Replication Data for: Understanding Spatial Regression Models from a Weighting Perspective in an Observational Study of Superfund Remediation

This data contains binary exposure and confounder data used in the preprint "Understanding Spatial Regression Models from a Weighting Perspective in an Observational Study of Superfund Remediation" (Woodward, Dominici, Zubizarreta). The final dataset, named buffers, is at the level of the Superfund site (n = 1429). This dataset can be accessed by loading preprocessed_superfunds.RData into R. The binary exposure data, describing whether a Superfund site was remediated and removed from the National Priorities List between 1991 and 2015, is derived from publicly-available data on Superfund site status (source: EPA ArcGIS). Sociodemographic confounder variables were obtained from the 1990 Decennial Census at the census tract level using NHGIS. Confounders related to chemicals, exposure media, and Site Score were obtained from publicly available Superfund site information and the EPA’s Contaminant of Concern document 406203. The R code used to curate this dataset directly from the publicly available data sources is provided (preprocessing.R).