The current study examined the prototype-matching technique for using the five-factor model (FFM) of personality to assess personality disorders (PDs) and their correlates. The sample was composed of 69 psychiatric patients, most of whom suffered from affective or anxiety disorders. The participants were predominantly outpatients (78%), Caucasian (94%), and women (58%). NEO Personality Inventory (NEO-PI) data from these patients and informants were available for deriving PD prototype scores. The results supported four conclusions. First, the FFM prototype-matching method is useful for clinical samples. Second, agreement between self-reports and information from significant others (SOs) using this method is good compared to previous results. Third, this agreement varies systematically with the observability of the PD criteria. Fourth, the value of self- and other-reported personality information depends on the rating source of the outcome variables. Other-reported information provides incremental utility in predicting impairment and interpersonal distress rated by clinicians and significant others.