The problem of predicting the average effect of a new training program using experiences with previous implementations was investigated. The ability to adjust for population differences depends on the availability of characteristics of the two populations and the extent of overlap in their distributions. It was found that adjusting for pre-training earnings and individual characteristics removes various differences between control units that have some previous employment experience. It was also found that adjusting for individual characteristics is more successful at removing differences between control group members in different locations with some employment experience.