Research That Matters (January 17 - 20, 2008)

Purpose: Research demonstrates that self-esteem play a significant role in maintaining good health among older adults. Older adults with high self-esteem tend to actively engage in health-promoting behaviors (e.g., exercise). Facing age-related losses (e.g., functional impairment and spousal loss), however, not all older adults may successfully maintain self-esteem in late life. There may be a wide variability in older adults' capability to maintain self-esteem in old ages. Is there any variability in trajectories of self-esteem within- and between-older adults and, if yes, what factors account for the variability? To answer this question, we examine conjoint trajectories of self-esteem and of social relations over time that have been rarely explored in previous research. Based on socioemotional selectivity theory (Carstensen, 1995) that suggests emotional goals are preferred in late life, we hypothesize that rate of change in self-esteem is associated with positive social exchanges and emotional social networks in old ages.

Methods: Data are from the nationally representative Americans' Changing Lives (ACL) study (House et al., 1990), for the subgroup of older adults aged 60 to 95 years old who participated in the baseline interview (n=1,669) and following two interviews in 1989 (n=1,279) and 1994 (n=889). To examine within- and between-individual change in self-esteem with longitudinal data, we conduct growth curve analyses using Hierarchical Linear Models (Bryk & Raudenbush, 1992). Self-esteem is measured by a self-esteem index comprising three items from the Rosenberg Self-Esteem Scale (Rosenberg, 1979). Positive exchanges are measured by an index of positive social exchanges with friends/relatives. Emotional social networks are assessed by the number of persons sharing private feelings and instrumental social networks are measured by the number of persons receiving help/advice. Social integration is measured by the two indexes: formal and informal integration. We include other important variables (e.g., gender, education, and health status) to control for spurious effects.

Results: Results show that older adults as a whole experience linear decline in self-esteem across the three waves. Outcome data also demonstrate significant variation in trajectories of self-esteem between older adults. As hypothesized, rate of change in self-esteem is significantly associated with positive exchanges with friends/relatives. Especially positive exchanges with friends/relatives are associated with a slower increase in self-esteem for female older adults. The number of persons sharing private feelings is associated with a slower decline in self-esteem for female older adults. On the other hand, there is no significant association between rate of change in self-esteem and changes in social integration and instrumental networks. Finally, education and gender are significant predictors for inter-individual differences in trajectories of self-esteem.

Implications: The findings suggest the importance of taking into account gender and positive and emotional aspects of social relations when we develop interventions for older adults' resilience in self-esteem. As positive and emotional social networks appear to be more important than social integration and instrumental social networks, we need to have more attention to social work practice or interventions that employ emotion-focused strategy rather than problem-solving strategy with regard to the resilience of self-esteem of older adults.