We highlight that analyses using interaction terms to study treatment effect heterogeneity are susceptible to a form of omitted variable bias that is often overlooked in economics. Unlike most instances of omitted variable bias, the omitted variables in this case are available to the researcher but were not included in the model. We demonstrate that this exclusion matters based on a replication of 205 estimates across seventeen papers published in the American Economic Review over a five-year period. For approximately 60% of these papers, failing to account for the omitted variables changes the majority of estimates by more than 100%.

School districts historically approached conflict resolution from the perspective that suspending disruptive students was necessary to protect their classmates, even if this caused harm to perceived offenders. Restorative practices (RP)—focusing on reparation, accountability, and shared ownership of disciplinary justice—are designed to address undesirable behavior without harming students. We study Chicago Public Schools’ adoption of RP and find that suspensions and arrests decreased, driven by effects for Black students. We find null effects on test score value added, ruling out meaningful average declines. We estimate a 15 percent decrease in out-of-school arrests, consistent with RP substantively changing student behavior.