Fig. 5: Coefficient plot of Clean Water Act sensitivity test results.
From: The widespread and unjust drinking water and clean water crisis in the United States
Descriptive regression model results. Different colors for plotted coefficients represent separate blocks of variables. Models are linear probability models with state fixed effects and Huber/White/Sandwich cluster-robust standard errors at the state level. All tests are two-tailed. Dots indicate point estimates and lines represent 95% confidence intervals. Models predicted whether or not there were greater than 1% of Clean Water Act permittees being considered in Significant Noncompliance in the county. First model excludes counties in states with CWA data issues (N = 2261), second model includes all counties reported by the EPA (N = 3206), third model duplicates counties in the top and bottom 10% of CWA Significant Noncompliance within states without data issues (N = 3151). Full model results, confidence intervals, and exact p values available in SI.
