A Conversation with Althea Nagai
Althea Nagai, Ph.D., is our research fellow at CEO and the author of our studies of preferences in admissions in higher education, the most recent one of which examines their use at five public universities in Virginia. The other day I caught up with her in the hope of learning more about how she thinks about these matters. I was not disappointed, and trust that you will not be either. Here is my conversation with her, which treats issues in the news, including Harvard’s use of race in admissions, the consistently discriminatory fate of Asian American applicants, and Dr. Nagai’s preoccupation with odds ratios.
Q: Are more Asian Americans going to college and are they facing significant discrimination in admissions because of their race?
A: Yes, an increasing number of Asian Americans are going to college. In 1960, roughly 16,000 were in college, rising to roughly 250,000 by 1980. In 35 years, more than a million Asian Americans were attending college—quadruple the number in 1980. It’s hard to say how much discrimination is out there against Asian Americans in college admissions. One needs data and statistical analysis to even get close to estimates for individual schools.
Q: Are there schools with race-blind admissions that Asian Americans proportionately are admitted to, without regard to race?
A: There are state universities that are not allowed to discriminate on the basis of race such as UCLA and Berkeley, two schools that saw a significant rise in the number of Asian Americans on campus after Proposition 209. Also, the California Institute of Technology (Cal Tech), an elite private university, does not use race as a factor in admissions. It also does not consider legacy connections. As a result, Asian Americans as a percentage of undergraduates enrolled at Cal Tech grew steadily. By 2015, they were 40% of the student body.
Let’s contrast this with two other elite schools. MIT does not consider legacy but does consider the race of applicants. From the 1980s, the number of Asian Americans at MIT was increasing but basically stopped in the 1990s, peaking at 29 percent of the student body. That number has drifted downward and stalled at about 26 percent in 2016.
Harvard looks at both legacy and race in admissions. The number of Asian Americans sharply increased from the 1980s to 21 percent in the early 1990s, and then significantly dropped. It has stayed at roughly 17 percent until 2016. The numbers have risen since the Harvard lawsuit. (See my CEO study, “Too Many Asian Americans: Affirmative Discrimination in Elite College Admissions.”)
Q: Why do Asian Americans do so well in terms of academic preparation?
A: This has been a source of much conjecture and there are some academic studies, but I can’t review them here. We have to be careful about stereotypes, and I am not sure that it’s true of the Asian American college-age population as a whole. Moreover, whatever I say would be anecdotal and speculative. And, of course, it’s always self-selected applicants that apply to certain colleges and universities.
Q: In the studies you’ve authored for CEO have you seen an effort to base admissions to a greater extent on non-academic criteria, such as extracurricular activities and legacy status?
A: Harvard, of course, places great weight on legacy status. As for public universities, legacy status is not as much of as factor as race at the more competitive ones. In-state versus out-of-state residency is also a major consideration for public universities. And at some state universities like Virginia’s George Mason University, legacy data are not collected nor used in admission decisions.
Q: Is the treatment of Asian Americans in admissions an issue in the Harvard case? Would more Asian Americans be admitted under a race-neutral process?
A: Yes, being Asian American is a negative factor in Harvard’s decision-making. And, at least for now, the diversity defense makes it legal. As for admitting more Asian Americans under a race-neutral process, probably, if one is to believe the statistician in Harvard’s Office of Institutional Research. Before the lawsuit, the statistician created four models of what Harvard would look like in terms of race/ethnicity if the school looked only at academic factors, which would have made Asian Americans the largest group by far (43%). When statistically adding athletic and legacy status, and then extracurricular activities and a personality assessment, whites become the majority and Asian Americans drop to 26%. When finally taking race into consideration, Asian Americans drop even more, to 18%. The statistician also showed that blacks and Hispanics only increase significantly when race is taken into account. In later statistics, he found that “Asian American” was the only negative among more than 10 factors. (See my “Harvard Investigates Harvard: ‘Does the Admissions Process Disadvantage Asians?’”)
Q: Why are “data” important in the debate over race preferences in admissions?
A: One needs to have data, including their academic credentials, on all those who applied, which would include those who were admitted but chose to go elsewhere and those who were turned down. CEO always asked for the admissions status, race, residency, gender and legacy status test scores and grades of all individual applicants, in the form of a large spreadsheet, so I can do my own statistical calculations. The spreadsheets for one year at one public university may run up to 35,000 applicants.
Q: What is a logistic regression analysis and what is an odds ratio, and why are these tools of use to you?
A: Logistic regression is a statistical technique that allows the mathematical comparison of multiple factors when looking at the admissions process, such as academic credentials, residency, gender, race and legacy status. An “odds ratio” is a ratio of odds, in the CEO studies, a ratio of the odds of one group being admitted over the odds of another, e.g., Asians versus whites, blacks compared to whites, females to males. Odds ratios are part of the output of running logistic regression equations regarding admissions. I’ve used odds ratios in the statistical studies for CEO since the 1990s.
Q: Concerning the CEO study this year of preferences at five Virginia public universities, is there a growing problem with how Asian American applicants are treated with regard to their race?
A: I was surprised to find that logistic regression analyses showed all universities treating Asian American as a negative factor (but not a large one), when controlling for all other variables. (See “Preferences in Virginia Higher Education.”)
Q: Are there reasons to think that preferences are in decline—as the case of Texas Tech suggests?
A: The Supreme Court upheld the use of racial preferences in pursuit of diversity in college student enrollment opinion in the Fisher v. University of Texas case, but the Court narrowed its use to “a factor of a factor of a factor.” On a positive note, the use of race as a factor among the Virginia schools seems to have declined a lot from when CEO first examined the schools in the late 1990s, although I found that statistically the University of Virginia and the College of William & Mary still grant very large preference to blacks over whites. These were significantly more than preferences given to in-state residents.
