Time and the convoy model: social network turnover and mental health in older adulthood

OBJECTIVES: The convoy model posits that continuity and lack of turnover in ties is key for good mental health in older adulthood. However, this proposition is rarely tested with data on the addition and loss of specific people from a person's social life. We undertake that analysis using a nationally representative, longitudinal survey of older adults, and examine the correlates of turnover (especially race). METHODS: We draw on three rounds of the National Social Life Health and Aging Project (3,005 cases in 2005/2006, 3,363 in 2010/2011, and 4,602 in 2015/2016) and model addition/loss of confidants based on their closeness to the respondent using population-averaged models. We model changes in mental health using lagged dependent variable models, predicting those changes using the loss/addition of people somewhat close, very close, or extremely close to the respondent. RESULTS: Black respondents, compared to White respondents, were less likely to have "somewhat" close people in their network, and were less likely to add "somewhat close" people over time. Total losses were associated with greater levels of depression but not with changes. Loss of "somewhat close" confidants was associated with greater levels of depression (but not changes). DISCUSSION: We found evidence for disparities in social network turnover by race, but no evidence that these disparities impact changes in depression. We conclude with implications for the convoy model and syntheses between convoy perspectives and theories of network advantage.