Search

Although retirement is generally thought to lead to a decline in social support due to a loss of social contacts with coworkers, the evidence for this is at best contradictory. This longitudinal study examined change in social support among 1,311 men, participants in the Veterans Administration Normative Aging Study, over a 3-year period. In general, long-term retirees reported the least quantitative social support, and the continuing full-time workers the most; however, change in workforce status produced no apparent effect on quantitative support over the duration of this study. In general, qualitative support showed no retirement effects. We also examined more specific patterns of coworker friendship and again found a pattern similar to that of general quantitative support. The findings support our earlier suggestion of a distinction between retirement as a transition and retirement as a state, and also support the convoy model of social support in late life and the selectivity theory of developmental adaptation to aging.

The stressfulness of retirement both as a transitional event experienced during the past year and as a life stage was investigated. Transitional stress was assessed using a life events approach, and stage stress using a "hassles" approach. Respondents were 1,516 male participants in the Normative Aging Study, 45% of whom were retired. Among those retiring in the past year, respondents' own and spouse's retirement were rated the least stressful from a list of 31 possible events. Only 30% found retirement stressful. Retirement hassles were also less frequently reported and were rated less stressful than the work hassles of men still in the labor force. The only consistent predictors of both transitional and stage retirement stress were poor health and family finances; personality did not predict retirement stress.

We investigated the relationship between the personality traits of emotionality and extraversion and symptom-reporting, both cross-sectionally and longitudinally. The sample consisted of 1,034 men, initially screened for good health, who completed both the 16PF and at least two CMIs, (M = 5, range = 2-8) over an average period of 17 years (range = 2-25). Using a two-stage growth model, we obtained a predicted baseline and slope for each man for self-reported physical and psychological symptoms. Correlations revealed that emotionality was positively related to the numbers of both physical and psychological symptoms reported at baseline, while extraversion was negatively related. However, regression analyses including age and both personality variables indicated that only emotionality was associated with level of symptom-reporting. Longitudinal analyses revealed that emotionality was unrelated to physical symptom slopes and only weakly related to psychological slopes. We interpret these results as showing that, while personality and symptom-reporting are related cross-sectionally, personality does not predict changes in symptom reporting.