The most noteworthy development in recent years with regard to evaluation has been the considerable extension in the field covered. In the case of mental health care, the concept of evaluation has accordingly broadenend out into policy evaluation, where originally it was limited to the measurement of effect. At the same time the experimental research design has given way to a longitudinal approach. In the English-language literature, the term `evaluation research' has been replaced by `program evaluation'. With reference to mental health care, this latter term now includes all kinds of evaluation, from the evaluation of a particular type of treatment to the evaluation of government policy. In this article we reduce all these forms to a single basic model: a cyclical process of exploration, setting of standards, operationalization, verification and adjustment. We then elaborate the model so that it can serve as a guideline for the establishment of any form of program evaluation. The model can also be an aid to the analysis of problems which occur in the course of ongoing evaluation processes.