공간패널모형을 활용한 절도 및 폭력범죄와 사회구조적 변인 간의 관계 분석

This study examined the relationship between crime and social structural variables using a spatial panel dataset collected in Y district, X city from 2015 to 2017. A volume of study has analyzed such a relationship with different statistical tools but failed to consider both cross-sectional and longitudinal features inherent in criminal phenomena. Even panel data analysis models considering both aspects simultaneously failed to consider spatial dependence inherent in the spatial panel dataset. Thus, among various spatial panel data models, this study applied a ‘dynamic spatial Durbin model’ that simultaneously considers both spatial and temporal dependence within variables. In this process, this study analyzed the discriminatory effects of social structural variables on the two types of crime, theft and violent crime. As a result, this study found that each type of crime is temporally and spatially dependent, but the spatial dependence is partially or completely offset by temporal dependence. This indicates that crime in a specific time and space is influenced by crimes in the same space, but with different time rather than crimes in adjacent areas. Moreover, this study estimated that theft decreases as CCTV density, alcohol outlet density, and multi-family housing increase whereas violent crime decreases as population density and multi-family housing increase, and single-household decreases. Based on these results, this study suggested that customized crime prevention activities, such as the installation of CCTVs in the spatial and temporal hotspot of theft, be implemented depending on crime types and socioeconomic conditions.