Using geographically weighted regression for environmental justice analysis: Cumulative cancer risks from air toxics in Florida

Previous quantitative research on environmental justice has been limited by simplistic assumptions used to measure health risks and traditional regression techniques that fail to discern spatial variations in statistical relationships. We address these gaps through a case study that examines: (a) whether potential health risks from exposure to hazardous air pollutants in Florida are related to race/ethnicity and socioeconomic status, and (b) how the significance of statistical associations between health risk and race/ethnicity or socioeconomic status vary across the state. This study integrates census tract level estimates of cumulative cancer risk compiled by the EPA with Census 2000 data and a spatial statistical technique known as geographically weighted regression that allows us to explore spatial variability in analytical results. Our findings indicate that while race and ethnicity are significantly related to cancer risks in Florida, conventional regression can hide important local variations in statistical relationships relevant to environmental justice analysis.