Ecological Networks and Urban Crime: A New Approach to Neighborhood Effects on Offending

• In Event: 533 - Section on Crime, Law, and Deviance Invited Session. Social Networks and Crime

Abstract

Despite decades of research, our knowledge of why and how neighborhoods influence local crime rates is still emerging. We hypothesize that the aggregate network structure of shared local exposure among neighborhood residents is consequential for crime. Neighborhoods in which residents intersect in space more extensively in the course of conventional routine activities will exhibit higher levels of familiarity, beneficial (“weak”) ties, trust, collective efficacy, and more consistent monitoring of public space. We introduce the concept of ecological networks – two-mode networks capturing links between residents and activity locations – to formalize our hypotheses. Specifically, we expect a number of ecological network structural characteristics (e.g., density, centralization, and local clustering) will be negatively associated with crime. We use unique travel diary data from the 1999 Mid-Ohio Area Household Travel Survey on all members of 2,705 households (including over 20,000 trip destinations) to construct ecological networks for Columbus, OH and extract network structural properties by census tract. We then examine associations between ecological network characteristics and geo-coded administrative data on crime in Columbus. We employ multilevel, spatial autoregressive models of neighborhood crime rates adjusting for spatial and ecological network autocorrelation. These analyses hold the potential to enrich insight into the spatially anchored network origins of neighborhood crime control.

Authors

• Christopher R. Browning, The Ohio State University

• Jonathan Dirlam, The Ohio State University

• Bethany Boettner, The Ohio State University