Elderly confidants in geriatric assessment

This study focuses on patterns of consensus between linked samples (focal person and confidant; total n = 124) for the purpose of explaining divergent reports and of assessing this design's utility in geriatric case assessment. Correlations, difference scores and group contrasts suggest that elderly informants who are also confidants can describe most internal states of their partners as these states are self-reported. Partners differed however, on ratings of self-esteem; confidant scores were significantly higher than self-ratings. Under certain conditions self-other discrepancy also occurs on spouse ratings of functional health: in early stages of decline and in marriages judged by clinicians to be of poor quality. This pattern is absent in pairs with women friends. The findings suggest that losses and age-related limitations have the potential to erode the self-esteem of older people and that linked samples would strengthen the process of functional assessment by bringing special information to the clinician.