Space and Gang Crime: Modeling Social Processes in the Spatial Autocorrelation Matrix

Local community levels of violence cannot be adequately explained by internal conditions such as the socio-economic composition of place. Instead, spatial analyses have demonstrated that violence in one area is dependent upon the level of violence in other areas.  The precise nature of the underlying processes through which places are linked, socially as well as geographically, remains an open question. The standard approach of conceptualizing "space" when modeling the impact of spatial effects crime has been to rely on contiguity or distance-based measures of space.  In recent years, some attention has been paid to moving beyond geographic space to be include also include measures of social similarity or the socio-spatial dynamics of urban street gangs. The current work briefly reviews these approaches and introduces a third method - the employment of block-modeling techniques which are commonly used to model social influence within the social network.  The current presentation will examine the spatial distribution of gang violence in Los Angeles and compare the results across these various methods and demonstrate that the choice of model is dependent upon one's theory of why space matters.

This presentation adopts a “social influence” modeling approach to understanding the spatial distribution of violence. This approach offers an important advance over previous efforts to model the spatial dynamics of violence because it requires a more careful consideration of the social processes driving spatial dependence.  In doing so, I combine the spatial methods with methods common in the social network literature.