Formative evaluation seeks to measure whether a program achieves its goals rather than whether it is effective by absolute standards. Such evaluation focuses on the process rather than the results and requires collaboration between practicing therapists and researchers. Results should contribute directly to improvement of the program. The method was used by the Canadian National Institute of Scientific Research to evaluate the eight Montreal drug abuse and alcoholism programs. The research model permitted assessment of elements common to all the clinics as well as elements particular to separate clinics. Each clinic was questioned about client types, the nature of the treatment programs, and the influence of such factors as patient motivation, age, previous treatment, and treatment duration. Survey instruments were constructed to permit comparison of the programs. The Community-Oriented Program Environment Scale was employed to measure the diverse programs. The evaluation process aided various programs in clarifying their objectives, thus establishing the advantages of the model. These include the opportunity for programs to participate in their own evaluation, therby integrating evaluation permanently into the program and into the actual process of therapeutic intervention and sensitizing practicing therapists to certain aspects of their programs through the questions and demands of the research team. Improved reliability of data resulted from the cooperative atmosphere; a therapeutic environment was mobilized; and information procured on success and failure factors of treatment programs. Disadvantages of the approach are the dependence and constraints facing researchers in trying to work with busy therapists, reduced objectivity because of researchers' immediate proximity to the situation, and the slowness and relatively great expense of the technique. Several tables and a 9-item bibliography are supplied.