An expanded version of the Rutter parent questionnaire was administered to 320 schoolchildren with hemiplegia. Scores and caseness categories derived from the original Rutter items corresponded well with the findings of independent psychiatric evaluations. A measure of prosocial behaviour derived from 10 novel items had promising internal consistency, factorial independence and inter-rater reliability. Another six novel items loaded on the same factor as the three original Rutter hyperactivity items. The modified questionnaire may be useful with a wide range of children, particularly if the focus on children's strengths as well as weaknesses increases parental compliance and reduces halo effects.