In both industrialized and less industrialized countries, persons of higher socioeconomic status (SES) live longer and have lower rates of most diseases than their less favored counterparts (Behm, 1980; Grosse and Auffrey, 1989; Holzer et al., 1986; Department of Health and Social Security, 1980). Some studies from less industrialized countries, such as a study of Nigerian civil servants (Markovic et al., 1998), have found a positive association between SES and chronic disease. This may reflect differences in the historical time period across societies in the secular distribution of disease. In the United States, for example, higher position in the SES hierarchy was associated with greater prevalence of heart disease earlier this century, but SES is currently inversely related to cardiovascular risk (Morgenstern, 1980). Thus, apparently discrepant findings highlight the importance of attending to broader social and historical contexts.

Of particular importance is a gradient in the relationship between SES and health: each level of the hierarchy exhibits less morbidity and mortality than lower levels (Adler et al., 1994, 1999; Marmot et al., 1991). Most of the evidence supporting this relationship derives from European and North American populations, where the data are consistent and robust. At the same time, several studies document that the gradient is nonetheless characterized by a threshold, usually around the median for income, where additional increases in SES have a diminished effect in reducing morbidity and mortality rates (Kitagawa and Hauser, 1973; Pappas et al., 1993; McDonough et al., 1997; Wilkinson, 1986). Research is needed to provide greater understanding of the conditions under which particular markers of SES manifest patterns of linear or nonlinear associations with health status. We need to identify the thresholds after which weaker SES effects are observed and to characterize the social, psychological, and material risks and resources that are associated with each level of the SES hierarchy.

A growing body of research also reveals that even though overall mortality rates have been declining, socioeconomic differentials in mortality have been widening in recent decades. Comparing data from the 1960s to those for the late 1970s and 1980s, U.S. studies reveal that income and educational differentials have widened over time (Duleep, 1989; Pappas et al., 1993; Feldman et al., 1989). Similarly, widening socioeconomic differentials in mortality have been observed in England, Wales, France, Finland, Norway, and the Netherlands (Department of Health and Social Security, 1980; Kunst and Mackenbach, 1994; Mackenbach et al., 1989). Widening health disparities appear to be primarily driven by larger improvements in the health of high-SES groups compared to their lower-SES counterparts. For some health conditions, however, there has been no change in health or worsening health status over time for economically disadvantaged populations (Williams and Collins, 1995). Although differences between SES groups in access, utilization, and the quality of medical care probably play