Project START, a pilot program operating between February 1976 and March 1978, was a cooperative effort among law enforcement agencies and citizens. It aimed to reduce the repeat rate of nonviolent property offenders through an enriched probation progam offering client diagnosis, service brokerage, employment counseling and referrals, community education, and training for agency representatives, probation officers, and community volunteers. The evaluation design used random assignment of probationers to the project or to a regular probation control group, a purposively selected group of parolees for comparison with probationers, multivariate statistical methods to generate a predictive model of the various program outcomes, a posttreatment interview of a sample of clients, and an ongoing process feedback to the program regarding its inner functioning. Rather than attempting to declare whether Project START works or does not, the evaluation sought to discover what kinds of effects if any, the project had on the various classes of clients served. The evaluation found that START increased service usage by about 350 percent over regular probation when only probation-referred services were considered; including nonprobation sources, START's increase was 29 percent. The project also exhibited a superiority in client tracking after referrals were made, and clients spent more time on jobs than did the controls. However, no effects on recidivism could be discovered attributable to service usage, employment, or even START clienthood, and no client characteristics were discovered bearing on recidivism. An independent cost-benefit analysis yielded a favorable social net present value for START. It is thus concluded that intensive service delivery to probationers cannot be expected to reduce or increase recidivism rates or crime rates. Tables, figures, and footnotes are supplied. Appendixes contain an evaluation report of the Citizen Power in Criminal Justice Conference organized by START and evaluation research instruments.