45 patients hospitalized for depression who had received double-blind trials with imipramine were examined for possible associations between pretreatment responses to the Minnesota Multiphasic Personality Inventory (MMPI) and their behaviorally-rated responses to this drug. Each patient was randomly assigned to one of two groups with the restriction that the number of responders and nonresponders be balanced for sex. Results from the group A patients suggest that responders and nonresponders to imipramine were best identified by using two sex-specific, empirically-derived, MMPI scales. In contrast, 12 of 13 regular validity and clinical scales and all 49 of the selected special scales of the MMPI failed to separate responders from nonresponders. In the cross-validation study with the group B patients, we were able to predict female and male responders from nonresponders by the new imipramine response scale with accuracy rates of 93 and 100%, respectively. The implications of these results are discussed.