The current study examines the social, racial, and economic determinants of homicide rates in the City of St. Louis using census tracts and 1000-m raster grids as geographic units for analysis. We used a meso and micro-level geographic scale to evaluate potential impacts associated with the modifiable areal unit problem. Using spatial interpolation, we generated redistributed values, based on area weighting, for explanatory variables. We applied spatial dependence regression models to assess the relationship with social, economic, and racial determinants in the analysis of homicide rates. At the census tract level, we do not find significant relationships between social, racial, and economic explanatory variables and homicide rates. Significant relationships are observed at the 1000-m grid level with income inequality, public assistance, and racial diversity and homicide rates. The results indicate a trade-off between the explanatory power of the census tract and the precision of the 1000-m raster grid.

中文翻译：

---

圣路易斯凶杀案的空间分析：规模的重要性

当前的研究使用人口普查区和1000米栅格栅格作为地理单位来分析圣路易斯市凶杀率的社会，种族和经济决定因素。我们使用了中观和微观的地理尺度来评估与可修改的面积单位问题相关的潜在影响。使用空间插值，我们基于面积权重为解释变量生成了重新分配的值。我们使用空间依赖性回归模型来评估凶杀率分析中与社会，经济和种族决定因素的关系。在人口普查领域，我们没有发现社会，种族和经济解释变量与凶杀率之间存在显着关系。在1000米网格级别观察到与收入不平等，公共援助，种族多样性和凶杀率。结果表明，在人口普查区域的解释能力和1000米栅格网格的精度之间要进行权衡。