Do ACT and SAT Favor Older Students?

The pandemic has been painful for many involved in higher education, especially the standardized testing industry.

Most colleges have gone test optional for at least a few years, and many have done so permanently. The colleges that have studied going test optional report that their admitted applicants are doing well. A major criticism of the tests is that students who are wealthier (and who are white or Asian) do better, on average, than students who are Black or Latinx.

Now comes a new criticism: age discrimination.

The critic is Pablo A. Peña, an assistant instructional professor in the Kenneth C. Griffin Department of Economics at the University of Chicago. He makes his case in an article out today in Education Next, which is published by the program on education policy and governance at Harvard University’s Kennedy School.

The article, “End the Birthday Bias,” reviews evidence from a variety of settings in the United States and Britain about the uses of standardized tests in elementary and secondary schools and in college admissions.

“Older children typically perform better on academic achievement tests than younger students in the same classes,” Peña writes. “Time and again, studies looking at an array of countries, grade spans, and subjects have found that age differences of even a few months do matter.” When he says “older” children, he’s talking about months, not years, making a difference.

Peña argues that testing companies should grant “age allowances,” or some extra points for younger people, on tests taken up until the age of 18, including the ACT and the SAT.

But he acknowledges that it won’t be easy. “No institution or school district operates in isolation, and many use the same or similar admissions exams,” he writes. “So adopting age allowances unilaterally may be a bad idea. Imagine that one selective school decides to make an ‘in-house’ age allowance in its admission process while comparable institutions don’t, but they all use the same test. The institution adopting the age allowance would experience a drop in unadjusted test scores. Of course, admissions to that institution would be fairer. But the average quality of incoming students as measured by test scores would look worse relative to both past incoming classes and peer institutions.”

He adds, “Despite a growing movement toward ‘test-optional’ admissions, average SAT scores remain a high-profile metric for many institutions, and any school that adopted age allowances would mechanically fall in college rankings. It’s unlikely that any one institution, even if interested in fairness in admissions, would want to be the first to adopt age allowances.”

How would Peña solve the problem? “To overcome our collective action problem, we can make age allowances at the source,” he writes. “Test creators and test administrators don’t face the tradeoff between fair admissions and institutional ranking. They also observe all test takers and are well-positioned to determine how big or small the ‘bump’ to younger students should be.”

He cites a study by Steven Hemelt and Rachel Rosen that “found that 12 months of age bump scores on the ACT by as much as three percentiles.” In addition, he has done “preliminary analysis” on SATs and age. His findings: “Students who retake the test one year after their first time gain about eight percentiles. To be sure, second-time testers may be more familiar with the SAT format and have undertaken more preparation than students sitting for the test the first time. But … the fact that they are one year older also would seem an important factor.”

Peña also notes that redshirting, in which some parents chose a later date for their children to start kindergarten, “is more prevalent among white children from high-income families” and “it contributes to the gaps in test scores observed along income and racial or ethnic lines.”

He adds that “by making redshirting less appealing, age allowances could simultaneously save resources and help level the playing field—a rare chance to enhance efficiency and equity at the same time.”

Inside Higher Ed asked the College Board and the ACT to comment on the article, and the College Board declined to comment.

ACT issued this statement: “ACT research has not established evidence of the age effect currently described in the paper. More research may be needed to understand what, if any, age effect exists for the ACT test.”