Rationally and empirically derived dimensions of children's symptomatology: Expert ratings and confirmatory factor analyses of the CBCL

Expert ratings and confirmatory factor analyses were used to develop an alternative system for scoring the Child Behavior Checklist (CBCL; T. M. Achenbach, 1991) to measure specific dimensions corresponding to current conceptualizations of child symptomatology. Data were from a nonclinic and 2 independent clinic samples. Subscales measuring Anxiety, Attention Problems/Hyperactivity, Conduct Problems, Depression, Oppositional Defiant, Social Problems/Immaturity, and Somatization were created. Logistic regressions were conducted to evaluate the diagnostic efficiency and discrimination of the new and original approaches to scoring the CBCL. Some of the new subscales demonstrated better sensitivity, positive predictive power, and discriminant validity than the original CBCL subscales; however, subscales from both approaches demonstrated low sensitivity. Results support the use of the new subscales for specific research purposes.