THE SUPPORTED WORK DEMONSTRATION WAS A PILOT FOR POSSIBLE FURTHER PROGRAMS, CONDUCTED AT 15 SITES, 10 OF WHICH WERE USED FOR THE RESEARCH. A TOTAL OF 6,616 INDIVIDUALS MADE UP THE RESEARCH SAMPLE, WITH 3,214 AS PARTICIPANTS AND 3,402 AS CONTROLS. EACH PERSON RECEIVED A BASELINE INTERVIEW AND UP TO 4 FOLLOWUP INTERVIEWS AT 9-MONTH INTERVALS. EXPENDITURES FOR THE DEMONSTRATION AMOUNTED TO $82.4 MILLION ($11 MILLION FOR THE RESEARCH). EACH SUPPORTED WORK PROGRAM HAD A RANGE OF WORKSITES FROM CONSTRUCTION PROJECTS TO DAY-CARE CENTERS. AFTER 5 YEARS, THE GREATEST PROGRAM SUCCESS WAS IN PREPARING FOR EMPLOYMENT A SUBSTANTIAL NUMBER OF WOMEN WHO HAD BEEN ON AID TO FAMILIES WITH DEPENDENT CHILDREN (AFDC) FOR MANY YEARS; THESE WOMEN SOUGHT AND MAINTAINED EMPLOYMENT. THE PROGRAM ALSO HAD AN IMPACT ON A SIGNIFICANT SEGMENT OF THE EX-ADDICT POPULATION, WHO DID BETTER IN GETTING JOBS AND IN EARNINGS THAN MEMBERS OF A CONTROL GROUP NOT IN THE PROGRAM. IN ADDITION, PROGRAM EX-ADDICTS WERE LESS PRONE TO COMMIT DRUG-RELATED AND OTHER CRIMES. FOR EX-OFFENDERS, THE PROGRAM HAD A MARGINAL IMPACT--THEY DID NOT SHOW LESS CRIMINAL BEHAVIOR, AND THEIR EMPLOYMENT RATE AND EARNINGS WERE ONLY SLIGHTLY BETTER THAN THAT OF NONPARTICIPATING EX-OFFENDERS. NO LONG-TERM POSITIVE RESULTS WERE REPORTED FOR THE YOUTH GROUP. THE USE OF A NONPROFIT INTERMEDIARY ENHANCED THE ABILITY OF GOVERNMENT AGENCIES TO PURSUE A COMMON OBJECTIVE AND TO HARNESS GOVERNMENTAL AND PHILANTHROPIC SUPPORT TO AN AGREED SOCIAL GOAL. RECOMMENDATIONS POINT TO MORE USE OF OTHER NATIONAL WORK PROGRAMS, CONTINUED USE OF SUCH SUPPORTED WORK PROGRAMS, AND FURTHER RESEARCH WITH OTHER HARD-TO-EMPLOY GROUPS.