Agreement among teachers' behavior ratings of adolescents with a childhood history of attention deficit hyperactivity disorder.

Examined agreement among secondary school teachers' behavior ratings for 66 adolescent boys with a history of attention deficit hyperactivity disorder. Behavior ratings consisted of the Teacher Report Form, Iowa/Abbreviated-Conners, and the Disruptive Behavior Disorders Rating Scale. Ratings from 2 to 5 teachers were collected for each adolescent. In contrast to previous studies, agreement was examined using statistical indices that corrected for chance agreement and discrepancies in scores (i.e., intraclass correlation [ICC], kappa) in addition to traditional indices (i.e., Pearson correlation and percentage agreement) typically used in the relatively sparse literature on teacher agreement for adolescent behavior ratings. Agreement was poor for dimensional subscale scores (Pearson correlations were in the .40-.50 range, and ICCs were in the .20-.50 range) as well as for categorization of youth as above or below clinical cutoffs (percentage agreement was between 52% and 96%, but ICCs and kappas ranged from .17 to .57). Findings suggest that, regardless of behavior rating scale used, a multiple teacher assessment strategy should be adopted for clinical assessment, treatment design, and evaluation of treatment efficacy.