It is difficult now to reconstruct the logic that led us to believe that the nature of human relationships—the degree to which an individual is interconnected and embedded in a community—is vital to an individual's health and well-being as well as to the health and vitality of entire populations. In retrospect, a combination of observations and reading the rich theoretical literature on social integration, attachment, and social networks led us to test these ideas empirically. Now, almost 25 years after John Cassel (1976), Sidney Cobb (1976), and other seminal thinkers in social epidemiology suggested that this was a critical area of investigation, and 20 years after the earliest studies in Alameda County, California; Tecumseh, Michigan; and North Carolina revealed the influence of social relationships on mortality (Berkman and Syme 1979; Blazer 1982; House et al. 1982), it is time to take stock of the vast literature on this topic. Our aim is to revisit some of the seminal theories that have guided empirical work, revise and reformulate some of those ideas, and point the way toward productive lines of inquiry for the future.

When investigators write about the impact of social relationships on health, many terms are used loosely and interchangeably, including social networks, social support, social ties, and social integration. A major aim of this chapter is to clarify these terms under a single framework. We discuss (1) theoretical orientations from diverse disciplines that we believe are fundamental to advancing research in this area,(2) an overarching model that integrates multilevel phenomena,(3) a set of definitions accompanied by major assessment tools,(4) some of the strongest findings linking social networks or support to morbidity, mortality, or functioning, and finally,(5) a series of recommendations for future work. Since there are now numerous books and literature reviews on networks, support, and health (Cohen and Syme 1985; Broadhead et al. 1983; House et al. 1988; Sarason et al. 1990; Thoits 1995; Berkman 1985, 1995; Seeman 1996), our aim is not to be all-in-