Contrasting perspectives on personality problems: descriptions from the self and others

Self-reported personality disorder (PD) traits have low to moderate correlations with informant and peer-reports of the same traits. The current study investigates the source of disagreement in correspondence and proposes that in some cases it is due to systematic differences in perceptions of traits by self and others. Military recruits (N = 2013) were administered self-report measures and an inventory in which they nominated peers who displayed DSM-1V PD traits. Correspondence between self- and peer-report for the same diagnostic categories was low (Rs ranging from 0.11 to 0.22). Semipartial correlations were used to find the 10 items which best predicted each diagnostic category over and above the corresponding scale. Adding these supplemental items to the original regression significantly increased the amount of variance explained (Rs ranged from 0.15 to 0.35) in a cross-validation sample. Although peer- and self-report differed in content, the relationship between self and peer perceptions appears to be systematic and meaningful (e.g., people whose peers describe them as paranoid describe themselves as angry and hostile). (C) 2003 Elsevier Ltd. All rights reserved.