The Longitudinal Study of Aging (LSOA) assessed the health and social functioning of a representative sample of 7527 American community-dwelling older people (>70 years). We tested the hypothesis that frequent volunteering is associated with less mortality risk when the effects of socio-demographics, medical status, physical activity and social integration are controlled. We used Cox proportional hazards analyses to assess the unadjusted and adjusted associations between frequency of volunteering and time-to-death (96-month follow-up). Death occurred in 38.3 percent of the sample. After adjusting for covariates, frequent volunteers had significantly reduced mortality compared to non-volunteers. This association was greatest for those who frequently visited with friends or attended religious services.