Individuals traditionally have been diagnosed as having Attention Deficit Hyperactivity Disorder (ADHD) by using a combination of rating scales of criteria that describe symptoms of the disorder and case histories. Cut-off points are suggested in the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1994) for the number of criteria above which individuals are considered to have the disorder. An alternative approach would be to use Rasch scaling to determine the relative frequency of each behavioral symptom. The cut-off points used at the present time imply that the more symptoms an individual shows, the greater will be their impairment. The possibility that some of the criteria might be more indicative of the severity of impairment than other items and more predictive than others of learning difficulties of young children was explored in this study. Data came from schools that were part of a study of performance indicators in elementary schools in England. Participants were 1,821 children from 70 schools. Principal components analysis was used to explore in detail the separate dimensions of inattention, hyperactivity, and impulsivity. The less frequently met items of the behavior rating scale did not appear to indicate a more severe level of impairment in terms of academic achievement . The total number of criteria met appeared to be more important than their frequency, and items related to inattention were more strongly related to achievement than criteria relating to hyperactivity/impulsivity. The highest correlation between behavior and achievement was found when a combination of a few behavior items were used. These items were within the attention subscale, but did not form a specific factor in the analysis. An appendix contains the Behavior Rating Scale used. (Contains 2 tables, 10 figures, and 19 references.) (Author/SLD)