“NIH Stands Against Structural Racism in Biomedical Research.” This was the title of a statement released on March 1st, 2021, by Francis S. Collins, who was then the Director of the National Institutes of Health. The statement continued:
As a science agency, we know that bringing diverse perspectives, backgrounds, and skillsets to complex scientific problems enhances scientific productivity. NIH has long supported programs to improve the diversity of the scientific workforce with the goal of harnessing the complete intellectual capital of the nation. These efforts, however, have not been sufficient. To those individuals in the biomedical research enterprise who have endured disadvantages due to structural racism, I am truly sorry. NIH is committed to instituting new ways to support diversity, equity, and inclusion, and identifying and dismantling any policies and practices that may harm our workforce and our science.
This is just the beginning of an effort that has a concrete goal of achieving racial equity but has no scheduled end point.
In previous months and years, other influential voices in the world of science and medicine have issued similar statements. On June 6th, 2018, the journal Nature announced that “Science benefits from diversity” and that “Statistics from the US National Science Foundation show that the representation of minority ethnic groups in the sciences would need to more than double to match the groups’ overall share of the US population.”
On March 31st, 2021, the Journal of the American Medical Association reported, “This study found that Black, Hispanic, and Native American people were underrepresented in the 10 health care professions analyzed. Although some professions had greater diversity than others and there appeared to be improvement among graduates in the educational pipeline compared with the current workforce, additional policies are needed to further strengthen and support a workforce that is more representative of the population.”
On April 9th, 2021, the journal Science declared that “The lack of diversity in the scientific and health professions—a longstanding manifestation of racism—can no longer be ignored, excused, or attributed to uncontrollable factors.” Then, a perspective entitled “Racial and Ethnic Diversity at Medical Schools—Why Aren’t We There Yet?” published the New England Journal of Medicine on November 4th, 2021, argued that “In medicine, diversity at all levels—from the frontline workforce to executive suites and from classrooms to laboratories—is an essential component of efforts to achieve equity.”
These voices are not calling for equality of opportunity, but equality of outcome (commonly referred to as “equity”)—proportional representation for minorities in the fields of science and medicine. So, if blacks make up 13 percent of the US population, then advocates for equity expect blacks to make up 13 percent of the workforce in these fields. (Statements deploring the deficit of minority racial and ethnic groups in STEM fields do not typically include Asian Americans—hereafter referred to as Asians—since they are well-represented. Calls for diversity also demand greater representation for women, but I will not be discussing that issue here.)
There are facts available, however, that can help us to determine whether or not the stated goals of diversity advocates are actually attainable. For example, how many minority students in the STEM pipeline today could realistically become the medical practitioners and medical and scientific researchers of tomorrow? The claim that “systemic racism” is the main reason for minority under-representation in STEM fields cannot be proven unless there is first an assessment of how many academically qualified minority students are available to meet equity goals. Data are hard to come by, but there is a way of estimating the percentage of minority students with the academic potential to succeed in a STEM field.
Most readers will be familiar with the SAT (formerly known as the Scholastic Aptitude Test), a college admissions test administered by the College Board since 1926. A less publicized test is the ACT (formerly known as the American College Test), which originated in Iowa City, IA, in 1959. The SAT and ACT are administered to high-school students to assess their academic readiness for college. At first, this test was mostly administered to high-school students in the Midwest and South. (I took the ACT as a high-school senior several decades ago.)
More students take the SAT than the ACT. But over the years, the ACT has gained in popularity, and by the mid-2010s, approximately two million high school students were taking the test each year. The SAT consists of two parts: Evidence-Based Reading and Writing, and Math. Scores on each section range from 200–800. Adding the two parts of the test yields a total score ranging from 400–1600. The ACT consists of four parts: English, Reading, Math, and Science. Each test has a range of scores from 1–36. A composite score is calculated by averaging the scores from the four parts of the test and rounding to the nearest whole number. The highest possible composite score is 36. An ACT to SAT score converter is available online.
The SAT total score and the ACT composite score are highly correlated (approximately .86 where a correlation coefficient of 1.0 represents perfect agreement). In other words, the two tests measure very similar academic skills. Since 2007, all US colleges and universities that accept college admission test scores have accepted both the SAT and the ACT (although many schools have recently gone “test optional” or do not require either test). Some schools outside of the US also accept SAT and ACT scores as part of the college application process.
Opponents of standardized tests (including the SAT and ACT, but not limited to them) have claimed that the tests are biased against minority students because those students get lower scores, on average, than white and Asian students. However, a sizable body of scientific literature—examining tests such as the SAT, ACT, Graduate Record Examination (GRE), and Medical College Admissions Test (MCAT)—demonstrates that this is not the case. An article in the American Psychologist concluded that:
(a) tests of developed abilities are generally valid for their intended uses in predicting a wide variety of aspects of short-term and long-term academic and job performance, (b) validity is not an artifact of socioeconomic status, (c) coaching is not a major determinant of test performance, (d) tests do not generally exhibit bias by underpredicting the performance of minority group members, and (e) test-taking motivational mechanisms are not major determinants of test performance in these high stakes settings.
Other studies (see here and here) have reported similar results. The ACT’s own research has shown that white students get better overall college grades than minority students at any given level of ACT composite score up to approximately 31; at that level and above, white and minority students earn the same grades. This means that, at most levels of ACT composite scores, minority students actually underperform expectations when they get to college. At no level do minority students get better college grades than those predicted by their ACT composite scores, which is what we would expect to see if the ACT were biased against them.Source: ACT
The ACT also has an extensive process to ensure that its test is fair and unbiased, and has conducted extensive research on the reliability with which its test predicts college grades. That research has found that both high-school grades and ACT scores predict success in college, and a combination of high-school grades and ACT scores is better than either one alone in predicting success.
The ACT has also conducted empirical research into what score on each of its four component parts makes a student “college ready”—when the student has a 75 percent chance of getting a C or higher grade, or a 50 percent chance of getting a B or higher grade in a freshman-year college course in the relevant subject area. Students taking the ACT can be “college ready” in anywhere from zero to four of the subject matter areas. Obviously, being “college ready” in all four subject areas of the ACT provides a student with the best chance of success in college. “Students who earn first-year grades of B or higher, on average,” reports ACT, “are much more likely to complete a postsecondary degree.”
The data in the following table have been extracted from Table 1, ACT College Readiness Benchmarks:Source: ACT
For STEM majors, the ACT score requirements established by its empirical research are 25 in Science and 27 in Math (ACT now defines “STEM ready” as scoring an average of 26 in these two subject areas). Not all “college ready” students will be “STEM ready,” and a student who does not meet ACT’s STEM benchmark could still earn a degree in a STEM field. However, based on ACT’s research, students who met the STEM benchmark were more than twice as likely to earn a degree in a STEM field than students who failed to meet the benchmark. Meeting the STEM benchmark also means that the student will likely graduate with a higher grade point average. They will, therefore, be a more competitive candidate when applying to medical school or to a graduate degree program in a STEM field.
College readiness and STEM readiness vary considerably between demographic groups. The data in the following table are extracted from Table 3.3, Percent of Students Who Met ACT College Readiness Benchmark Scores by Race/Ethnicity:Data based on 1,670,497 high-school students from the graduating class of 2020 who took the ACT. The data only show results for all students and for black, Hispanic, white, and Asian students, and include both male and female students. Source: ACT.
Black students made up less than three percent of the approximately 330,000 students from the graduating high school class of 2020 who took the ACT and were considered to be “STEM ready” based on their Science and Math scores. Hispanics made up eight percent of the “STEM ready” students, whereas white students and Asian students made up 64 percent and 11 percent, respectively.
Long-term trends in ACT composite scores by race/ethnic group are shown in the following table:Source: ACT.
Data are for 2005, 2010, 2015, and 2020. Only Asians increased their average composite score over this period. The percentage of students in each group who were “STEM ready” is unknown for the years before 2020. But because STEM scores are positively correlated with composite scores, these data strongly suggest that the percentage of students who were “STEM ready” has increased only among Asian students since 2005.
This analysis of the percentage of minority (specifically black and Hispanic) students who are academically prepared for careers in STEM fields is sobering, and highlights the difficulty of achieving equity in STEM fields. It contains information, all in the public domain, of which advocates of equity in medicine and science seem to be unaware or which they refuse to acknowledge.
The instrument most commonly employed in recent years to increase minority representation in STEM fields has been affirmative action, which has been widely practiced in college admissions and in faculty hiring. However, despite these initiatives, minorities remain under-represented in the STEM workforce. The group Students for Fair Admission, representing Asian students, has filed a lawsuit against Harvard College alleging discrimination in its admissions practices. A federal district court judge has ruled against the plaintiffs.
The case has been appealed to the US Supreme Court, which has accepted it for review. The outcome of this litigation is difficult to predict. (See here for an explanation of the quantitative aspects of Harvard’s alleged discrimination co-authored by Duke Economics Professor Peter Arcidiacono, an expert witness for the plaintiffs. See here for a critique of the lower court’s ruling upholding Harvard’s admissions policy.) The Supreme Court’s decision could determine the future of affirmative action for college admissions in the US.
Certainly, non-cognitive factors are also important for college and career success. Conscientiousness, for instance, is a personality trait associated with college success. Although conscientiousness can be measured by psychometric tests, those tests are not generally used by college admissions committees or by employers. Instead, the “holistic” approach to college admissions and hiring used by many institutions probably takes indirect indicators of conscientiousness into consideration. However, a high level of conscientiousness cannot make up for academic underperformance.
How, then, is it possible to achieve the goal of equity in medicine and science given the low percentage of minority students who have demonstrated the academic potential for successful STEM careers? There is no easy answer. It is not my intention to make policy recommendations here. Nothing in this article should be interpreted as suggesting that minority students in general are not academically capable of succeeding in STEM fields. Nor am I blaming minority groups for their academic underperformance or suggesting any specific cause. I am simply pointing out that academic underperformance by those minorities under-represented in STEM fields is well-documented, that it is not an artifact of biased testing, and that it has shown no sign of improvement over a 15-year time span.
It is therefore inappropriate and irresponsible to attribute the under-representation of minorities in STEM fields to “systemic racism” without first addressing the issue of minority academic underperformance. “Reality,” the science fiction novelist Philip K. Dick once remarked, “is that which, when you stop believing in it, doesn't go away.” The reality of minority academic underperformance is not about to go away but it is something that proponents of equity don’t want to face. It is time they did.
This is a companion discussion topic for the original entry at https://quillette.com/2022/03/11/it-may-not-be-possible-to-achieve-racial-equity-in-american-scientific-research/