Doubt has been cast recently on the usefulness of the Wechsler P > V sign test because of a failure to find significant correlations between it and many personality traits measured by the Minnesota Multiphasic Personality Inventory and the Milton Clinical Multiaxial Inventory. This article takes issue with this doubt and argues that intellectual imbalance in either direction; that is, V>P or P>V, is a useful predictor of the probability of becoming delinquent. It is also shown that P>V was not an artifact of (a) lower overall IQ, (b) lower verbal IQ, or (c) socioeconomic status (SES) among a sample of 513 male juvenile delinquents. Finally, it appears that, once a youth becomes delinquent, the frequency and severity of delinquent involvement can be predicted by the degree of P>V discrepancy, holding SES and race constant.