Most instruments that assess personality disorder rely principally on self-report. However, there are major limitations to using self-report for the diagnosis of personality pathology. First, the self provides only one opinion. Second, like everyone else, people with personality disorders are frequently unable to view themselves realistically and are unaware of the effect of their behavior on other people. Using informant data, therefore, may improve the reliability and validity of diagnostic assessment. A review of the literature reveals that agreement between informant- and self-reports of personality disorder is modest at best, even though informants tend to agree with each other. Self-informant concordance appears to be higher for older subjects and for Cluster B traits (excluding narcissism). Further research should focus on methods of resolving discrepancies between self- and informant-reports of personality disorder and determining when self or informant data are more valid.
