theory development as goal displacement

William McKinley has an interesting essay about the state of organization theory in the most recent Organization Studies. I’m not sure that I agree with McKinley’s sentiments, but his argument is provocative and worth considering seriously. Here’s a portion of the abstract:

In this essay I argue that organization theory has witnessed a significant displacement of ends over the last 30 years.Whereas in the 1960s and 1970s the dominant goal of the discipline was achieving consensus on the validity status of theories, today the overriding goal appears to be development of new theory. Formerly new theory development was considered a means to the end of attaining consensus on theory validity, but was not the only activity deemed necessary to accomplish that goal. In addition, instrumental standardization and replication were viewed as important. The contemporary displacement of
ends toward new theory development creates the paradox that organization theory today is both epistemologically simpler (in terms of the intellectual activity deemed desirable) and more complex theoretically than it was 30 years ago.

McKinley’s argument is similar to something Donald Hambrick wrote in AMJ a few years ago, in which he claimed that organizational scholars have a “theory fetish” that “prevents the reporting of rich detail about interesting phenomena for which no theory yet exists.” McKinley takes it a step further and claims that we have too many theories and that now would be a good time to start testing some of them. While it’s true that most theoretical advancements are accompanied by empirical analysis (e.g., ASQ rarely, if ever, publishes a paper without some empirical analysis), he argues that these findings are not often replicated, leading to a proliferation of new theoretical propositions that never get past the initial stage of investigation. The lack of replication is not limited to organizational theory, of course. Jeremy Freese has claimed that sociology suffers from the same avoidance of replication.

Rather than focus so much on theoretical development, McKinley would have us focus more (or at least as much) on theory testing and replication. He even suggests that we might consider dividing the labor of the field between theorists and empirical researchers. I remain skeptical, in part, because I’m not sure that trying to develop theoretical consensus is a good thing or even practical. In his essay he actually illustrates the difficulty of a “validity status” approach to theories. In the 1960s a rash of scholarly activity developed around testing hypotheses about the relationships of different structural elements of organizations. Debate erupted when the Aston group and John Child disagreed about how activity structuring related to authority concentration. I don’t want to lay out the whole history of the brouhaha here, but the point is that the groups couldn’t come to an agreement about why or if the two structural elements were correlated. McKinley reports that despite a decade’s worth of research on the matter, the disagreement was never resolved and eventually scholarship moved in a different direction.

So there are a couple of takeaways from this. The first is that empirical investigation in organizational scholarship will not always lead to the establishment of theory validity, or at least not consensus, because organizations are so contextually different from one another, making it difficult to build or validate theoretical propositions that hold true in every context.  Even when one does find a seemingly valid theoretical proposition, e.g., density–dependence, scholars will continue to debate about what the mechanisms are underlying the empirical relationship until eventually one or both groups get tired of arguing about it and move on to something else. My second takeaway is that it’s not clear that replication or theory validity testing is always very theoretical. Looking back at the Aston group, we wonder if there was any kind of theory in their project at all.  The main purpose of the project, from what I can tell, was descriptive not explanatory.  Once Child found an anomaly in his replication of their study, the real theorizing began (hmm, how do we explain this apparent difference in findings???).  So maybe I just have a difference in opinion with McKinley about what constitutes theory. Theory is not about establishing relationships between variables; theory is developing a view of the world that explains why things work the way they do. If you have good theories, you can start to derive hypotheses about particular empirical relationships. Or alternatively, you may have to alter your theories to make room for deviations observed in empirical research. This is why inductive research is important. But the ultimate goal of theory isn’t to just build more propositions; it’s about creating paradigms that make sense of our world.