Association for Public Policy Analysis & Management

There has been a renewed interest in understanding prison experiences as researchers and policy makers address the “age of mass incarceration” (National Research Council, 2014). One key element of the prison experience is the maintenance of social bonds and relationships to outside individuals during the period of incarceration. Recent prison visit and prisoner social network studies have explored the importance of social relationships on in-prison misconduct and health, as well as post-release outcomes. Theoretically social networks and their maintenance have relevance for prisoner success both during and after release. For example, external social networks may provide instrumental (e.g., access to a job or housing upon release) or expressive support (e.g., source of emotional encouragement) (Lin, 1986). As such, policy audiences would benefit from knowing which prisoners have quality social supports and which may require additional services or intervention. 

The present study will explore the bonds inmates have with individuals outside of the facilities. Detailed longitudinal visitation records will be used to model the heterogeneous forms of prisoner external egocentric social networks that are present in a statewide prison system. The data consist of New York State Department of Corrections and Community Supervision (DOCCS) electronic visitation records for all facilities between 2000 and 2013 combined with prisoner-level data. Analyses will be restricted to those prisoners for which complete admissions, releases, and transfer history data are available. Depending upon computational resources, random sub-samples may be selected for modeling. Statewide criminal history records from the Division of Criminal Justice Services (DCJS) will be used to examine pre-incarceration criminal records, as well as post-release recidivism.

Specifically, four research questions will be examined. 1) How do prisoner egocentric social networks vary in terms of size and composition in a diverse statewide prison system? 2) Are these variations correlated with prisoner attributes such as demographics or length of criminal record? 3) Are these variations correlated with prison policy characteristics such as length of stay or distance from home? 4) Are these variations correlated with prisoner outcomes? Although not causal, patterns observed in the relationships between these individual and policy variables and various types of prisoner-visitor networks may provide insight into future areas of inquiry. Social network analysis is premised on the inference that network perspectives may provide important insights on behavior (Knoke & Yang, 2008) and, therefore, examining network types against prisoner behavior may identify areas for future causal studies.