Morgan Polikoff and Paul Bruno

Using data from Milwaukee Public Schools (MPS), this report analyzes the association of high school computer science (CS) coursetaking with subsequent performance in mathematics and science courses. The report includes CS, mathematics, and science coursetaking and grade data for approximately 60,000 high school students between 2014-15 and 2021-22, of whom approximately 21% took at least one CS course. In addition to looking at an overall CS effect, we consider the effect separately for different types of CS classes—EC, IB, and other CS classes. 

            To analyze the data, we use econometric techniques, with two main forms of statistical models. In the first model, we leverage variation among students in a given math or science course in terms of their prior CS coursetaking, controlling as well for their prior math/science GPA. In the second model, we examine variation in grades within individual students before and after having taken CS courses.

            Looking across the results, we generally find small negative associations of CS coursetaking with subsequent grades, typically in the range of .01 to .03 GPA points. These associations are sometimes, but not always, statistically significant. There is some suggestive evidence that these negative associations are driven by non-ECS CS courses, especially in mathematics. We do not find strong evidence that these associations differ across the types of science and mathematics courses. We conclude by discussing limitations of the available data for answering the questions of interest, contextual differences between this study and prior similar studies of CS coursetaking, as well as potential threats to the validity of the findings.