The JOBSTART demonstration program provided education and vocational training, support services, and job placement assistance to educationally disadvantaged dropouts aged 17-21 at 13 sites. An evaluation assessed whether helping disadvantaged dropouts increase their educational attainment led to increased earnings. A total of 2,312 people were randomly assigned to an experimental group that received JOBSTART services and a control group that did not. Data from 1,941 youth for whom 48 months of follow-up data were available were analyzed. Findings indicated that JOBSTART led to a significant increase in the rate at which the youths passed the General Educational Development examination or completed high school. Youths in the experimental group earned less on average than those in the control group during the first year of follow-up. In the final 2 years, experimentals' earnings appeared to overtake those of controls for the full sample. Encouraging earnings impacts included those for young men who had been arrested between age 16 and program entry and for young men and women who had dropped out of school because they had educational difficulties. Earnings impacts were very large for one site: Center for Employment Training, San Jose, California. Overall, JOBSTART led to little change in youths' receipt of public assistance. From the perspectives of taxpayers and society as a whole, the investment in JOBSTART services was not repaid through increases in earnings or other quantified benefits by the end of the follow-up period. (Appendixes include descriptions of data sources, methodological issues of the JOBSTART impact analysis, description of the cost estimation, data tables, and 84 references.) (YLB)