The Structured Interview for DSM-III Personality Disorders was used to interview 82 patients. In addition, a close relative or friend of the patient was interviewed regarding each patient's normal personality. After the informant interview, the diagnosis of the presence or absence of any personality disorder was changed in almost 20% of the sample. In general, the information given by the informants revealed additional pathology in the patients and was less frequently used to retract a diagnosis. Despite conflicting information from the patients and informants, the reliability of personality disorder assessment remained high.