Using ratings of the Group Assessment of Interpersonal Traits (GAIT) as a criterion of psychotherapeutic skill, the relative value of academic criteria versus therapy-analytic criteria in predicting therapeutic skill was examined. Two measures of traditional academic criteria were used: scores on the psychology subtest of the GRE and course exam scores. Academic measures did not predict therapeutic skill whereas therapy-analytic ability did. This finding argues for the use of therapy-analysis in undergraduate courses in mental health delivery disciplines and in selecting therapeutically-talented nonprofessionals. Therapy-analytic ability was also found to be more highly correlated with therapeutic skill for the GAIT performance done in the first week of a 10-week course on clinical psychology than for the GAIT performance done in the tenth week of the course, suggesting that the predictive value of therapy-analysis is greater for the untrained than for the trained nonprofessional. This study also analyzed for the effects of training by using Pre- and Post-GAIT ratings.