Network Mechanisms in a Prison-Based Therapeutic Community

? DESCRIPTION (provided by applicant): The proposed project will use theory and methods from the burgeoning area of network science to investigate the structure and dynamics of a Therapeutic Community (TC) alcohol and drug treatment unit in a Pennsylvania medium security men's prison. The project will focus on identifying and exploring the peer- related mechanisms thought to underlie TC effectiveness. Of particular interest are peer role-modeling and social influence processes assumed to promote successful treatment engagement and identity transformation. Specifying these system-based mechanisms will transform our understanding of prison drug treatment and create a tool for evaluating and improving TC operations. To understand TC peer processes, a multidisciplinary group of criminology, network, substance abuse, and criminal justice scholars will design and field monthly network and behavioral surveys to all inmates residing in a prison-based TC (N=60 inmates) for a nine-month observation period. Peer networks will be constructed using inmate nominations of friendship and respected peers in the unit. The resulting cross-sectional tie matrices will be analyzed to understand global characteristics of TC peer networks, including 1) how network status correlates with TC engagement and 2) determining if sociodemographic subgroups (e.g. by race, age, or criminal history) exist that potentially undermine TC treatment effectiveness. The project will use longitudinal network data to examine peer influence dynamics as related to TC engagement. Peer reinforcement (i.e., mutual self-aid) is presumed essential for effective treatment but has never been rigorously tested using a systems framework. The project will use sophisticated actor-based models of network and behavioral dynamics (i.e., SIENA) to examine if network status facilitates the diffusion of positive treatment attitudes and behaviors over time These analyses will also examine if changes in inmates' network positions correlate with patterns of sent and/or received peer affirmations and corrections. The latter will directly test i embeddedness in friendship/respect networks has intended effects of increased peer reinforcement and role-modeling behaviors. Finally, the project will examine whether TC inmates' network positions predict drug relapse, criminal recidivism, and incarceration one year following treatment. Time-to-event survival analyses will compare outcomes across TC graduates to identify potential structural sources of treatment success. In sum, this project will demonstrate the feasibility and contributions of a systems approach to prison-based drug treatment. Research products will help develop and evaluate treatment procedures that improve the health of drug-addicted prison inmates, a population of substantial policy interest. Results will have a sustained impact on theory and practice, while also forming the foundation for a future R01 proposal across multiple TCs.