We use high resolution data to investigate the association between crime incidence and proximity to different types of public schools over the past fifteen years in the city of Philadelphia. We employ two statistical methods, regression modeling and propensity score matching, in order to better isolate the association between crime and school proximity while controlling for the demographic, economic, land use and disorder characteristics of the surrounding neighborhood. With both of these approaches, we find significantly increased crime incidence near to public schools regardless of crime outcome, educational level and time period. The effect of school proximity on crime varies substantially depending on whether or not school is in session, as well as between different types of crime and educational levels of the school. We see the largest effects of school proximity on crime for violent crimes near to high schools during their in-session time periods. Our results support several theories which suggest that crime should be elevated near to schools, as well as finding significant associations between crime and other aspects of the built environment.

翻译：我们利用高分辨率数据，调查过去十五年间费城犯罪发生率与不同类型公立学校邻近性之间的关联。采用回归建模和倾向得分匹配两种统计方法，通过控制周边社区的人口、经济、土地利用及失序特征，更好地分离犯罪与学校邻近性的关联。两种方法均发现：无论犯罪类型、教育阶段还是时间段，公立学校附近犯罪发生率均显著升高。学校邻近性对犯罪的影响在学期内与学期外存在显著差异，且因犯罪类型及学校阶段不同而有所变化。研究观测到最大效应出现在高中学期内，暴力犯罪受学校邻近性影响最为显著。研究结果支持多项理论——即学校周边犯罪率应高于一般区域，同时证实犯罪与其他建成环境要素存在显著关联。