model he added demographic considerations.

For each model, he calculated what proportion of the admitted class would be formed by applicants of different racial and ethnic backgrounds. Asian-American students would have made up 43 percent of the class when only academics were considered, according to the analysis. When the other factors were added, they would have dropped to 18 percent of the class.

During questioning by Harvard’s lawyer, Hansen said that his goal in making the chart was not to model the actual admissions process. Many other factors go into that process, other witnesses have testified, such as what part of the country applicants are from or whether their parents went to college.

Allison D. Burroughs, the Federal District Court judge presiding over the case, asked several questions about how to interpret the chart’s analysis. Among other things, she wanted to know what the chart does show, if not how the admissions process actually works.

“Hypothetical admitted-student pools, if you considered different factors,” Hansen said. The models could not support a conclusion, he said, that there was bias in the admissions process.

Nell Gluckman writes about faculty issues and other topics in higher education. You can follow her on Twitter @nellgluckman, or email her at nell.gluckman@chronicle.com.

Correction (10/24/2018, 6:28 p.m.): This article originally mischaracterized the nature of the analysis Hansen performed. For each model, he calculated what proportion of the admitted class would be formed by applicants of different racial and ethnic backgrounds, not their probability of admission. The article has been updated to reflect that.