The Role of Community-level Conditions in Producing Street Segment Level Crime: An Exploration of Direct Effects, Moderating Effects, and Model Fit

• In Event: Crime and Place: Advancing Place Based Theory

Abstract

In this study, we test a multilevel theoretical model that incorporates measures of routine activities and social disorganization at the street level and measures of community structural conditions at the neighborhood level. The main purpose of the study is to determine whether community level conditions (1) have direct effects on street segment crime (2) moderate the effects of street segment conditions on crime and (3) improve overall model fit of street segment-only models. To test the multilevel theoretical model we constructed a multilevel dataset that nested 36,541 street segments in 451 census block groups in Portland, Oregon. To account for complexities of a multilevel spatial dataset, we estimated models using multilevel negative binomial regression with controls for spatial autocorrelation. Finally, we used likelihood ratio tests to explore how model fit changed when community level conditions were included in street segment level crime models.

Author

• Roderick W Jones, University of North Carolina at Wilmington