This study was designed to identify correlates of success in Supported Employment(SE) programs for persons with psychiatric disabilities. Indiana policy-makers are seriously considering a managed care, or "capitated," system of payment to make SE provider programs more efficient economically. However, many agencies are concerned about providing services to more severely impaired individuals because of the potentially higher costs of serving these individuals. Two studies are included in this project. The goals of the first study were to identify SE consumer (clinical) characteristics that predict (1) successful outcomes, defined as whether the consumer achieves gainful work, and (2) program costs, defined as the amounts of SE service hours utilized by consumers who obtain work. In two large samples of SE consumers with serious mental illness, no clinical characteristics (e.g., diagnosis, rated functioning, hospitalization history) were associated with vocational outcome or service costs. The goal of the second study was to describe the types and amounts of services utilized by SE consumers who obtain work. Specific service categories associated with obtaining work were travel, training, and advocacy that was unrelated to the consumer's job. The implications of these findings are discussed in the framework of the debate over clinical versus empirical prediction. The need for a theoretical model of SE services that allows the use of predictive clinical and consumer driven services is also discussed. / Department of Journalism