Objectives <p>Review and assess the primary metrics used in the crime and place literature to measure concentrations of crime, discussing their utility and limitations. We also propose a new metric that meets the criterion of Ockham’s razor, and is based on the 50-X metric proposed by Weisburd (2015) in identifying the law of crime concentration—the percentage of street segments that account for 50% of crime.</p> Methods <p>Calculations of metrics using simulations, theoretical examples, and empirical examples are used to compare 7 different metrics of crime concentration: our new metric, the 50-X metric, the 5-X metric (Hipp and Kim 2017), the adjusted 50-X metric (Andresen and Malleson 2011), the Generalized Gini coefficient and the standard Gini coefficient (Bernasco and Steenbeek 2017), and the marginal crime concentration metric (Chalfin et al. 2021).</p> Results <p>We show that different metrics can provide very different information regarding crime concentrations, and some of these metrics are quite sensitive to sample size issues, computational complexity, and interpretability. A new metric, the generalized 50-X measure, is developed that addresses the issue of small Ns of crimes relative to streets, and that is easy to compute and interpret.</p> Conclusions <p>The utility of different measures of concentrations relates both to the principle of Ockham’s razor, and to the specific goals of the metric. We argue that the generalized 50-X measure is consistent with policy interest in hot spots of crime, and that it adheres to the principle of Ockham’s razor.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Ockham’s Razor and the Measurement of Crime Concentrations at Places

  • Martin A. Andresen,
  • David Weisburd

摘要

Objectives

Review and assess the primary metrics used in the crime and place literature to measure concentrations of crime, discussing their utility and limitations. We also propose a new metric that meets the criterion of Ockham’s razor, and is based on the 50-X metric proposed by Weisburd (2015) in identifying the law of crime concentration—the percentage of street segments that account for 50% of crime.

Methods

Calculations of metrics using simulations, theoretical examples, and empirical examples are used to compare 7 different metrics of crime concentration: our new metric, the 50-X metric, the 5-X metric (Hipp and Kim 2017), the adjusted 50-X metric (Andresen and Malleson 2011), the Generalized Gini coefficient and the standard Gini coefficient (Bernasco and Steenbeek 2017), and the marginal crime concentration metric (Chalfin et al. 2021).

Results

We show that different metrics can provide very different information regarding crime concentrations, and some of these metrics are quite sensitive to sample size issues, computational complexity, and interpretability. A new metric, the generalized 50-X measure, is developed that addresses the issue of small Ns of crimes relative to streets, and that is easy to compute and interpret.

Conclusions

The utility of different measures of concentrations relates both to the principle of Ockham’s razor, and to the specific goals of the metric. We argue that the generalized 50-X measure is consistent with policy interest in hot spots of crime, and that it adheres to the principle of Ockham’s razor.