In a sample of 77 dyads, involving depressed patients at least 50 years of age and their family or friends (informants), patient illness burden and cognitive decline were associated with selfinformant rating discrepancies for facets of NEO-PI-R Openness and Extraversion. Informant judgments about Neuroticism and Conscientiousness were not associated with illness burden or cognitive function, underscoring the potential utility of risk-detection strategies that rely on informant-report in these two domains. Findings suggest the need for research on how patient illness severity and cognitive function affect how friends and family use or misuse information when making judgments about older depressed patients.
KeywordsInformant reporting; Personality; Medical illness; Cognitive function; Older adults The reliable judgment of others' personality characteristics is important for guiding social interactions in everyday settings (Funder, 1995;2003;McAdams & Pals, 2006) and in research contexts where self-report data are unavailable (Useda et al., 2007) or of dubious reliability (Benedict et al., 2009;Duchek, Balota, Storandt, & Larsen, 2007;Siegler, Dawson, & Welsh, 1994). In clinical practice, informant ratings of personality are useful as a complement to self-report (Costa & Piedmont, 2003;Ganellen, 2007;Klein, 2003;Vazire & Mehl, 2008), but only if the ratings are reliable and discrepancies between self-and informant-report can be probed.Research on the reliability of informant ratings of personality traits has expanded rapidly since the 1980s. Many studies have been conducted on undergraduates, non-patients, and fictitious patients (Costa & McCrae, 1988, 1992Funder, 2003;Funder & Dobroth, 1987;Watson, 1989). The extent to which findings are applicable to older adults, clinical samples, and real patients-those at greatest risk for morbidity and mortality-is uncertain. In fact, correlations between self-report and informant-report traits have been lower in the few studies involving adult patients (Bagby et al., 1998;Ready & Clark, 2002;Yang et al., 1999), perhaps due to their higher levels of medical comorbidity or cognitive impairment.Despite the potential value of informant reports, and the need to learn more about their reliability, few studies have either attempted to examine correlates of self-and informantreport discrepancies (Benedict et al., 2009;O'Rourke, Neufeld, Claxton, & Smith, 2010) or understand how discrepancies between self-and informant-report arise (McCrae, Stone, Fagan, & Costa, 1998 (McCrae et al., 1998), but psychologically substantive considerations may also be important (O'Rourke et al., 2010), particularly in the context of age-or disease-associated changes in cognition and illness burden (Benedict et al., 2009).Prior research suggests that patient illness burden or cognitive function may reduce selfinformant agreement of personality ratings. In a study of female multiple sclerosis patients and their significant others, self-informant agreement varied across the course of the disease (B...