Abstract
A growing literature documents the positive effect of social relationships on health, in general, and in reducing mortality, in particular. Much remains to be learned about which relationships have this effect, particularly among the elderly. This research, a secondary analysis of a national sample of community-dwelling elderly aged 70 and over, addresses these questions. Data come from the Longitudinal Study on Aging. Five scales constructed from relationship questions asked in the survey are related to respondent mortality over a 4-year follow-up using multiple logistic regression . The research confirms that certain relationships appear to reduce subsequent mortality: going to church/temple, volunteering, seeing friends or neighbors, and talking with them on the phone (socioexpressive relationships). The effect can be seen among healthy and more infirm elders. Helping relationships and household and kin relationships do not seem to reduce the risk of elder moriality.