Racial and Socioeconomic Fairness of Area-Level Traffic-Related Air Pollution Measure Aggregation.

Abstract

Background: The Environmental Protection Agency's Environmental Justice Screen traffic proximity (EJ Screen) and the Department of Transportation's Average Annual Daily Traffic (AADT) commonly serve as proxies of traffic-related pollution exposure. However, the methods used to aggregate to area-level measures have been untested for bias.

Methods: Using a parcel-level measured developed for Hamilton County, Ohio, agreement was determined with both above measures at three geographic levels: census block group, census tract and zip code tabulation area (ZCTA). Fairness was assessed using linear regression.

Results: Generally, the values of AADT were in significant agreement with the parcel proximity measure while the EJ Screen was not. Racial and community deprivation bias was widely detected for EJ Screen.

Discussion: While the biases detected were not directly against majority black and materially deprived neighborhoods, the biases could manifest in negative downstream effects. These manifestations include suppression of known traffic-related pollution effects in subsequent research.

Impact statement: The Environmental Protection Agency's Environmental Justice Screen traffic proximity (EJ Screen) and the Department of Transportation's Average Annual Daily Traffic (AADT) are widely used traffic related pollution proxies however, with common aggregation techniques largely untested for fairness, this research has detected potential biases in the EJ Screen product.