Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Gabriel P. Kuperminc

Second Advisor

Wing Yi Chan

Third Advisor

Dominic Parrott

Fourth Advisor

Robert Latzman


Substance abuse treatment programs are a promising approach to reducing criminal behavior. However, these programs are not equally e↵ective for all participants. Research shows that individuals at high risk for criminal recidivism usually benefit from high-intensity treatment programs, while those at lower risk benefit more from low-intensity, community-based approaches. In addition, there is evidence that individuals at moderate risk for criminal recidivism can actually be harmed by high-intensity treatment programs, leading to higher than expected rates of recidivism. In addition to overall risk for recidivism, other factors such as drug of choice, substance abuse severity, and psychological problems are known to co-occur in ways that are associated with different patterns of criminal behavior, and may influence how participants respond to treatment. Because of this high degree of co-occurrence of psychopathology (characterized by criminality, externalizing behaviors such as aggression and inattention, and internalizing problems such as depression and anxiety) with substance use, it is possible that low intensity interventions that are e↵ective in reducing substance use among individuals at moderate substance abuse risk could also be e↵ective in reducing criminal recidivism, at least for some subgroups. However, despite e↵orts in the criminal justice field to tailor interventions to levels of risk and characteristics of individuals that influence responsiveness to treatment, interactions between individual and program characteristics are rarely tested in evaluations of intervention programs, due in part to methodological challenges that arise when testing them using variable-based approaches. Some of these challenges can be overcome by using person-centered approaches, which allow for comparisons to be made across classes of individuals that di↵er on several variables simultaneously. Research on the ecological predictors of crime, such as neighborhood disadvantage and alcohol outlet density, has also been limited by a lack of attention to interactions. These neighborhood-level factors are strong predictors of criminal behavior, yet little is known about whether they a↵ect treatment outcomes. The present study used person-centered and ecological approaches to explore whether a moderate risk sample of clients receiving treatment for alcohol and/or drug abuse demonstrated lower levels of criminal and other externalizing behavior following treatment, and if so, whether their outcomes differed depending on individual and neighborhood-level characteristics.