populations than communities with no commercial hazardous-waste facilities. In the United States in 1993, for example, the percentage of people of color (defined as everyone except non-Hispanic whites) was 14.4 percent in zip code areas with no commercial hazardous-waste facilities, 29.5 percent in areas with one facility, and 45.6 percent in areas with three or more facilities, an incinerator, or a large landfill. A similar trend had been evident in 1980. Communities with lower per capita incomes were also more likely to be situated near commercial hazardous-waste sites as well. The original study was one of the important factors in motivating a substantial response to the environmental justice issue from the federal government.

Another national study of hazardous-waste sites used census tracts as the unit of analysis as opposed to zip codes (Anderton et al., 1994). Anderton and colleagues examined the differences in race, class, and economic indicators between the populations in census tracts with treatment, storage, and disposal facilities (TSDFs) and those without TSDFs. Although they found no difference in the mean percentage of the population that was African American, that had incomes below the poverty level, or that was on welfare, they did find differences in the percentages of the population that were Hispanic (9.4 percent in census tracts with one or more TSDFs and 7.74 percent in those with none) and that were employed in manufacturing (38.6 and 30.6 percent, respectively). Additionally, Anderton and colleagues (1994) found that the differences became more noticeable in the areas surrounding the census tracts with TSDFs—that is, tracts with TSDFs and ''other tracts that have at least 50 percent of their area within a 4-kilometer (2.5-mile) radius of the center of a tract containing a TSDF" (Brown, 1995, p. 18). Using that level of analysis, Anderton and colleagues found that populations of the areas surrounding TSDFs have higher mean percentages of African Americans (24.7 percent, compared with 13.6 percent in census tracts outside the larger unit of analysis), of Hispanics, (10.7 and