DIAGNOSING ATTENTION-DEFICIT DISORDERS WITH THE BEHAVIORAL-ASSESSMENTSYSTEM FOR CHILDREN AND THE CHILD-BEHAVIOR CHECKLIST - TEST AND CONSTRUCT-VALIDITY ANALYSES USING OPTIMAL DISCRIMINANT CLASSIFICATION TREES

The usefulness of the Behavioral Assessment System for Children (BASC) and Child Behavior Checklist (CBCL) Parent scales was examined with r espect to (a) differentiating students with attention deficit-hyperact ivity disorder (ADHD) from non-ADHD students and (b) discriminating be tween the predominantly inattentive-type and combined-type ADHD-afflic ted students. For both the BASC and the CBCL, a different optimal disc riminant classification tree analysis (CTA) model was developed for ea ch of the 2 diagnostic predictions. For distinguishing ADHD students f rom non-ADHD students, the BASC model was more parsimonious and accura te than the CBCL model. Toward the goal of differentiating between pri marily inattentive and combined types,the CBCL's model was superior fo r predicting primarily inattentive students. The results demonstrate t he diagnostic utility of the BASC and CBCL and describe salient behavi oral dimensions associated with subtypes of ADHD.