Age differences in the reliability of the psychiatric interview of the child.

The NIMH Diagnostic Interview Schedule for Children, a highly structured interview covering a broad range of clinically relevant symptoms and behaviors, was administered to 242 disturbed children and their parents. Parent and child were interviewed separately and were assessed twice at a median interval of 9 days. Intraclass correlations between symptom scores derived from the interviews indicated that parents were generally more reliable than children in reporting child symptoms. However, test-retest reliabilities showed an opposite age pattern for parent and child. The reliability of the child's report increased with age and was lower for children aged 6-9 than those aged 10-13 and 14-18. Conversely, the reliability of the parent's report decreased with the age of the child and was slightly higher for children aged 6-9 than those aged 10-13 and 14-18. These findings were interpreted in terms of children's cognitive development and age-related shifts in parents' perceptions and awareness of their children's behavior.