Using spatial econometrics and GIS to analyze robbery in medium-sized cities

Some criminological theories look at the place, space and/or location where crime occurs. Several of these theories are geographic in nature and are reviewed and extended in this dissertation. Social disorganization theory notes that the social structure of a neighborhood influences the crime rate. Neighborhoods or communities that are disorganized will have weak informal social control or 'collective efficacy' that does not deter crime from occurring (Sampson and Raudenbush, 1999). The routine activities theory assumes that there is a motivated offender, attractive target and the absence of a capable guardian that come together in space (Felson, 1999). The physical structure of a neighborhood influences the daily activities of people occupying the space, which in turn influences the crime rate. The broken windows hypothesis/theory states that incivilities and minor disorders, if left unchecked, will serve as a cue that residents do not care about the neighborhood, and will influence crime (Wilson and Kelling, 1982). Each of these theories tackles one or more of the social, physical and behavioral dimensions of a neighborhood. However, most research has not incorporated the most current methods of spatial analysis for neighborhood based analysis of crime. With few exceptions, much of the 'spatial analysis' relies on multivariate regression analysis with units of analysis such as neighborhoods (defined as either census tracts or block groups) and does not include the influence of space through the dependency or heterogeneity of the neighborhoods. This dissertation examines the use of spatial econometrics in conjunction with geographically-based criminological theories of crime. The crime used for this dissertation is robbery, and the study areas are three mid-sized cities, Albany, NY, Dayton OH, and Hartford, CT. The data for this research is obtained from public sources (local and regional governmental agencies, US Census) and is also collected as primary data in the field. The unit of analysis is the census block group and a multivariate regression analysis incorporating spatial econometrics is used. The results from this dissertation demonstrate that there are three environmental dimensions along with spatial effects that influence the occurrence and distribution of robbery.