A personalized power scale for the prediction of alcoholism

An objective test measure of personalized power was developed to test McClelland's hypothesis. Four existing personality questionnaires that have been considered to be relevant to personalized power were used. These scales were administered to 50 patients in a Veterans Administration alcoholism program and to 50 VA medical outpatients. Items which discriminated the groups at the p less than or equal to .20 level were included in a new scale, which was then cross-validated on a new group of alcoholics and medical outpatients. The resulting 34-item scale was internally consistent (.83) and correctly classified 76% of the alcoholics and 94% of the medical outpatients. Overall, 86% were correctly classified as compared to an overall correct classification by the MacAndrew Scale of 67% on the same population. While the personalized power dimension seemed to add to the scale's predictive ability, the contribution of that dimension appears to be modest at best.