The dual imperatives of action research

TLDR

Critics argue that action research is merely consultancy, lacks causal inference, generalizability, is prone to bias, and falls short of rigorous research standards, concerns the authors acknowledge while noting deficiencies in information systems practice. The authors contend that these problems stem from how action research is currently conceptualized. They propose a deeper reflective analysis that culminates in a model incorporating both a problem‑solving interest cycle and a research‑interest cycle. The model’s implications are illustrated with examples drawn from a real‑life action‑research study.

Abstract

Action research (AR) is not without its critics, and those who reject some of the paradigmatic assumptions embodied in AR maintain that AR is little more than consultancy, that it is impossible to establish causal relationships, that it is difficult to generalize from AR studies, that there is a risk of researcher bias, and that generally speaking, it lacks some of the key qualities that are normally associated with rigorous research. The authors are sensitive to such criticisms, for although they are committed action researchers, they have elsewhere voiced their concerns about the quality of AR practice in the field of information systems. The authors argue that part of the issue concerns the way in which we currently conceptualize AR. In this article, the argument for a deeper and more reflective analysis of the meaning and full implications of AR is developed, culminating in a model of AR being developed that explicitly includes both a problem solving interest cycle and a research interest cycle. Important implications of this new model are articulated, with examples to illustrate these points being drawn from a real‐life AR study.