Multimethod assessment of psychopathology among DSM-IV subtypes of children with attention-deficit/hyperactivity disorder: self-, parent, and teacher reports.

Using data based on self-, parent, and teacher reports, we assessed various aspects of psychopathology in a large sample of control children and those with ADHD. Confirmatory factor analysis was employed to extract response bias from latent constructs of aggression, anxiety, attention problems, depression, conduct disorder, and hyperactivity. These latent constructs were then entered into logistic regression equations to predict membership in control versus ADHD groups, and to discriminate between ADHD subtypes. Results of the regression equations showed that higher levels of attention problems and aggression were the best predictors of membership in the ADHD group relative to controls. Logistic regression also indicated that a higher degree of aggression was the only significant predictor of membership in the ADHD-Combined group compared to the ADHD-Inattentive group. However, when comorbid diagnoses of Oppositional Defiant Disorder and Conduct Disorder were controlled for in the logistic regression, greater hyperactivity rather than aggression was the sole variable with which to distinguish the ADHD-Combined from the ADHD-Inattentive subtype. Results are discussed in the context of the DSM-IV ADHD nosology and the role of instrument and source bias in the diagnosis of ADHD.