CCTV의 범죄예방효과 분석

: 연속적 실험처치(continuous treatment)를 위한 용량반응 모형(dose-response model)의 적용

Examining the census output areas of Gwanak-gu, Seoul, South Korea, this study analyzed crime deterrence effects of CCTV in the application of a dose-response model proposed by Cerulli(2015). This study observed the number of CCTV installed in the output areas as a zero-inflated count variable, which cannot be fully estimated with traditional propensity score or dose-response models. On contrary to the traditional models, which assume the treatment variable as either binary or normal, the revised dose-response model can solve the zero-inflated and non-normal distribution issues in the treatment variable. Using the revised dose-response model, this study found that a CCTV installed in an output area is estimated to decrease 1.23 cases of theft. On the other hands, this study did not find any statistically significant effect of CCTV on violent crime. Moreover, this study estimated that the decrease in the total population and the increase in the alcohol outlet density are estimated to positively impact the number of theft. Based on these results, this study suggested that customized crime prevention activities, in particular, the installation of CCTVs in the hotspot of theft, should be implemented in each output area.