Moving Beyond the Average Gang Membership Effect: Investigating Treatment Effect Heterogeneity in Concurrent and Long-Term Outcomes

Abstract

Objectives: This study investigates whether there is treatment effect heterogeneity in the “gang effect” for both concurrent (offending, drug use) and longer-term (offending, drug use, incarceration, education) outcomes according to baseline risk. Methods: Using panel data from the Rochester Youth Development Study, we first investigate whether the within-individual and between-individual changes in gang membership and deviance are conditional on an individual's baseline propensity to join a gang. Then, we investigate whether the relationships between gang membership and several young adult outcomes are conditional on this baseline propensity to join a gang. Results: Within-individual changes in gang membership status are associated with increased offending and drug use, but the effect does not vary across baseline propensity to join a gang. However, there is between-individual evidence that the effect of gang membership on offending and drug use decreases in magnitude as baseline propensity to join a gang increases. For each longer-term outcome, the negative effects of prior gang membership decrease as baseline propensity to join a gang increases. Conclusions: This study suggests that the average treatment effect of gang membership on several outcomes masks important heterogeneity, indicating it may be misleading to portray it as a general, universal negative turning point.