OBJECTIVE: Difficulties in the assessment of personality disorders and the burgeoning interest in axis II have led to increased use of informants when studying these conditions. The present study sought to evaluate the correspondence between patients and their informants on symptoms of personality disorders. METHOD: A total of 105 outpatients and knowledgeable informants were independently interviewed by using the Personality Disorder Examination, a widely used instrument for the full range of personality disorders. RESULTS: Diagnostic concordance between interviews was low (median kappa = -0.01), while correlations between dimensional scores were somewhat higher (median kappa = 0.36). Overall, patient interviews showed more pathology than interviews with informants. However, many of the symptoms obtained from informants were not reported by patients. CONCLUSIONS: These data suggest that patient- informant concordance for axis II disorders is poor for diagnoses but somewhat better for dimensional scores. There was no evidence that low agreement can be explained by patients attempting to present themselves in a favorable light. Further work is necessary to elucidate the reasons for discordance and determine which data source provides the most valid information.