We begin with a review of the standard criticism that AR is unscientific. Some of our AR colleagues address this issue by accepting the idea that science itself is necessarily inhumane and alienating. They are proud to be called unscientific. We disagree with them vehemently and argue instead that AR can, does, and should produce valid and meaningful social research results.
In academic circles, AR, applied research, and most qualitative research are generally denigrated as “unscientific.” Although conventional social science researchers occasionally admit that some AR is useful, they generally argue that AR findings are anecdotal, based on telling stories rather than on doing science. Indeed, most conventional researchers behave as if they believe that useful work is, by definition, scientifically trivial. In these circles, doing science is equated with being objective and rigorous, using statistical tests, using at least quasi-experimental controls, and staying away from the world of application. In this regard, they throw their entire weight behind what Nowotny, Scott, and Gibbons (2001) call Mode 1 knowledge production—“reliable knowledge”—while AR produces knowledge in the context of application or Mode 2 knowledge that is “socially robust.”
It is not as if we are inventing criticisms that have never been leveled at the social sciences. The conventional social sciences have been criticized over the years for their often-questionable scientific practices. Critics of the contemporary social sciences often claim that these fields erred in accepting classical physics and chemistry as the model of science (see Clifford & Marcus, 1985; Geertz, 1973; Porter, 1995; Rabinow & Sullivan, 1987). For some of these critics, the proper response is to repudiate science and to advocate perspectives that challenge the very existence of generalizable knowledge. Others argue for various reforms in social science practice, such as doing more “relevant” research, but without understanding that a serious engagement with relevance requires a fundamental change in social research approaches.
Many critics say that the social sciences have been captured by some kind of mechanistic and ritualistic error in conceptualization (for example, Barnes & Shapin, 1979; Mills, 1959), to be remedied by a variety of cures ranging from hermeneutics to structuralism to deconstruction. According to this view, the social sciences have become derailed from methods appropriate to them. This is a “tragic” narrative, and it usually argues that the social sciences have not paid sufficient attention to the dimensions of social phenomenon that do not exist in the sciences (for example, intersubjective understandings, language). Of course, a particular academic agenda emerges from these criticisms, including a struggle for power and influence in the academy, a struggle that is increasingly intense as we see in battles between the rational choice positivists and the cultural studies practitioners in the social sciences.
We do not dispute the need to change the agenda, but this imagined history of the social sciences does not correspond to the findings from research on the history of the conventional social science disciplines nor to the situations we have experienced. This is not the place to rehearse the purging of engaged and reformist work from the core social sciences (see Madoo Lengermann & Niebrugge-Brantley, 1998; Messer-Davidow, 2002; Ross, 1991) and their retreat into jargonized irrelevance, but this is a significant part of the story. The other part of the argument that we wish to concentrate on arises from a misunderstanding of science itself.
Research in the physical and biological sciences does not match the stereotype of scientific research that these critics unknowingly (we suppose) use. Rather, we believe that much research in the sciences can best be understood as a successful and disciplined form of repeated cycles of testing and reformulation and on a clear relationship between thought and action. In other words, the sciences indeed are radicaUy different from the contemporary social sciences, but only because the contemporary social science practices have very little to do with scientific practices.
One mistaken notion, incorrectly borrowed from the sciences, is that social scientists should be completely disengaged (actually and intentionally) from the phenomena they study. Equating social disengagement with objectivity, impartiality, and the requirements of scientific practice, these practitioners systematically distance themselves from their research subjects. Having achieved a distance from which it is all but impossible to understand human actions, they then further insist on separating science from action. This move severs the connection between thought and action that permits the testing of results, making the conventional social sciences quite unlike either the physical and biological sciences and quite unlike AR.
We do not take on the larger issues about the meaning of science itself and simply assume that it is useful to consider the physical and natural sciences to be scientific in some meaningful sense.3 Our focus is that AR’s pursuit of constant and disciplined interactions between thought and action resembles research in the physical and biological sciences far more closely than do the practices of conventional social science.
At the heart of this problem is the tremendous emphasis conventional social scientists place on their claim that being scientific requires researchers to sever all relations with the observed and to avoid being co-opted by the seduction of their own prejudices. Such social scientists equate objectivity with disengagement from the phenomena under study and demonstrate both arrogance about their own capacities to understand other human beings and about the incapacities of their research subjects to offer conceptual analyses of their own behaviors and situations. This belief and practice undermines the argument that conventional social science practices scientific methods precisely because biological and physical scientists do not disengage themselves from the phenomena they study to be objective. The experimental method requires just the opposite—it requires engagement, albeit on particular terms. The scientific method and its experimental apparatus are a form of praxis on and in the world, though certainly not one oriented around democratic social change.
Viewing social research this way is not a new idea, but it has been suppressed as conventional social scientists and the social interests their work serves have turned away from social engagement and social reform. Kurt Lewin, the social psychologist we introduced earlier (Chapter 2) and an early proponent of AR, operated with a view of social research as both scientific and socially engaged. As we stated there, his view of the matter is summed up in the two often-repeated statements attributed to him: “Nothing is as practical as a good theory” and “The best way to understand something is to try to change it.” He articulated these views in the 1930s and 1940s, echoing the earlier ideas of the famous pragmatist philosopher of democracy and education, John Dewey.
We make a fuller presentation of these arguments in Chapter 6. What is important about Lewin and Dewey in the present context is that their approaches to social research are in concert with the way contemporary scientists think and behave, and Lewin understood clearly the link between AR and the scientific method.
Rather than pursing the contention about AR and scientific method in the typical manner of such discussions (that is, through more pronouncements), we make the case by narrating an episode that illustrates our claims.4 Our purpose is to clarify the implications of our argument that AR is more capable of producing valid results than is conventional social science5 and then to examine why conventional social science has deviated from this course.
Source: Greenwood Davydd J., Levin Morten (2006), Introduction to Action Research: Social Research for Social Change, SAGE Publications, Inc; 2nd edition.