The program gives temporary employment (unskilled jobs) to participants, who work in crews of peers. Work standards are gradually increased, and good performance and attendance are encouraged by a system of bonuses. In 10 of the demonstration's 15 sites, a sample of applicants were randomly assigned either to a program or a control group. Both groups were followed, through interviews, from the time of assignment for up to 3 years. All benefits and costs were measured and compared. The demonstration appears to be cost effective for long-term AFDC recipients and for ex-addicts. The long-term benefits exceed costs by over $8,000 per AFDC participant and over $4,000 per ex-addict participant. However, Supported Work for youths is not economically efficient. Long-term social benefits fall short of costs by about $1,500 per youth participant. The results for ex-offenders are incomplete, and definite conclusions cannot be drawn. Both ex-addicts and ex-offenders received substantially more income because of the program. Taxpayers are better off as a result of the Supported Work participation of AFDC and ex-addict target groups and worse off with respect to the participation of youths. Data tables, footnotes, and about 50 references are supplied.