Having concluded that there are clearly disparities in health status and that there are significant disparities in exposure to environmental hazards, the next step in analyzing the issue of environmental justice is to determine whether there are causal relationships between exposure to these hazards and health outcomes and whether the disparities in health status can be attributed to the disparities in exposure. For many of the identified potential environmental hazards, the lack of published research makes it difficult to draw a strong conclusion that disparate exposures result in disparate health outcomes.

The difficulties encountered in this type of research are illustrated in the work of Baden and colleagues (1996), who tried to examine both the interactions of multiple environmental hazards and the possible connections of these hazards to adverse health outcomes. They included in their analysis the siting of several possible environmental hazards in the Chicago metropolitan area: the large areas of waste regulated under the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (the Superfund law), waste-generating sites regulated by the Resource Conservation and Recovery Act, older existing waste disposal sites such as landfills, and emissions from highway exhaust. They conducted regression analyses of census-tract data examining the relationships between proximity to waste disposal sites and poor infant health. The research was unable to account for "causal links between hazardous-waste and ill health effects, … occupational exposure to toxins, parental bioaccumulation of toxins, [or] medically risky behavior such as drinking and use of illegal drugs while pregnant" (Baden et al., 1996, p. 25). The researchers also noted that "use of the census tract for analysis may introduce some aggregation bias" (p. 25). Within these constraints, they concluded that "there was no relationship between the presence of waste sites and low birth weight for children born in Chicago in 1990. … Neither was there one single type of waste site associated with adverse health outcomes, nor did a combination of all of the waste sites (insignificant in themselves) combine to produce adverse infant health outcomes" (p. 26). The researchers' final conclusion was, ''Future research must address these issues" (p. 25). Ecologic studies such as these are subject to a variety of potential biases in addition to the aggregation bias noted by the authors, and are not by themselves adequate for making or excluding causal connections (Greenland and Robins, 1994). In spite of the general lack of published research linking disparate exposures to disparate health outcomes, some well documented links do exist. One such study used field epidemiology methods combined with a prevention intervention trial to document definitively a link between disparate exposure to dimethylformamide and disparate prevalence of toxic liver disease (Friedman-Jiménez and Claudio, 1998; Redlich et al., 1988).