Few studies have compared self-report and clinician-administered measures of youth psychopathic features in juvenile-justice settings in terms of antisocial behavior and treatment indices. In a sample of 85 adjudicated delinquents, the predictive validities of the Antisocial Process Screening Device (APSD), the modified Childhood Psychopathy Scale (mCPS), and the Hare Psychopathy Checklist: Youth Version (PCL:YV) were tested. Three indices of institutional antisocial behavior (physical aggression; verbal aggression; administrative infractions) and two indices of treatment progress (time to treatment level promotion; whether treatment levels were dropped) were used as external correlates. The self-report measures (mCPS more so than APSD) were more consistently and strongly related to antisocial behavior and to the days required to progress in treatment than the PCL:YV. The following issues are discussed: (i) implications of the potential impact of measurement format on the understanding and predictive validity of youth psychopathy features and measures; (ii) the differential predictive validity of self-report versus clinician-administered measures; and (iii) the potential practical utility of measures of psychopathic features in youth. Copyright © 2004 John Wiley & Sons, Ltd.