Results were compared from independent interviews using the Schedule for Affective Disorders and Schizophrenia for School-aged Children-Epidemiologic Version and DSM-III with 220 subjects (ages 6 to 23 years) and their parent informants. In agreement with results from studies using a variety of structured diagnostic interviews or symptom scales, considerable discrepancies were found between parents' and children's reports on the degree and nature of the child's psychopathology. The children reported more illness about themselves than their parents reported about them. The parents' reports were primarily a subset of the children's reports. Various factors that might affect agreement, including demography, parental clinical status, severity of illness, and treatment, were also explored. The findings that parents under-report psychiatric disorders in their children are comparable with those reported in studies of adults when family informants are used to obtain diagnostic information. Until these parent-child discrepancies can be resolved by longitudinal, family, and other research, diagnostic assessment of children should include direct interviews with them. An independent assessment of the child's diagnosis based on information from multiple informants, including the child, may be the best estimate.