Each patient, along with the members of the treatment team (eg, parents, teachers, case workers, group home staff, workshop staff), participated in 2 detailed assessments of behavioral symptoms at baseline and again at the end of weeks 4, 8, and 12 of the controlled trial. Repetitive behavior was rated with a modified version of the Y-BOCS.33- 34 Each of the 10 items on the Y-BOCS is scored on a 5-point scale from 0, indicating "least symptomatic," to 4, indicating "most symptomatic," so that the total Y-BOCS score ranges from 0 to 40. The first 5 items of the Y-BOCS are designed to assess the severity of repetitive thoughts, whereas the last 5 items determine the severity of repetitive behavior. Because 16 (52%) of the patients were nonverbal and thus unable to provide information about repetitive thoughts, we chose to assess only repetitive behavior for each patient; thus, the maximum Y-BOCS score for each patient was 20. Based on previous findings,35 the ego-dystonicity diagnostic criterion for obsessive-compulsive disorder was eliminated in rating the repetitive behavior of the patients with PDD. Aggression was rated with the SIB-Q, a 25-item clinician-rated instrument that assesses self-injurious behavior, physical aggression toward others, destruction to property, and other maladaptive behavior (T. Gualtieri, MD, unpublished data). Patients could receive a score from 0, indicating "not a problem," to 4, indicating "severe problem," on each of the 25 items (range for total score, 0-100). The Ritvo-Freeman Real-life Rating Scale36 served as an observational measure (30 minutes during each behavioral rating session) of a variety of symptoms of autism, including subscales for assessing sensory motor behaviors (eg, hand-flapping, rocking, pacing) (subscale I), social relationship to people (eg, appropriate responses to interaction attempts, initiating appropriate physical interactions) (subscale II), affectual reactions (eg, abrupt changes in affect, crying, temper outbursts) (subscale III), sensory responses (eg, agitated by noises, rubbing surfaces, sniffing self or objects) (subscale IV), and language (eg, communicative use of language, initiating appropriate verbal communication) (subscale V). Each of the 5 subscales contains a number of individual items that are scored on a 4-point scale: 0 indicates "never"; 1, "rarely"; 2, "frequently"; and 3, "almost always." The average score from the mean value of each of the 5 subscale scores was determined to yield an overall score on the Ritvo-Freeman scale (range for overall score,−0.42 to 2.58). The Ritvo-Freeman scale score increases as the number and frequency of symptoms of autism increase. A mathematical sign correction to subtract normal behavior is necessary on subscales II, IV, and V; this results in some scores with a negative value. Different mood states were assessed on 10 clinician-rated visual analog scales,37 scored on a 100-mm line where 0 indicates "not at all," and 100 indicates "most ever." The 10 items included "anxious or nervous," "calm," "depressed," "eye contact," "happy," "irritable," "restless," "social interaction," "talkative," and "tired." Finally, the CGI global improvement item—7 indicates "very much worse"; 4, "no change"; and 1, "very much improved"30 —was recorded at the end of weeks 4, 8, and 12 of treatment (compared with the predrug baseline). Global improvement ratings were based on an aggregate assessment of the patient's social behavior and the level of the patient's interfering repetitive and aggressive behavior. For the 15 patients who received treatment for 12 weeks with open-label risperidone following the double-blind placebo phase, the rating scales were administered in a manner identical to that described above.