ABSTRACT
Action research (AR) can be defined as a systematic attempt to move a social system towards a more desirable system state in such a way that participants learn something relevant from the process and are thus empowered. It is reasonable to assume that with such an exercise, learning will take place anyhow; the question, more likely, is to what extent this learning is explicit rather than implicit and to what extent it can be applied to other situations (‘portability’). Two approaches are developed to make AR more systematic, increasing the potential for explicit learning as well as its portability. The first resides in a rigorous thinking through of the AR process; the second is to adopt the principles and mindset of realist evaluation (RE) in the design and implementation of AR, acknowledging that observed outcomes are the result of context interacting with results-producing mechanisms, which can be captured in coherent theory explaining outcomes in causal terms. As a full-fledged application of realist principles is resource intensive, we conclude by examining some AR scenarios with variable requirements for depth and portability of knowledge produced. Some expert guidance may be indispensable in the case of ambitious AR based on realist principles.
