On both sides of the pond, we are very used to uncritical media reporting on claims made about the unique and extensive discrimination facing black ethnic minority students with little or no evidence. For example, in the UK, few questioned the Runnymede Trust’s 2015 The School Report’ in which the authors claimed that ‘educational success for minority ethnic groups also needs to address broader issues – for example, cultural capital, role modelling and the transition from school to university or work’.
While these factors are worth considering when researching social facts, too often the subjective variable at the level of individuals is ignored. Below, Frederick Prete, based in the USA, analyses data from the ACT assessment used by American universities to assess candidates for further study. The data considers both student performance and their aspirations. Frederick’s careful interpretation of the data offers hope for those who believe that while social and economic contexts are important, they are not all-determining.
Claims that the ACT is racially biased ignore the data, and the effects of hard work and high aspirations
Frederick R Prete
I taught ACT classes to high-school students for over a decade. To keep abreast of changes in the test, I took more than a dozen myself. So, I know the ACT better than most. I’ve also published several historical analyses of the ACT results (here, here, and here) that convinced me that the ACT is a fair, unbiased, accurate measure of what students have — or should have — learned in high school, despite oft-repeated but erroneous claims to the contrary.
The ACT has four multiple-choice subject sections: English, math, reading and science. The number of correct answers on each section is converted to a scaled score from 1–36 and averaged into a student’s composite score. Thirty-six would be perfect.
In 2005, the ACT established four, empirically-based College Readiness Benchmark Scores, which were updated in 2013 to 18 for English, 22 for math and reading, and 23 for science. These are the minimum scores that predict a 50% chance of earning B, or a 75% chance of earning C in a corresponding credit-bearing college course.
To understand trends in the ACT, you have to look beyond year-to-year fluctuations. (The ACT suggests five-to-10-year intervals.) So, although I’ve done finer grained analyses elsewhere, here I’ve compared the most recent (2023) scores to those from one and two decades prior (2013 and 2003).
In 2023, approximately 1.4 million students took the ACT, intermediate between the two earlier cohorts (1.2m and 1.8m, respectively). In 2003 and 2013, the average composite scores were virtually identical (20.8 and 20.9), and the subject scores differed by less than 0.3 point. However, in 2023, the average composite score was just 19.5 and subject test scores dropped by as much as 9%. In addition, there was a 21% decline (from the 2003/2013 highs) in the number of students meeting any of the College Readiness Benchmarks. Although the latest declines caused a lot of handwringing, and wrongheaded claims about testing biases, there is a more nuanced — and positive — story in the data.
First, it’s clear that at least one key reason for the 2023 decline is that students are taking a much less rigorous course load than in previous decades. In 2023, only 41% of students had taken what the ACT calls ‘Core or More’, four or more years of English and three or more years each of math, social studies and natural science. In 2003 and 2013, 57% and 74% did so. The effects of this change are clear: over all three cohorts, students who took Core or More scored an average of 11-19% higher on subject tests, and 14% higher overall than did students who took a less-rigorous mix of courses. These differences are greater than the overall (aggregate) decline in scores over the same decades, suggesting that the lower 2023 scores are attributable (at least in part) to a decline in academic rigour.
More importantly, however, Core or More was most beneficial for students who, historically, have fared poorly on standardised tests. For instance, in 2023, the percentage of students self-identifying as Black/African American, American Indian/Alaska Native or Native Hawaiian/Other Pacific Islander who met at least three of the ACT’s College Readiness Benchmark scores was 100-150% higher for those who took Core or More than for those who did not. Similarly, 32-72% more students in the ACT’s other racial/ethnic categories met at least three of the College Readiness Benchmarks if they took Core or More. In other words, rigorous coursework dramatically increases test performance for all groups of students irrespective of race or ethnicity.
If one considers only overall average composite scores — as most do — the ranking of the ACT’s ethnic/racial categories has remained unchanged for decades. For instance, in the three cohorts considered here, students identifying as Asian (or Asian-American/Pacific Islander in 2003), White and Two or More Races (in 2013 and 2023) scored as much as 24% higher than their cohort’s overall average. In contrast, those identifying as Hispanic/Latino, Native Hawaiian/other Pacific Islander, American Indian/Alaska Native, and Black/African-American scored as much as 18% below the average. These aggregate data are what’s used (erroneously) to claim that the ACT is racially biased.
However, as I have argued elsewhere, grouping scores by these arbitrary and anachronistic racial/ethnic categories not only perpetuates misleading stereotypes, it hides the key predictor of ACT performance: Student Post-Secondary Aspirations. The ACT reports student aspirations as their self-reported plans to pursue a Graduate, Professional Level, Bachelor’s, Two-Year College, or Vocational-Technical degree after high school.
In 2003, the ACT only reported the overall composite scores in terms of student aspirations. Nonetheless, the data were amazing. Scores were 34% higher for students aspiring to a Graduate Degree versus those intending to pursue Voc-Tech, and the intermediate scores declined progressively between the two extremes. The same pattern repeated in 2013 and 2023. More importantly, however, the composite scores of students in the top aspirational categories (Grad. Study, Prof. Level Deg., Bachelor’s Deg.) actually increased each decade by as much as 8%.
In 2006, the ACT began reporting scores in terms of both aspirations and ethnic/racial categories. Again, the results were remarkable. In 2013 and 2023, students in the four historically lowest performing groups (Hawaiian/Pacific Islander, Hispanic, American Indian/Alaska Native, Black/African American) who aspired to a Graduate or Professional degree actually outscored students in the historically highest performing groups (Asian, White, Two or More Races) who aspired to a Two-Year College or Voc-Tech degree by as much as 42%. In other words, aspirations were a far better predictor of student performance than were the ACT’s outmoded racial/ethnic categories.
So, despite declines in overall average performance, from 2003 to 2023, students who took more rigorous coursework and had higher postgraduate aspirations not only improved, they outperformed other students irrespective of their race or ethnicity. In fact, in 2023, students who planned to pursue Graduate or Professional degrees outperformed their cohort’s average by as much as 41%, again, irrespective of race or ethnicity.
In other words, hardworking, ambitious Black and Hispanic students did better on the 2023 ACT than Asian and White students with more modest ambitions. Did anyone notice?
Frederick R. Prete received a PhD in biological psychology from the University of Chicago, and is a Senior Instructor in the Department of Biology at Northeastern Illinois University, USA. He studies the neurobiology of vision, and the use of haptics in assistive technologies for people who (like him) are visually impaired.