Abstract. Action Research was used at the Saskatchewan Hospital, Weyburn in the 1950s. The research influenced the hospital and the field of psychology. At the heart of the approach was intense collaboration between the researcher and hospital staff who were the end users of the information. Action Research is contrasted with two other models, the Academic and the Applied. Two applications of action research at Weyburn are described, followed by subsequent usage in California with consumer cooperatives and an organization for families of mental health clients. Collaboration with lay individuals changes the research process and requires trade-offs between rigour and relevance.The Saskatchewan Hospital, Weyburn provided my first opportunity to engage in action research (AR), an approach also known as action science (Argyris, Putnam, & Smith, 1985), participatory action research (Whyte, 1991) and praxis research (Honadle, 1996). There are similarities between AR and Schon's (1983) model of reflective practice, in which a person simultaneously engages in research and practice using theory as a framework while fully respecting the discrete nature of each case. AR emphasizes the practical use and dissemination of research products and builds utilization strategies into the overall research design (Ketterer, Price, & Politser, 1980). My usage can be traced back to Lewin's (1946) conception of AR as requiring the active involvement of the potential users of the information throughout the research. This is not business as usual for the researcher but a different kind of business. According to Parlee (1983), Psychological knowledge would be dramatically changed if it were consistently developed through interaction with its 'subjects' and its intended audiences instead of being developed with professional colleagues in mind and 'given away' (p. 1). AR tends to be value-driven rather than value-neutral and has several related objectives: to improve the lives of the participants (a good test of the success of AR is whether the participants are better off after the research than they were before); to advance knowledge; and to improve the practice of AR through a critical examination of the collaborative process (Sanford, 1970).Prior to coming to the Saskatchewan Hospital, I followed the conventional academic model of research. It took many years before I realized the epistemological breakthroughs made at Weyburn, not so much in content but in approach. I became aware of the assumptions underlying the academic model and how these differed from a collaborative approach. Table 1 shows three research models, the Academic, the Applied, and Action Research, as ideal types. There are hybrids and variants of each model, omitted here in the interests of parsimony.Let us travel back in time over 40 years to Saskatchewan Hospital, Weyburn. In 1957 this was one of the most intellectually exciting places in the mental health landscape. The provincial CCF (now NDP) government encouraged experimentation with new social programs. Cutting edge work was undertaken in community mental health, architectural psychology, psychedelic drugs, operant conditioning, and hospital administration. Psychologists were part of the intellectual ferment and the research done at Weyburn influenced both the hospital and the field of psychology. Ayllon's behaviour modification ward sparked interest in applied behavioural analysis in institutional settings (Ayllon & Michael, 1959) which stimulated the creation of new journals and organizations. The focus on the behavioural effects of architecture gave impetus to a new specialty of environmental psychology.With a vigorous commitment to the value of research-based knowledge, this large Kirkbride-style mental hospital, probably the largest single building between Winnipeg and Vancouver, came very close to being what Campbell (1969) called an experimenting society and Hutchins (1968) described as a questioning community. …