This report critically reviews past benefit-cost studies of manpower programs and offers an alternative methodology for program evaluation. Because of selective admission criteria and a tendency to attribute all income changes to training, previous studies have had an upward bias in measuring the effectiveness of training. This study evaluates a community effort manpower program in which trainees were not selected on the basis of aptitude or intelligence. Nevertheless, the trainees were far from typical, so multiple regression analysis was used to isolate the effects of training on income. The results show a higher benefit-cost ratio for on-the-job training than for institutional training, especially for women. The report concluded that the program as a whole represents an efficient allocation of resources. (BH)