Segregation and Local Crime: An Integrated Spatial Analysis

美国城市地区的特点是社会群体在地理空间上分布不均，这体现在普遍存在的种族、民族和经济地位的居住隔离水平上。对于弱势群体和少数族裔居民高度集中的社区来说，这种隔离与犯罪和其他有害条件有关，但对优势群体和白人社区却有好处。这些隔离的不同结果被解释为少数民族社区的空间隔离通过制造和加强社会和结构隔离而产生社会问题的证据。然而，大多数犯罪研究依赖的措施没有考虑到隔离区域相对于其他区域的位置。因此，它们没有回答隔离的固有空间性质如何与犯罪活动的地理分布联系起来的问题。本研究借鉴了对种族隔离及其后果的研究、对犯罪地域差异的分析、以及对总体和多层犯罪的研究，开发了一个将空间引入犯罪分析的模型。我们的大问题是：种族、民族和相关的经济隔离如何影响犯罪的地理分布？回答这个一般性问题需要解决四个具体目标：(1)使用最好的地理信息系统（GIS）数据和技术，开发更精细的基于空间的局部隔离（沿着各种维度）区域措施；(2)确定当地社区物质基础设施的哪些方面会影响犯罪；(3)确定城市内不同街区对附近地区犯罪的影响；(4)评估地方隔离对不同城市社区犯罪的影响。因此，这项研究的学术价值在于，通过实现这些目标，它将揭示城市范围内的种族隔离、地方种族隔离和社区犯罪之间的联系。该项目将分析2000年前后美国10个大城市的数据：奥斯汀、波士顿、芝加哥、哥伦布、沃斯堡、杰克逊维尔、密尔沃基、俄克拉荷马城、凤凰城和波特兰。这10个城市涵盖了一系列的黑人-白人隔离，结合了地区差异，并有相对较多的西班牙裔和/或黑人人口。国家社区犯罪研究（NNCS）数据将与基于地理信息系统（GIS）的地方隔离措施、区域物理和生态特征相结合，这些数据来自地理信息系统数据库中的数字交通网络、地块和建筑层面的土地使用，以及其他公共资源。最先进的GIS分析和层次贝叶斯空间统计建模技术将用于评估我们的理论模型。该建模策略旨在突出不同类型的地方隔离（如种族、西班牙裔、经济地位）对犯罪的实质性空间影响，同时控制由未测量条件引起的犯罪的剩余空间依赖性。分层贝叶斯方法还允许以综合的方式探索城市内部和城市之间的实质性问题，从而解释剩余的空间依赖性。更广泛的影响。实际上，这项研究将阐明一个主要的社会进程（即隔离）如何助长城市内不同群体之间犯罪流行的不平等。除了其实质性影响之外，这项工作将：(1)整合三个不同领域关于社区犯罪来源的思考，为未来在这一重要政策主题上的跨学科工作奠定基础；(2)制定方法学战略，作为各个领域寻求整合多种类型空间数据和概念的研究模型；(3)向科学界提供基于地理信息系统（gis）的新改进的局部隔离和邻里物理和生态特征测量，作为NNCS的补充；(4)为有色人种和经济背景不利的本科生和研究生提供有吸引力的主题和多样化的导师，以提高他们进行研究的能力。

SES- 0528232Ruth PetersonChristopher BrowningCatherine CalderLauren KrivoMei-Po KwanOhio State University Urban areas in the U.S. are characterized by the uneven distribution of social groups across geographic space as evidenced in pervasive levels of residential segregation by race, ethnicity, and economic status. This segregation is associated with crime and other deleterious conditions for communities with high concentrations of disadvantaged and minority residents, but carries benefits for more advantaged and White localities. These differential outcomes of segregation have been interpreted as evidence that the spatial isolation of minority neighborhoods produces social problems by creating and reinforcing social and structural isolation. However, most crime studies rely on measures that do not take into account the location of segregated areas relative to one another. As such, they leave unanswered the question of how the inherently spatial nature of segregation is connected with the geographic distribution of criminal activity. This research draws on studies of segregation and its consequences, analyses of geographic variation in crime, and aggregate and multilevel crime research to develop a model that brings space into the analysis of crime. Our broad question is: How does racial, ethnic, and associated economic segregation affect the geographic distribution of crime? Answering this general question entails addressing four specific objectives: (1) developing more refined spatially-based measures of locally segregated (along a variety of dimensions) areas using the best Geographical Information Systems (GIS) data and techniques; (2) identifying what aspects of the physical infrastructure of local communities affect crime; (3) determining how the effects on crime of nearby localities vary across neighborhoods within cities; and (4) evaluating how the effects of local segregation on neighborhood crime varies across cities. Thus, the intellectual merit of the study consists of the light it will shed on the linkages among city-wide segregation, local segregation, and neighborhood crime through meeting these objectives. The project will analyze data for 10 large U.S. cities for circa 2000: Austin, Boston,Chicago, Columbus, Fort Worth, Jacksonville, Milwaukee, Oklahoma City, Phoenix, and Portland. The 10 cities cover a range of Black-White segregation, incorporate regional variation, and have relatively large Hispanic and/or Black populations. National Neighborhood Crime Study (NNCS) data will be combined with GIS-based measures of local segregation and physical and ecological characteristics of areas derived from GIS databases on digital transportation networks and parcel- and building-level land use, and other public sources. State-of-the-art GIS analyses and hierarchical Bayesian spatial statistical modeling techniques will be used to evaluate our theoretical model. The modeling strategy is designed to highlight the substantive spatial effects of different types of local segregation (e.g., by race, Hispanic origin, economic status) on crime, while controlling for residual spatial dependence in crime caused by unmeasured conditions. The hierarchical Bayesian approach also permits the within-city and between-city substantive issues to be explored in an integrated manner that accounts for residual spatial dependence. Broader Impacts. Substantively, the study will shed light on how a major social process (i.e., segregation) fosters inequality in the prevalence of crime among diverse groups within cities. Beyond its substantive impact, this work will: (1) integrate thinking in three different fields about the sources of neighborhood crime, laying a foundation for future advances in interdisciplinary work on this important policy topic; (2) develop methodological strategies that will serve as models for research in a variety of fields seeking to integrate multiple types of spatial data and concepts; (3) make available to the scientific community newly refined GIS-based measures of local segregation and neighborhood physical and ecological characteristics as supplements to the NNCS; and, (4) provide an attractive topic and diverse mentors for undergraduate and graduate students of color and from economically disadvantaged backgrounds seeking to enhance their ability to conduct research.