When ‘Black’ & ‘Hispanic’ Students Outscore ‘Asian’ & ‘White’ Students on the ACT, Nobody Notices

In biology, “race” is a taxonomic term, and like all biological terms, it’s a bit ambiguous. Sometimes it’s used as a synonym for “subspecies.” Most often, however, "races” are defined as groups of plants, animals, or bacteria within a species that are genetically, geographically, or physiologically distinguishable but still able to reproduce with each other.

As currently applied to humans, however, contemporary scientists consider ‘race’ a social construct based on societal definitions without any inherent physical or biological meaning. One of the reasons is that there is much more genetic (biological) variability within human racial categories than between them. In other words — to paraphrase the evolutionary biologist Richard Dawkins — whereas human sex is binary (there are only two), race is a spectrum.

Ironically, however, despite the fact that academics consider human racial categories to be socially constructed and ambiguous, they continue to use the categories as if they represented well-defined, homogeneous groups of people. (Think of the controversies over college admissions.) I find this especially confusing when racial categories are used to explain differences in educational achievement such as performance on standardized tests like the ACT.

I’ve published several articles (here, here, and here) pointing out that race or ethnicity — as the categories are currently used — are neither the most reliable predictors of, nor the key factors affecting student performance on the ACT. Others have, too. Yet, the race-based narrative continues to dominate the discussion.

A large part of the confusion is that the colloquial racial categories used by the ACT (and other such tests) are biologically arbitrary, ambiguous, or contrived, and they are often confounded with other variables such as self-reporting biases and socio-economics.  In other words, the categories, themselves, don’t refer to any well-defined groups of students. Even worse is the fact that they perpetuate a variety of inaccurate stereotypes.

The focus on race notwithstanding, the ACT does report student scores in terms other than “Race/Ethnicity.” The most important of these is “Postsecondary Educational Aspirations” (what the student is planning to do after high school graduation). Unfortunately, educational aspirations are generally ignored in discussions about standardized tests despite the fact that they’re far better predictors of student performance than self-reported race or ethnicity.

The ACT is a standardized test that predicts the likelihood of a student's success in their first year of college. The overall ACT score is based on the results of four subject tests; English, math, reading, and science. The number of correct multiple-choice answers for each section is converted to a scaled score ranging from 1–36, and the four scores are averaged and rounded to a whole number. (For clarity, I’ll follow the rounding convention here.)

Thirty-six is a perfect score. In 2022, approximately 1.35 million students took the ACT. Overall, the average score was 20, the lowest since 1991. To put this into perspective, between 1990 and 2021, the average was 21. That one-point decline in 2022 represents a difference of 3 to 4 correct answers out of 215 total questions.

In terms of the ACT’s racial categories, in 2022, the highest scores were earned by students identifying as Asian (25), White (21), and Two or More Races (20). The next three groups were Hispanic/Latino, Prefer Not to Respond/No Response, and Native Hawaiian/Pacific Islander (18). The two lowest-performing groups were American Indians/Alaskan Natives, and Black/African-American (16).

This ranking by race has remained virtually unchanged for a decade, contributing to the belief that the test is racially discriminatory. However, it’s impossible to know exactly how students sort — or choose not to sort — themselves into any of the inherently diverse racial/ethnic categories, or the demographics of the ambiguous Two or More Races category (first introduced in 2012). Even more confusing is the fact that the largest decline in scores last year, and over the last five years (by 11%) was in the Prefer Not to Respond group the demographics of which are unknown.

The arbitrariness of the ACT’s racial categories is especially evident in the creation of the Asian student category. In 2012, the ACT separated the Asian-American/Pacific Islander group into two cohorts that turned out to be dramatically different academically. That year, there was a 13% increase in the number of students in the newly formed Asian group and a 12% decrease in the number of Pacific Islanders who met the ACT’s College Readiness Benchmarks compared to 2006. Further, since 2012, the Asian group’s ACT performance has steadily improved while all other groups have declined. This suggests that the original Asian-American/Pacific Islander group was a demographic of convenience that masked large within-group differences in academic performance. Further, given the broad diversity of the two new groups, it’s unclear why their performances diverged so radically. Assuredly, the same would happen if any of the other student groups were similarly subdivided.

When viewed in terms of the ACT’s racial/ethnic categories, student performance appears to be largely determined by self-reported race and ethnicity. However, in 2006, the ACT added seven Postsecondary Educational Aspiration categories to their analyses. These included Vocational-Technical Training, Two-Year College Degree, Bachelor's Degree, Graduate Study, Professional-Level Degree, and two which I won’t consider here, Other and No Response. When the ACT data are considered in terms of these categories, a very different picture emerges.

The effects of educational aspirations were nothing less than remarkable. For instance, in 2006, 2012, and 2022, across all student groups, ACT scores were 31-51% higher for those aspiring to a professional-level degree compared to those planning on vocational-technical training, and scores were 12-17% higher for those aspiring to a graduate degree compared to a bachelor's degree.

These effects have been even more pronounced within racial groups every year since 2006. For instance, in 2022, students aspiring to a professional-level degree earned ACT scores averaging 47% higher than those planning on vocational-technical training. The largest differences - 53% and 61% - were in the Native Hawaiian/Pacific Islander and Prefer Not to Respond/No Response groups. The smallest difference - 35% - was in the American Indian/Alaska Native group.

More importantly, in 2022, students in the four lowest-performing groups who aspired to graduate study or a professional-level degree earned scores as much as 39% higher than students in the highest-performing groups who aspired to vocational-technical training or a two-year college degree.

In other words, Black, American Native, Hispanic, and Hawaiian/Pacific Islander students with high educational aspirations outscored Asian, White, multiracial, and non-identifying students with lower aspirations. Unfortunately, no one seems to have noticed.

Epilogue

Over the past 15 years, educational aspirations have consistently had a greater impact on ACT performance than self-reported race or ethnicity. This suggests that by increasing students' aspirations, regardless of their demographics, their academic performance can be improved. One way to do this is by teaching students that they can — and will — succeed if they work hard, and that they are not at the mercy of uncontrollable external circumstances.

As a Biological Psychologist — and someone who taught ACT prep courses for over a decade — I understand that the factors influencing student performance are complex and often the result of long-term trends that vary within and between groups, regardless of how the groups are defined. However, using outmoded, ambiguous, or contrived racial categories to interpret educational performance is misleading and unhelpful.

Racial categories have external validity only to the extent that each represents a meaningfully homogenous, culturally and/or genetically distinct group of people. However, none of these criteria are met by the colloquial racial categories currently applied to people. Distinguishing students by, for instance, family or socioeconomic demographics, internalized cultural beliefs, personality traits (such as resilience or aspirations), school district, or ZIP code would all be more informative and useful in improving educational policy than the current crude, indeterminate racial categories.