ArticlesAffirmative Action Harms Minorities

Recently, the non-profit group Students for Fair Admissions (SFFA) sued Harvard university for racial discrimination against Asian-American. Ever since, Harvard has been a breeding ground for controversy. It has caused some Asian Americans to rally against affirmative action as they see it as a platform to keep Asian Americans out of elite institutions. Others defend it to promote the interests of people of color.

It’s no secret that although Asian Americans consistently score higher on the SAT, achieve higher GPAs, do more extracurriculars they still face the lowest admission rates out of the racial groups. According to Harvard’s internal records, the percentage of Asian Americans currently enrolled in one year was 19%, whereas if admissions were based purely on academic that percentage would 43%, more than doubling their numbers. Clearly, this harms Asian-Americans at the cost of pursuing a more diverse student body, which is something that Harvard and supporters of affirmative action present as a goal worth pursuing. However, something rarely discussed how affirmative action may actually harm the under-represented minorities (URM) supposedly benefiting from it.

In this article, Edmund Zhen and I will layout first how it can harm minorities, especially African Americans in the STEM disciplines, and second how it harms Asian-Americans, which was briefly discussed in the introduction.

Part 1: Mismatch

Mismatch is the theory that when a student with lower academic credentials is placed in a school in which they are taught a curriculum catered towards students with significantly stronger credentials, that student struggles, resulting in outcomes like lower grades or lower rates of graduation. This is self-evident.  A less academically prepared student coming into college will likely do worse than a more academically prepared counterpart. Indeed, this phenomenon is readily observed in legacy students and student-athletes who get preferential treatment in the admissions process. When this same phenomenon is applied to racial preferential treatment via affirmative action, many are skeptical because upon initial impression affirmative action could only be of benefit to URM. However, there is evidence to suggest that such policy does harm to URM by pushing them from STEM fields to the humanities and softer sciences and lowering the number of degrees from STEM disciplines.

Two papers were published by Duke economists Peter Arcidiacono and Esteban Aucejo, one detailing the racial disparities between GPA and major choice, and the other talking about the effect of proposition 209, which banned affirmative action in the UC system, on graduation rates. In What Happens After Enrollment? An Analysis of the Time Path of Racial Differences in GPA and Major Choice, the economists argue that the GPA of Duke black and Hispanic students compared to their white and Asian peers is due to a lack of academic preparedness. This disparity can be seen in the chart below.

They note that the initial difference, which was the largest in the first semester decreases substantially by the last semester. It is possible that this gap is reduced because the students acclimate to their more rigorous academic environment, eventually catching up to their peers. However, they explain that this reduction in difference could be explained by decreased variance in grading in upper-level courses, and different grading standards in the different majors. They observed that in Duke, STEM courses have average grades 8% lower than grades given in humanities and social sciences. This deflated grade is also observed in STEM across universities in the US. Upon coming into college, black and white students show similar interest in STEM with blacks, but by the time they decide on their final major, 68% of the latter choose humanities and social science majors compared to less than 55% of the former. The retention rate is astonishingly low, with black males dropping from 76.7% initially choosing STEM to 35% getting a degree, and for black females, 56% and 27.7%. In comparison, the proportion of whites that stay in STEM are 10% smaller than those that came in with an initial interest in STEM, dropping from 60.8% to 50.5%. Their findings are replicated in the National Longitudinal Survey of Freshman, showing a similar pattern across 28 selective colleges and universities.

Table 7 also shows that among who entered college with an expected major, there is a higher initial commitment to STEM fields by black students than by white students. That trend reverses by the time they receive degrees due to major changes. Table 8 shows that among students that do not report “Do not know” for initial major, the initial interest rates are similar for blacks and whites, both around 60%.

Once considering smaller grading variance, and major choices between racial groups, one can explain an alternative hypothesis as to why the GPA gap is smaller by the last semester. This gap is also confirmed by studying legacy students who often come in with lower academic achievement than the average student, who suffer from similar phenomena of lower GPA, and lower retention rates in the STEM fields. They clarify that the problem of mismatch is less a matter of race and more a matter of qualification.  They conclude, “affirmative action may be working to increase the number of non-science majors at top schools at the expense of science majors at less-selective schools.”

If the American people want to encourage more minorities into the STEM fields, they should be attentive to this study, which concludes that affirmative action may harm their prospects in the field by pushing them out of STEM and into the humanities. On the other hand, the study didn’t account for factors like racism, lack of mentorship and research opportunities, all of which wouldn’t be present in historically black colleges and universities(HCBUs). In fact, the number of graduates with STEM degrees from HCBUs are overrepresented when one considers that HBCUs make up 3% of the nation’s institutions of higher education. They produce 38% of blacks graduating with biological science degrees, 31% in mathematics, 35% in computer science, and 22% in engineering. In such institutions where racism is likely absent, and where faculty makes active efforts to foster mentor-mentee relationships, and to provide research opportunities, students would likely be more encouraged to stay in STEM majors. However, HCBUs do not preferentially select for their students based on their race thus this lack of mismatch may account for their impressive rates of STEM graduates.

This is but one paper in the literature opposing affirmative action. In another paper by the group of Duke economists, they propose that graduation rates of minorities in the UC system under Proposition 209, which banned racial preferences in the UC school system, raised by up to 4.4 percent. They hypothesize that the higher rate of graduation was because admitted minority students were better matched to their school and performed at rates like the median admitted student. However, an article by the Brookings Institute responded to the latter article, discrediting it by saying that the increased graduation rate was a part of the trend of raising graduation rates for URM. It concludes that “the current weight of the evidence leans strongly against the mismatch hypothesis”, but fails to address the myriad of other evidence such as those provided by people like Dr. Richard Sander, and  Dr. Rogers Elliot in Report of the U.S. Commission on Civil Rights in 2010.

In the report Sanders present evidence such as the University of Michigan graduation rates for both black and white students were much lower for those who received large preferences than for those who received no preference, which you can see in the table below. Sanders went as far as to argued that racism was not a meaningful factor. In his analysis of Michigan, he demonstrates that for black and white students with the same credentials, black students were not only more likely to graduate, but also more likely to major in the sciences. This evidence, though limited, demonstrates that at least in U of M credentials are important in producing science degrees and contradicts the hypothesis that discrimination and inadequate academic support undermine black and Hispanic academic performance.

In the same report Dr. Elliot Rogers found that within four Ivy League schools, persistence rates (shown in table 2 below) in STEM were 34 percent for blacks and 56 percent for Hispanics, and 62 percent for the white/Asian group. He also found that URM had, on average, lower credentials than whites and asians, confirming Sander’s mismatch data. He too attributes the lower persistence rates to the level and pace of coursework, especially at selective schools. Rogers also finds that the students entering with the top third SAT math scores made up more than 50% of the science degrees, while those entering with the bottom third SAT math scores only made up 15.4% of the degrees (shown in slide 3 below). The majority of those in the bottom third were URM. His findings led him to conclude that although there is value of a degree from an elite school, it may be better for an aspiring minority scientist to enter a school where they would be more likely to get a degree than to suffer the consequences of being a relatively under credentialed student.  There is even evidence to suggest that if URM go to schools where their credentials would have been at least average, there would be more minorities with STEM degrees.

Even in the pro-affirmative action book The Shape of the River: Long-Term Consequences of Considering Race in College and University Admissions, which was cited in Justice O’ Connor’s decision in the
Grutter v. Bollinger case in 2003, admits that those who are recipients of affirmative action are plagued by below average grades during their undergraduate years. To reiterate, this is not strictly about race, but about qualifications, as most of the publications against affirmative action make it clear that mismatch can affect anyone regardless of race. If an individual enters a college where they have below the median academic credentials, they are likely to perform at the level that their credentials suggests. This happens on a near universal level in post-secondary institutions due to affirmative action, primarily making URM less likely to persist in STEM.

If the intention of affirmative action is to encourage upward mobility in URM then it is imperative for Americans to encourage more URM representation in STEM, as students with a degree in STEM consistently outearn those that have non-STEM degrees. A look at the literature shows that affirmative action is doing the exact opposite.  By placing underprepared URM in environments where they are unlikely to succeed, affirmative action only results in funneling would-be STEM majors to humanities majors. Policy should be reformed to reflect the most recent and compelling evidence so that America can ameliorate racial inequality.