A Study of Criminal Social Networks in the Federal Justice System

• In Event: Application of the Current Federal Justice Statistics Program (FJSP) to Guide Current Criminal Justice Policy

Abstract

Academic studies have shown that criminal social networks can affect homicide victimization and juvenile recidivism. Understanding criminal network structures and the effects of networks on recidivism could lead to more effective interventions for criminals on probation, better strategies for treating juvenile offenders, and improved sentencing approaches for co-defendants.  This study describes the structure of criminal connections using data from the Federal Justice Statistics Program (FJSP).  To date, all studies of which we are aware focus on particular criminal organizations, gangs, schools, neighborhoods, or states. Using data from the FJSP will allow us to generalize beyond these specific groups to other contexts.  We will use social network analysis (SNA) techniques to answer questions such as:  How many other federal defendants does a typical federal defendant know?  How many second-degree connections (“friends-of-friends”) does a typical defendant have in the FJSP?  Which individuals are important focal points within the network?  How large are connected groups of defendants?  How does the structure of criminal networks vary across types of crimes, in different geographic areas, and over time? To answer these questions, we will begin with the universe of individuals included in the FJSP dataset.  Each of these individuals represents a unique node in the network.  We will then identify connections between nodes based on whether two individuals appear as co-defendants in an investigation or trial.  Next, we will measure the intensity (or strength) of the connections between two individuals based on characteristics of the case in which they were both involved.  Finally, we will use this network model to answer the questions above. Although this analysis will describe only connections observable through the FJSP data, the results will still provide an important window into how criminal networks are structured in the United States.

Authors

• Matthew Ranson, Abt Associates Inc.

• Sharmini Radakrishnan, Abt Associates Inc.

• Holly Swan, Abt Associates Inc.