The literature evaluating agreement between parent and teacher reports on symptoms of attention deficit disorder (ADD) has repeatedly shown only modest correlations between these two sources of information. A more relevant statistic to this clinical issue, however, is the positive predictive power, the conditional probability that diagnostic criteria will be met by teacher's report if that diagnosis has been made based on parental report. In a population of 43 children (ages 4 to 17) satisfying clinical criteria for DSM-III ADD, parents and teachers independently completed them module on ADD derived from the Diagnostic Interview for Children and Adolescents. Although correlations for individual symptoms were low to moderate, there was a 90% probability that the teacher report would result in a positive diagnosis given a positive parent diagnosis, indicating a very high probability that, in clinically referred children, a clinical diagnosis of ADD based on parent report will be corroborated by a teacher report.