Ethical College Admissions: Recruit to Reject?

What is the proper response if your child receives a communication from the admissions office at Harvard University with the following message: “Your strong grades and standardized test scores indicate to us that Harvard and other selective institutions may be possibilities for you”? Frame it? Buy a Harvard flag to hang outside the front door or a Harvard decal for the family SUV? Start softening your R’s so that when you visit Cambridge you can sound like a local when you ask directions to “Hahvahd Yahd”?

A recent New York Times article suggests that you should not get carried away by a message it describes as “seductive and flattering,” The correct answer, stated clearly in the article’s title, is “That Recruitment Letter From Harvard Probably Doesn’t Mean Much.”

The Times article references a recent National Bureau of Economic Research working paper by three economists, Peter Arcidiacono from Duke University, Josh Kinsler from the University of Georgia and Tyler Ransom from the University of Oklahoma. The paper, “Recruit to Reject? Harvard and African American Applicants,” is one of two working papers produced by the threesome using admissions data from Harvard that was released in discovery for the court case Students for Fair Admission v. Harvard. Arcidiacono served as an expert witness for the plaintiffs, analyzing Harvard’s data, policies and procedures, and now he has turned that into scholarship.

The other paper produced by Arcidiacono and his colleagues (the topic of last week’s "Ethical College Admissions" column) concerned the admission preferences given by Harvard to “ALDCs” -- athletes, legacies, dean’s (donors) list and children of faculty and staff. Their research covering the classes of 2014 to 2019 showed that admit rates for each of those groups is substantially higher than the 4 to 5 percent overall admit rate, ranging from 33 percent for legacies to 86 percent for recruited athletes.

What’s the connection between that paper and “Recruit to Reject”? Arcidiacono’s research shows that ALDCs are overwhelmingly wealthy and white, leading (or leaping) to a possible conclusion that preferences for ALDCs function as a form of affirmative action for the already privileged. How does Harvard achieve diversity in its student body given the number of admission slots consumed by the ALDC categories?

“Recruit to Reject?” analyzes Harvard’s use of the College Board Student Search program to recruit prospective students. For the Class of 2018, Harvard sent out 114,000 letters encouraging students to consider applying to Harvard. Half of those went to underrepresented minorities. Arcidiacono argues that with Harvard admitting only around 2,000 students with an admit rate that currently stands below 5 percent, most of those receiving letters from Harvard have no chance of being admitted.

That is particularly true, the paper suggests, for African American applicants contacted through Student Search. Arcidiacono and his colleagues argue that Harvard recruits differently based on race, using lower score cutoffs for Student Search parameters for African American candidates, but doesn’t change its reliance on high scores in the admissions process. As a result, “Harvard is actively sending recruiting materials to prospective students who effectively have no chance of being admitted.”

I don’t have access to the data or the ability to evaluate Arcidiacono’s statistical analysis, but clearly at the trial Harvard’s expert witness, David Card, economics professor at the University of California, Berkeley, challenged and disagreed with Arcidiacono’s interpretation of the data on Asian Americans. Here, though, is a summary of the argument in “Recruit to Reject?”

Arcidiacono and his co-authors organized Harvard applicants in the classes of 2014 to 2019 into deciles based on the Academic Index, the formula used by the Ivy League for athletic admission involving grade point averages, test scores and class rank where available. Their analysis shows that African Americans made up 11 percent of applicants during that period, but 41 percent of the bottom decile. Of the 5,921 applicants in the bottom decile during the six-year span, only two were admitted, both ALDCs.

(Inquiring minds want to know who they were.)

The paper concludes that those numbers are related to recruiting decisions made by Harvard. It observes that between 2003 and 2007 the percentage of African Americans in the Harvard applicant pool increased from 6.4 percent to 10.1 percent. During the same period, there was no change in the percentage of those admitted. There is an irony there. One of the arguments made to show supposed discrimination against Asian Americans by Harvard was that the percentage in the admitted class didn’t match the percentage in the applicant pool.

Arcidiacono et al. suggest that during that period, Harvard expanded the number of African American applicants by lowering test cutoffs in recruiting. Their evidence is that the number of African American applicants with SAT section scores below 550 increased sharply at the same time the applicant percentage increased. In the Class of 2009, there were more than twice as many African American applicants to Harvard with scores above 640 as those with scores below 550. By 2012 there were more applicants with scores below 550 than those above 640.

The very title of the paper suggests that Harvard is giving false hope to those receiving recruiting letters, and by implication is unethical. But I think the issues are more complex.

I think recruiting-to-reject strategies exist and are clearly wrong. A college or university that actively works to increase application numbers from students who have no chance of admission is taking unfair advantage of vulnerable, naïve young people. At the same time, application numbers and admit rates are metrics used by boards, college rankings and bond rating agencies. The bigger wrong is the belief that selectivity equals quality.

Does Harvard need to send out 114,000 recruiting letters? Probably not, but that doesn’t mean that sending them is wrong. A Harvard spokesman told the Times that 60 percent of applicants and 80 percent of minority applicants are a product of the recruitment efforts. I’m skeptical about that, but the essence of the argument for “undermatching” is that there are bright students out there who don’t know that places like Harvard exist or are a possibility.

What about Arcidiacono’s contention that Harvard recruits differently by race? According to the paper, in 2013 white and Asian American males received letters only if they had PSAT scores of 1380, whereas the threshold for African American males was 1170. The reality is that lots of colleges use Student Search with different parameters for all kinds of factors, not only race. The argument for differing standards is the hope of finding “diamonds in the rough” whose test scores may not reflect their academic potential.

The broader issue here is how we identify talent in populations that are historically underrepresented. There are two differing approaches to affirmative action at play in Harvard’s recruiting practices. One is to expand the pool of candidates by expanding outreach, which is what Harvard is doing by sending out 114,000 letters.

The other is to take into account differences in background preparation and the connection between test scores and socioeconomic advantage in evaluating different groups.

The danger in the second approach is that you end up with two separate admission processes. That was the issue in the Bakke case, the original Supreme Court case dealing with affirmative action in college admission. In that case, the medical school at the University of California, Davis, set aside a certain number of places for minority applicants and used lower thresholds for both grades and test scores.

According to Arcidiacono, Harvard also lowers its grade threshold in its Student Search recruiting from A-minus to B-plus for African American applicants. I’m not sure why, but for some reason that seems less defensible than lowering the test score cutoff.

The Students For Fair Admission v. Harvard case has certainly not ended societal discussion of how to provide opportunity to underrepresented groups in higher education. The biggest long-term benefit of that case may be the data providing a glimpse into how selective colleges and universities balance competing priorities. The two papers by Arcidiacono, Kinsler and Ransom represent a first attempt to make sense of that data. I trust and hope they won’t be the last.