measured in terms of social activities with others—have been shown in other studies (Fabrigoule et al., 1995; Glass et al., 1999; Kiely et al., 2000). Greater social integration and more emotional support from others predict lower mortality after myocardial infarction (Berkman et al., 1992; Case et al., 1992; Williams et al., 1992). These factors also lead to lower mortality or better functional recovery after a stroke (Colantonio et al., 1993; Glass et al., 1993). More recently, higher levels of emotional support have been shown to protect against cognitive decline in the MacArthur Successful Aging Cohort (Seeman et al., 2001).
The importance of social engagement is also demonstrated in studies of the effect of religious involvement, which tends to increase with age. Compared with the younger population, older persons generally express a higher degree of religiosity (Greeley, 1989). Fellow church members provide emotional and material support, as well as information, advice, and spiritual benefits (Taylor and Chatters, 1988). Religious beliefs, it is argued, are an important source of hope and comfort and provide systems of meaning to help cope with stress, disability, the loss of loved ones, and the fear of impending death (Koenig et al., 1998).
A significant factor in religion is church attendance, which appears to be related to lower mortality and disability. Net of other demographic and socioeconomic factors, adults who attend church more than once a week have a one-third lower risk of death (in 8-year follow-up data of a nationally representative sample) than adults who never attend church (Hummer et al., 1999). An effect of similar magnitude has been verified for an older population (Koenig et al., 1999). Other cohort data also show that church attendance is associated with a lower risk of disability over a 12-year follow-up period, independently of the effects of a comprehensive set of other predictors of disability (Idler and Kasl, 1997). In contrast with the effects of church attendance, private religious activity and personal religiosity are not associated with better survival or reduced disability (Hummer et al., 1999; Idler and Kasl, 1997; Mendes de Leon and Glass, 2004).
Evidence regarding how social engagement affects health risks is beginning to emerge. Both community studies and laboratory work show that social relationships, particularly supportive ones, are associated with lower risk biological profiles. Community studies point to lower blood pressure, serum cholesterol, and levels of stress hormones (e.g., norepinephrine, epinephrine, cortisol) among people who report greater social integration or support than others (Seeman and McEwen, 1996; Seeman et al., 2002). Attendance at religious services, for instance, is associated with lower systolic and diastolic blood pressure among both black and white elderly people (Koenig et al., 1998). Laboratory-based experiments demonstrate that the presence of supportive others reduces cardiovascular and neuroendocrine reactivity among those exposed to challenging tasks (Seeman and