Using data from the 1992 and 1994 waves of the Health and Retirement Study (HRS), this study described reasons reported by pre- and early-retirement populations for perceived changes in global health status over a 2-year period. It then analyzed the association between self-perceptions of change and the actual changes in objective health conditions, controlling for demographics, emotional health status, and the changes in work status and health-affecting habits. The results were compared to the determinants of self-ratings of health at wave 2. Existing or increasing impairments in functional abilities were found to contribute to self-perceptions of decline. However, a diagnosis of new chronic disease and the experience of a major medical event per se did not universally contribute to self-perception of decline. The relationship between cross-sectional self-ratings of health and objective health conditions was more straightforward. Self-perception of improvement among people with serious health problems most likely owed to medical interventions and improvement in symptoms, the most frequently mentioned reasons for perceived improvement, and reflected the subjects' selective optimization and resiliency. (Contains 3 tables.)