The Psychometric Properties of the Vanderbilt Attention-Deficit Hyperactivity Disorder Diagnostic Teacher Rating Scale in a Community Population

Abstract

This study examined the psychometric properties of the Vanderbilt AD/HD Diagnostic Teacher Rating Scale (VADTRS). Information was collected from teachers and parents in 5 school districts (urban, suburban, and rural). All teachers in participating schools were asked to complete the VADTRS on all their students. Construct validity was evaluated through an exploratory factor analysis investigation of the 35 items that made up the 4 scales of inattention, hyperactivity, conduct/oppositional problems, and anxiety/depression problems. Convergent validity was assessed among a subsample of participants whose teachers completed the Strengths and Difficulties Questionnaire (SDQ). Finally, predictive validity was examined for another subsample of high- and low-risk children whose parents completed a structured psychiatric interview, the Diagnostic Interview Schedule for Children-IV. For construct validity, a 4-factor model (inattention, hyperactivity, conduct/oppositional, and anxiety/depression problems) fits the data well. The estimates of the KR20 coefficient for a binary item version of the scale ranged from .85 to .94. Convergent validity with the SDQ was high (Pearson's correlations > .72) for these 4 factors. For predictive validity, the VADTRS produced a sensitivity of .69, specificity of .84, positive predictive value of .32, and negative predictive value of .96 when predicting future case definitions among children whose parents completed a diagnostic interview. The confirmation of the construct and convergent validity and acceptable scale reliabilities found in this study further supports the utility of the VADTRS as a diagnostic rating scale for attention-deficit hyperactivity disorder. The low predictive validity further demonstrates the need for multiple observers in establishing the diagnosis.