Abstract A slightly modified form of the Quay and Peterson Questionnaire for the measurement of three personality dimensions associated with juvenile delinquency was administered to 69 institutionalized juvenile delinquents and 230 school children with behaviour problems. Results were factor analysed by an image covariance analysis, and it was found that although two factors seemed to match with those of Quay and Peterson, only 13.84% of the total test variance was accounted for by the first three factors. Coefficient α, an index of internal consistency reliability, was calculated for the three item sets proposed as homogeneous scales by Quay and Peterson, and in each case α did not reach the minimal acceptable level of reliability for scales of this type. Separate factor analyses of the three item scales proposed by Quay and Peterson showed that none of these scales was satisfactorily unidimensional, but there was a core of items on each scale which could perhaps be used as a nucleus for further revision...