Research using Social Security data to measure the effects of government manpower training programs on the earnings of participants is discussed. Previous studies employed a longitudinal set of Social Security summary earnings records of over 50,000 individuals who participated in Manpower Development and Training Act (MDTA) institutional training in 1964 and 70,000 individuals randomly selected from the same file. This study used Social Security's Continuous Work History Sample (CWHS) to demonstrate that these prior studies have substantially underestimated the impact of training on earnings. The studies being criticized employed a model of income determination that did not take into account the fact that trainees were induced to join the program because they were having difficulty finding or holding adequate jobs. The basic technique used in this paper was to compare actual post-training earnings of trainees to the earnings of a control group whose earnings potential at the time of initiation of training was identical to that of the participants. In order to compare the two groups, a model of income determination was specified and estimated using regression analysis. Since it was not possible to identify trainees in the CWHS or to obtain a reliable measure of unemployment and loss of job tenure in the summary earnings records, the estimation procedure was based on knowledge of the simple correlations among the variables. (Author/MV)