An important overview of theoretical problems about studies on socioeconomic aspects of population health. The authors point out that they all use data not originally collected for analysis and suggest socioeconomic status information should be routinely collected in all health studies. In a traditional prospective epidemiological cohort study, where exposure measures are obtained at the outset and related to relative probability of subsequent health events, this is not suited to understanding reciprocal ways in which wealth and health affect each other or the impact of these two factors on evolution and maintenance of risk factors for health. They suggest the need for longitudinal panel studies with repeated measures over time to sort out the causal texture of the relationship. Not much is known on the impact of socioeconomic factors over the different stages of the life course. There is no unanimity of the meaning of social cohesion or whether it is declining. Researchers need to be wary of romanticizing past eras when society "worked", periods or when it may have worked only for the dominant group (by race, religion, ethnicity, gender etc.). The health effects of social cohesiveness may vary widely depending on who is doing what to whom. There are studies suggesting community socioeconomic characteristics, independent of individual characteristics which are important to health outcomes, but studies do not really characterize communities except as geographic areas. Good measures of community life are needed in meaningfully defined communities for which there are measures of health outcomes. There is an urgent need to document the impact of economic and social policies on the health of populations. We know little about ways in which economic transitions, entry into and movement through the labor force, and income inequality influence health. For example what are the effects on health of consistent low incomes, or of large income reductions, or of parental social class influence on educational achievement or of movement into labor force, or changing occupations. They conclude that while it is tempting to focus on biological pathways that "explain" how inequalities in health are manifested in individuals, it is likely that only an understanding of economic, behavioral, social, psychological and community dynamics and their causes will lead us to remedies.