The Death of Confidants and Changes in Older Adults' Social Lives.

This study examines the consequences of confidant death for the social lives of older adults, testing hypotheses from socio-emotional selectivity theory and the hierarchical compensatory model. We draw upon longitudinal data from the National Social Life Health and Aging Project-a nationally representative survey of older adults ( = 2,261). We employ ordinary least squares (OLS) and ordinal logistic regressions in the context of multiple imputation with chained equations, checking our findings with doubly robust estimation. We find that the death of a spouse, but not the death of a family member or friend, was associated with increased support from friends and family, spending more time with family, and more frequent participation in religious services, but not volunteering. Death of other confidants also had little impact on older adults' social lives, suggesting the robustness of their networks to nonspousal loss.