The demonstration project is becoming a major instrument for social planning. In sponsoring demonstration projects the overall goal is for small scale "pilot" programs which include some form of research to contribute to program change and policy-making (14, 16, 19, 21). It is generally expected that the lessons learned from demonstrations, through the rigours of scientific research, will somehow result in large scale adoption and major shifts in aims, styles and resources, and effectiveness of social service programs. Models or prototypes for future operational programs are tested to determine their effectiveness in meeting states objectives. For this reason, they are undertaken in natural settings which presumably resemble the non-experimental conditions in which such programs might be later introduced.

This paper will attempt to elaborate the incompatibility between the two primary goals of a demonstration project - evaluating the program and using it as an instrument of change. It will show the manner in which the aim of using the demonstration project as a rational planning instrument is undermined when research and change strategies are pursued simultaneously rather than sequentially. By examining a nationally prominent demonstration - the St. Paul Family Centered Project - we can isolate relevant issues and illustrate how this incompatibility is manifested. Although there is limited generalizability from case studies writings by Marris and Rein (10) as well as Moynihan (11) support the findings of this analysis. The advantage of a detailed examination of a case study is that it provides an empirical and vivid illustration of a phenomenon which appears widespread. Moreover, it yields questions for large-scale research. Therefore, this paper is not meant to describe the history of the Family Centered Project. Rather, this Project provides the empirical data for grappling with the major issues - i.e., the demonstration project as a research and change strategy.