Deterrence and labeling theories make opposing predictions regarding the effect of sanctions on subsequent crime. Deterrence anticipates that sanctions deter, while labeling anticipates that sanctions amplify future crime. The knowledge base with respect to this question is vast, and while a handful of studies provide evidence of a deterrent effect, the majority of studies indicate a null effect. Our study examines whether an arrest leads to an increase in subsequent crime, but extends the knowledge base by considering whether an arrest has the same effect across offender trajectories and by employing techniques that deal with sample selection bias. Thus, we assess for whom sanctions deter or exacerbate subsequent offending. Results indicate that for greater risk youth, arrest amplifies subsequent delinquency, net of other effects, but not among lower risk youth. Thus, experiencing an arrest aggravates subsequent delinquency among some but not all persons. Implications and directions for future research are identified.