This study examined the relations between performance on alternative measures of childhood depression and diagnosis of depressive disorder. Hospitalized inpatient children (N = 170, ages 7-13) with a diagnosis of depressive disorder (Research Diagnostic Criteria) were compared with a matched sample of patients whose diagnoses excluded depression. Children and parents completed four standardized depression scales plus measures designed to assess associated features including hopelessness, self-esteem, and internalizing symptoms. The results indicated that alternative depression measures, whether completed by children or parents, yielded scores that were significantly higher for children with a diagnosis of depression; parent ratings of severity of depression were consistently higher among the measures than the child ratings; optimal cutoff scores derived for each measure correctly classified approximately 60% of depressed and nondepressed cases; different cutoff scores were required for the same measure, depending on whether children or parents were the raters; and using a battery of measures and combining these in a stepwise discriminant function yielded over 80% classification accuracy. The implications of the results for selection of cases for clinical research are discussed.