Award: National Institute of Justice contract 2006-IJ-CX-0022

Title: Investigating the Simultaneous Effects of Individual, Program and Neighborhood Attributes on Juvenile Recidivism Using GIS and Spatial Data Mining

Investigators: Philip Harris, Zoran Obradovic, Alan Izenman, and Jeremy Mennis
Temple University

Specific Aims:

We propose to investigate the coaction of neighborhood, program, and individual effects on juvenile recidivism using geographic information systems (GIS) and spatial data mining. Unlike conventional analytical approaches in the social sciences, spatial data mining is inductive and is intended to facilitate data exploration by domain experts. Exploration of these spatial patterns drives hypotheses as to why the observed spatial variation occurs, or in other words, why does one place differ from another? GIS provides the technology to integrate diverse spatial data sets, quantify spatial relationships, and visualize the results of spatial analysis. In the context of juvenile recidivism, this approach will facilitate the investigation of how, and why, recidivism rates vary from place to place, through different programs, and among individuals. Our methodology will be compared to more conventional statistical methods for purposes of evaluation.

The proposed study will apply spatial data mining to the analysis of adjudicated juvenile delinquents assigned to court-ordered programs by the Family Court of Philadelphia, Pennsylvania. This population encompasses all adjudicated delinquents committed to programs by the court during the years 1996 to 2002 - more than 26,000 cases. The proposed study makes use a population of relevant youths, a population of programs, and a population of neighborhoods. The analytical challenge we address is the complexity of relationships among individuals, programs and neighborhoods, where individuals are nested in neighborhoods, the individuals are nested in programs, and the programs are nested in neighborhoods.

Our research questions are:

1. How do individual youth and family characteristics, program characteristics, and neighborhood characteristics interact to produce specific program outcomes, such as recidivism and placement in a more secure facility?
2. Why is recidivism more common in some neighborhoods than others?
3. Why is drug reoffending more common in some neighborhoods than others?
4. To what extent is program impact a function of (constrained and enhanced by) neighborhood forces?