Be wary of unintended consequences: against test-optional

MCSG President Merrit Stüven, together with Ari Hymoff and Remy Eisendrath, published “Going test-optional: we are lagging behind” last semester in The Mac Weekly. As I care a great deal about the issues our president highlights in her article (chief among them diversity), I cannot help but be distressed by her proposition to eliminate the test requirement from admissions. Test-optional policies have a negligible effect (if any) on diversifying enrollment. Theoretically, they could even cause the opposite.

“Going test-optional” makes two evidence-based arguments: standardized test scores are poor measures of academic success and that test-optional policies improve a college’s capacity for reaching underrepresented groups. Both arguments are, in a very specific sense, true. The authors write, “[research showed] that SAT and ACT scores were, on their own, not predictive of academic success at Macalester.” This should be unsurprising—very few people would expect that standardized tests, on their own, to be predictive of academic success. But this conclusion is largely irrelevant, because no college decides admission solely on standardized tests. The question is whether standardized tests should be required as a factor in admissions, and the evidence cited, which states that SAT and ACT scores are only poor predictors “on their own,” seems to suggest that tests can be good predictors—as long as they are weighed alongside other factors, which, again, is already how college admissions operate.

“Going test-optional” quotes the work of Bill Hiss to claim that test-optional policies result in “an improved capacity to reach underrepresented students.” As with before, this is technically true—test optional increases the diversity of applicants. But more recent research published in Educational Evaluation and Policy Analysis by Belasco et al. demonstrates that the increase in applicant diversity is not reflected by an increase in enrollment diversity. Students of color and/or lower economic status may be more likely to apply, but they are no more likely to be accepted: “test-optional policies failed to effect a positive change in the proportion of low-income and minority students enrolling at test-optional institutions,” the report states. There is no evidence to suggest test-optional colleges have uniquely increased the diversity of their enrollment. Most test-optional colleges report that their campuses have become more diverse in the long run, but every college in the US has become more diverse in the long run. What real effect, then, do test-optional policies have?

The irony is that test-optional policies may actually suppress diverse enrollment. Research has demonstrated that the SAT and ACT are biased to favor the wealthy and white. But if an admissions office was to judge test-lacking applicants, they would compensate by placing more weight on factors like participation in extracurriculars and advanced courses (AP, IB, etc.) or college essays—factors that are just as biased in favor of the wealthy and white. Belasco et al. spectualate that “test-optional colleges may be inadvertently trading one inequitable policy for another.” I would extend their speculation a step further: test-optional colleges may be inadvertently trading an inequitable policy for a more inequitable policy. The bias in standardized tests is well-known and researched; it is quite feasible (and has, in fact, been done) to quantify the effects of race and wealth on test scores, and then weigh this effect alongside an applicant’s test results (this is also known as affirmative-action). But such a process is far less feasible for other admission factors—how does one quantify the impact of race or wealth on a college essay’s quality, or on whether a student became captain of a team? Discarding standardized tests means discarding one of the few factors of admission that can be held accountable to its bias.

The quality that standardized tests measure might be unimportant to whether someone is a worthy applicant. But that is a different debate. If the aim is to improve a college’s diversity, the evidence is clear: test-optional admissions are ineffective. If any group is to benefit from a test-optional policy, it would likely be wealthy whites that perform poorly on standardized tests. In the pursuit of a more diverse community, “Going test-optional” will take us, at best, nowhere, and at worst, backwards.