In order to establish whether subjects asked to simulate neuroticism (i.e. malinger) can be discriminated from control subjects and genuine neurotics, a sample of each type of subject was given the Maudsley Personality Inventory (MPI). Keehn's malingerer scale was included in the inventories to see whether it improved the discrimination. Discriminant function analyses were carried out, and they showed that, while the usual form of the MPI gave a misclassification rate of 25 per cent, inclusion of the malingerer scale gave a misclassification rate of only 11 per cent. The main improvement was to reduce the bad error of classifying neurotics as malingerers. Analysis of the scores of a second sample of the three types of subject showed that the predictive accuracy of the discriminant weights established with the first groups was slightly reduced, as would be expected, but still highly satisfactory for a personality questionnaire.