Test Scores Don’t Explain Themselves

A test score can tell us something about a student’s performance. It cannot, by itself, explain how that student arrived there.

In this Facebook clip from John Stossel’s interview with Heather Mac Donald, she presents lower average Black test scores as an alternative explanation to systemic racism for unequal representation in education and employment.

My objection is to that framing. A skills gap and systemic racism can coexist. Unequal educational opportunity can help produce the very gap being offered as an alternative explanation.

To be fair, Stossel raises the legacy of slavery and inherited wealth. But the clip does not adequately address the question his challenge raises: how much of the gap reflects unequal opportunities to learn?

The U.S. Government Accountability Office, known as GAO, is a nonpartisan agency that conducts investigations and evaluations for Congress. Its 2022 report explains how a history of discriminatory practices has contributed to educational inequality, intertwined with differences in housing, wealth and income. It also reports that Black students generally attend schools with fewer resources and worse outcomes. [1]

These conditions affect children’s opportunities. Recognizing them does not deny individual effort or responsibility. It means examining the circumstances in which that effort takes place.

Employment discrimination also requires its own examination.

In a study published in 2004, researchers sent fictitious applications to employers in Boston and Chicago, randomly assigning names associated with Black or white applicants. Applications with white-associated names received 50 percent more interview callbacks. [2]

More recent research provides further evidence. In a nationwide experiment published in 2022, economists Patrick Kline, Evan Rose and Christopher Walters sent more than 83,000 fictitious applications, with randomly assigned characteristics, to 108 major U.S. employers. Black-associated names reduced the probability of employer contact by 2.1 percentage points compared with white-associated names. The size of the gap varied substantially between companies. [3]

A follow-up published in 2024 used the same experiment to assess racial and gender differences in employer contact at 97 named employers. It made those company-level differences publicly identifiable. This was further analysis of the earlier experiment, not a separate round of applications. [4]

These studies do not measure every employer or explain every employment disparity. They do show why differences in actual qualifications cannot explain all racial differences in hiring. Randomly assigning names allows researchers to distinguish employers’ responses to perceived race from differences in applicants’ credentials.

I am not claiming systemic racism is the sole cause of every educational or employment disparity. A disparity alone does not prove discrimination. Equally, a test-score gap does not disprove it.

Mac Donald goes further, saying, “You can have meritocracy in an institution, or you can have diversity. You cannot have both.”

That is a sweeping claim. A racial average cannot establish an individual applicant’s competence. Nor does the presence of qualified Black people mean an institution has abandoned excellence.

I consider this framing destructive because it encourages viewers to treat unequal outcomes as a reason to dismiss the barriers that can help produce them. It risks turning the consequences of unequal opportunity into a justification for leaving those conditions unexamined.

We should expect competence. We should also ask who gets the opportunity to develop it, and whose competence gets recognized.

If we are going to discuss the scores honestly, we have to investigate what produced them. Stopping at the numbers leaves the central question unanswered.