The pernicious predictability of anti-testing advocates

14 Feb The pernicious predictability of anti-testing advocates

By Dale Chu

A viral TikTok video posted earlier this week served up some freshly warmed-over disparagement of annual assessments. The digital screed was prompted by an opinion column’s review of a new paper by Jamil Maroun and Chris Tienken purporting to debunk the use of standardized tests as a credible measure of student learning. Like many of these types of criticisms, there’s a kernel of truth, but the video’s creator and the study’s authors arrive at the wrong conclusion. No one likes to take tests, but testing—whether its the three Rs, cholesterol, or blood pressure—shows us the realities no matter how much we might not want to see it.

When it comes to reading and math, Maroun and Tienken suggest that standardized tests are biased against low-income and minority students—and because of this, they say, these tests should be abandoned. Specifically, the authors insist that standardized test scores are a function of wealth (or the lack of it), but that’s like saying high cholesterol is solely the result of a poor diet. In both cases, it’s a gross oversimplification, and in neither would it justify the elimination of testing. Such a move would be penny-wise and pound-foolish. Indeed, low academic scores are much more likely to be the result of inadequate instruction.

But Maroun and Tienken nonetheless posit that a student’s background knowledge—which is largely predicated on a child’s family and upbringing—is the most significant variable when it comes to standardized test results. In doing so, they paint a false dichotomy between assessments and background knowledge. Now to be fair, students from marginalized backgrounds often have limited access to the life experiences (e.g., travel) that feed into the social capital required to succeed in society writ large. However, the deficits created by these disparities is not a function of the tests themselves. No, standardized tests are rather one of many tools—albeit the best one we currently have—that we should avail ourselves of in order to help gauge academic performance.

What’s more, the paper itself is a mess. First, as a friend of mine astutely observed, the dataset used in the study included less than half (41 percent) of the high schools in the state of New Jersey. Second, Maroun and Tienken should have used individual data rather than aggregated school-level results. Third, the authors were essentially trying to predict a school’s proficiency rate based on its poverty level. They claim that seventy-five percent of schools were within one standard deviation of their “expected” result, but the numbers were in fact pretty close to a normal distribution—and any differences could be explained by the spread of scores.

Then there’s the well-established body of research itself, which—contrary to what Maroun and Tienken assert—shows that standardized tests are valid and reliable. Even so, the repeated and overheated dismissal of the power of these measurement tools has everything to do with politics and little to do with the evidence. To wit, instead of standardized tests, the authors argue schools should use “teacher-made curriculum-based tests.” But this is a red herring: different tests serve different purposes. It’s simply dishonest to conflate the annual exams administered at the behest of states and the locally developed, formative variety used to inform instructional practice. You need both.

But all of this is impossible to convey on TikTok, where a growing number of Americans are going for their news and information. One thing’s for sure though: this won’t be the last “research” paper to fall victim to confirmation bias, and the hackneyed notion that standardized tests are punishing students and schools.