Cohen and Wills (1985) distinguished between two models to explain the mechanisms by which social relationships influence health outcomes: (a) the main effects model and (b) the stress-buffering model. According to the stress-buffering model, social ties influence health outcomes only for individuals who happen to be experiencing stress - in other words, social resources buffer the individual against the deleterious effects of stress - whereas the main effects model posits that social relationships are beneficial regardless of the presence of stress. While these two models are not mutually exclusive, the emerging consensus in the field is that social networks operate via the main effects model, whereas social support is mobilized (and is most effective in promoting well-being) under stressful circumstances.
Berkman and Glass (2000) have gone further to identify a set of four distinct processes and mechanisms by which social networks exert their main effects on health outcomes. They are: (a) social influence over health-related behaviors, (b) social engagement, (c) exchange of social support, and (d) access to material resources. Social influence refers to the notion that our behaviors are influenced and regulated by others - an idea that harks back to Durkheim (1897). Socially more well-connected individuals tend to exhibit healthier profiles of lifestyle behaviors compared to socially isolated individuals, e.g., less smoking, better quality diet, more physical activity (Berkman and Syme, 1979; Eng et al, 2002; Kawachi et al, 1996). Social influence is particularly salient in marriage, which is for many people the most intimate of social ties. Longitudinal studies of marital transitions demonstrate the influence of marriage on health behaviors within spousal dyads. For example, when men become widowed or divorced, their alcohol consumption increases relative to men who stay married. Conversely, when widowed or divorced men become remarried, their alco hol consumption declines (Eng et al, 2005).2 Just the opposite is observed among women - that is, their level of drinking decreases when they become widowed or divorced, but rises when they re-marry (Lee et al, 2005). From the foregoing examples, it is evident that social influence does not uniformly promote healthier behaviors, and the intriguing gender differences may partly explain why the health benefits of marital ties have been reported to be stronger for men than for women (House et al, 1988).
A separate pathway linking social networks to health is through social engagement, which refers to participation in social activities through one's social relationships. Participation in turn defines and reinforces an individual's social roles, identity, meaning, and sense of belong-ingness. Lifelong engagement in social activities has additionally been linked to maintenance of cognitive ability at older ages (Bassuk et al, 1999) through what appears to be a "use it or lose it" mechanism.
Thirdly, social networks are the conduits through which the transfer and exchange of social support take place. Social support is further classified into several subtypes, ranging from emotional support (love and affection) to the exchange of information and advice, and instrumental support (cash loans, labor in kind). Of these different types, particular attention has been paid to emotional support because of its direct influence in producing positive affective states, which are in turn believed to dampen neuroendocrine responses to stress (the stress-buffering mechanism referred to earlier). We emphasize again that social networks and social support are distinct constructs. Hence, it is possible to receive emotional support from others who are not part of one's social network (e.g., a crisis interventionist volunteering on a suicide hotline). Conversely, social networks may produce differences in immune, inflammatory, or
2In popular parlance, this effect may be dubbed "nagging", but we will stick to the term "social influence".
neuroendocrine responses even in the absence of mobilizing emotional support. For example, in Sheldon Cohen's experiments exposing volunteers to an intranasal dose of the cold virus, individuals reporting high social network diversity (i.e., the presence of social ties in many domains including the work-place, community groups, churches) experienced roughly half the risk of succumbing to a symptomatic cold compared to more isolated individuals, even though the experiments did not involve any manipulation of social support in the laboratory (1997). Presumably, this finding is explained by some as-yet unaccounted for the effect of social network integration on immune functioning (i.e., the ability to fend off the cold virus). Furthermore, longitudinal data from the Framingham Study (Loucks et al, 2006) as well as the MacArthur Successful Aging Study (Loucks et al, 2006) have reported associations between higher levels of social networks and lower levels of inflammatory markers such as interleukin-6 and C-reactive protein, even after controlling for depressive symptoms, SES, smoking, body mass index, and physical activity.
The fourth distinct mechanism through which social networks may affect well-being is through improving access to material resources. This mechanism is distinct from the mobilization of instrumental social support (e.g. cash loans) during times of crisis. The idea here is that people with wider social networks are able to access more opportunities (e.g., job openings) by virtue of their connections.3 Crucially, the connections need not be particularly strong or intimate. As Granovetter (1973) demonstrated in his classic study of job seekers, most successful job hunters found their work not through their closest contacts, but through "friends of friends" - i.e., their so-called "weak ties". The explanation for this seemingly paradoxical finding is that close
3Sometimes, this concept has been referred to by the term "individual social capital" (see van der Gaag and Webber,
friends tend to share access to the same information, whereas novel information is likely to originate from more remote sources. In other words, social networks do not represent an undif-ferentiated source of "host resistance" to illness. Different aspects of social network promote health through distinct mechanisms. Thus, when a person is in need of emotional support, having strong social network contacts (e.g., confidants) matters the most, but when one is unemployed and seeking work, having a far-flung network of weak ties matters more.
Was this article helpful?