基于多层次模型的广州门禁小区家庭入室盗窃环境影响分析

基于环境设计预防犯罪理论（CPTED）和社会失序理论，利用2012年10―12月在广州门禁小区进行的大规模入户调查数据，采用多层次回归模型从社区内部环境角度对广州门禁小区家庭的入室盗窃受害情况及其发生的环境机理进行研究。通过5个多层次回归模型的对比分析发现：家庭被盗情况主要受到家庭的住房特征和人口特征的影响，小区变量对家庭变量有非常重要的调节作用。住房特征、公共空间使用率、家庭人口特征和小区的正式控制机构失效度对家庭是否被盗有着非常显著的影响；环境形象和人口流动性、小区开放度和功能混合度通过对家庭特征的间接影响而影响家庭的被盗情况。

Environmental Impacts on Burglary Victimization in Gated Communities: A Multi-level Analysis in Guangzhou

Based on large-scale household survey conducted during the period from October to December, 2012, in Guangzhou, the authors studied burglary victimization and its environmental leading factors in gated community. The research variables selection of this paper is under the guidance of the CPTED (crime prevention through environmental design) and social disorganization theory. From the perspective of community environment, the paper explores the impact of environmental factors on household burglary victimization. On the basis of literature review, this paper first introduced the core elements and main ideas of CPTED theory and social disorder theory, which are broadly accepted in community crime research. Under the guidance of these elements and ideas, we finished an investigation of gated community in Guangzhou, China. Related social, economic and environmental factors are collected. Principal component analysis and cluster analysis are performed to refine the features of these selected factors. Three main components at community level and four main components at family level were extracted. To check community environment and social-economic factor’s effects on household burglary condition，hierarchical logistic regression was performed to analyze the cross scale effects between community and family factors on family burglary condition. The results of the five multi-level regression models suggest that both the community environment and family housing characteristics have significant impact on residential burglary victimization. Family burglary victimization is mostly affected by family housing conditions and demographic characteristics, while community variables have a very important role in regulating family factor’s effect. Regression results show that housing conditions, public space utilization rate and family population characteristics have significant influence on whether a house is burglarized. Furthermore, neighborhood formal control mechanisms, community environmental image, residence turnover rate, community openness and degree of function mixing through indirect effects of family characteristics have influence on household burglary victimization.