A Critique of Generalizability in Interpretive Research

This paper is about generalizability in interpretive systems research. The authors are concerned that, in its search for appropriate inquiry methods, the discipline of Information Systems (IS) does not slip into the errors of other social disciplines and become a dismal science. Many systems thinkers have repeatedly argued that the purpose of IS research needs to be to produce what Argyris calls actionable knowledge and Ulrich, critique heuristics. That is, rules of thumb which managers can use to solve design problems, not too vague and not too detailed. This is a different concept from generalizability, which aligns with the scientific notion of seeking universal objective laws. The paper uses the argumentative inquiry in its reflexive capacity to critique knowledge claims on a chapter by Baskerville and Lee that discusses generalizability to bring out the differences between this cornerstone of universal truths and the more systems thinking concepts of actionable knowledge or critical heuristics. THE PROBLEM Baskerville and Lee (1999) wrote a chapter entitled "Distinguishing Among Different Types Of Generalizing In Information Systems Research" in which they refine the definition of "generalization" for the benefit of interpretive empirical inquiry. My9 concern is that they may have inadvertently led new researchers to think that generalizability is in some way central to interpretive systems research when this richer research tradition has a very different set of priorities. At least five different research perspectives provide examples of these priorities, notably in the now extensive argumentative (Rehg, 1999), systems inquiry (Churchman, 1971), perspectival thinking (Linstone, 1999; Haynes, 2000), actionable knowledge (Argyris, 1996) and critical social theory (Alvesson and Skoldberg, 2000) literature. Moreover, any discussion of generalizabilty might want to address Popper's (1963) critique of induction. The alternative priorities in interpretive systems research can be loosely clustered into one single claim that social inquiry is best seen as the justification of a perspective (insights, falsifiable conjectures). This justification may or may not use measured empirics, as empirics are understood to be actions to assist thinking, or merely illustrations intended to improve the human condition. Generalization is not central to this view of empirics. So, more specifically, this paper will explore the conjecture that interpretive inquiry offers alternative priorities than generalizabity that are more appropriate for the "wicked" problem of Information Systems (IS) related inquiry. Rather than merely list the attributes of these alternatives, they will be used as the lens with which to critique Baskerville and Lee's chapter constructively. It is hoped this will serve two purposes. The first is to reveal how the work of these five perspectives might be used systematically to critique any social inquiry and so reveal underlying assumptions. Encouraging critique methods lies at the core of interpretive inquiry. The secondary purpose is to give further airing to the evaluation criteria these five perspectives suggest for interpretive inquiry, empirical or not. In line with most inquiry methods, some upfront working definitions may be helpful to establish common ground. My understanding of generalization is that it is the concept of being able to extrapolate from a sample onto the population from which it is drawn, as a whole. Interpretive IS inquiry, whether empirical or not, is the search for justified interpretations (perspectives) of events or things, maybe distilled from the multiple interpretations of others in order to improve the appreciation of the researcher (system designer). By argumentation is meant the process of making a one line claim, conjecture, conclusion or proposition that has to be justified (well reasoned) to a knowledgeable cynical audience using convincing evidence. It includes anticipating the counter evidence. Objective knowledge is defined in terms of Popper's (1971) cultural world 2. It is knowledge agreed upon by the majority of a culture to the extent it has become independent of any one member of that culture, and science is a culture. Objective-empirics are typically observations that are repeatable by skilled people at any time in any situation. A critique is more useful if the reader is familiar with the proposition, in this case Baskerville and Lee's chapter. For those readers who are not familiar with this chapter, their "conclusion" is provided again here. 4. Conclusion