these and other relevant disciplines should be developed to address specific public health problems. It is important that research concerning environmental justice use both traditional and nontraditional methodologies to best serve the communities of concern. These include creating new epidemiologic tools to assess small populations better, addressing race and relevant socioeconomic considerations in the analysis, and involving the community in every stage of the research through participatory research.
Questions of environmental justice tend to be raised on behalf of relatively small populations. Many of the problems associated with the study of small populations, such as minority and economically disadvantaged individuals, have to do with the fact that study size requirements, (in addition to other requirements, such as isolation of exposures) for traditional epidemiologic studies can rarely be met. Populations that are (or that are believed to be) dealing with illness as a result of a high level of exposure to environmental health hazards are often isolated, either in urban or in rural areas, and are typically subject to other factors that affect health and well-being. Moreover, environmental and health data for populations whose health may be affected because of exposure to environmental health hazards are not routinely collected or analyzed by demographic categories (Environmental Protection Agency, 1992).
Various socioeconomic factors are associated with different rates of morbidity and mortality among different racial or ethnic groups (National Center for Health Statistics, 1998b; Warren, 1993). To the extent that they are correlated with environmental exposures, these socioeconomic factors could be considered confounders (in the epidemiologic sense) of the relationships between the environmental exposures and the disease outcomes. Typically, even in the event that they are measured in environmental epidemiologic studies, the studies' designers attempt to control for these potential "confounders" statistically to assess whether an association exists between the disease and the exposure variables of primary interest.
Several investigators have commented on reasons for the paucity of data on the roles of race, ethnicity, and other socioeconomic factors in studies of occupational diseases (Friedman-Jiménez, 1994; Kipen et al., 1991; Zahm et al., 1994). Some research deliberately excludes analysis of differential disease occurrences in minority workers because the small number of minority subjects would have provided an unacceptably low statistical power to test the primary hypotheses of the study. This is unfortunate because the small body of published occupational health literature that does explicitly include data from studies with these populations suggests that racial, ethnic, and economic disparities continue to influence the risk for adverse health effects due to environmental hazards in the workplace.
Research is needed to improve the capacity of epidemiologic studies to detect adverse health impacts in small populations and to evaluate clusters of effects