Ho do you actually go about creating a theory? While struggling with this question in a practical sense, I remembered Henry Mintzberg. Somewhere along the way, I learned about his dissertation and first book (The Nature of Managerial Work) in which he revolutionized the theory of what managers do. I remember being struck by the simplicity of his research design: “I will follow senior executives around for a while and see what they actually do.” Okay, what he actually said was: “I sought to develop by the process of induction a statement of managerial work … using a method called “structured observation”. “Structured observation”; it sounds so simple. Like Yogi Berra said: “you can observe a lot just by watching.” But the simplicity and understatement with which Mintzberg described his research does not begin to describe the intensity of his observation and record keeping, the categorizing and re-categorizing, the analysis and interpretation that pointed towards the concluded work.
In 1971, Mintzberg wrote: “We must observe, describe, and understand the real work of managing; then, and only then, shall we significantly improve it.” He was dealing with a theoretical framework for management that was 50 years old, that was based on thinking about how managers should act rather than observing how they do act, and that used words such as plan, organize, coordinate and control to describe objectives of managerial work rather than its actual content. It is my contention that this is where we find ourselves now with respect to the policy sciences and a theory of policy analysis communication.
The question of “the interesting question” leads me back to Mintzberg. I was pleasantly surprised to find that HM had written recently about the process of theory creation. Here we have a sketch of a map of the inside of the theorist’s mind, and I commend this article to you (see Henry Mintzberg: Developing Theory about the Development of Theory).
Some observations on what I thought were the key elements of this article:
1. Theories aren’t just falsifiable – they are implicitly false. Because theories are only a description of reality, not reality itself, they cannot be said to be true. I find this very comforting, and greatly reduced my hesitancy to embark on a process of theory creation. Think about how this idea can help ground the budding theorist, to diminish any sense of self-importance. You may think you’re pretty hot stuff – the young thinker knocking off the dusty old theories of dusty old professors – but you’re not. If you are lucky enough to create a theory, the same thing will happen to you someday – and soon. This, for me, makes the idea of theory creation exquisitely less daunting.
While were at it, the question of falsifiability needs to be addressed. My understanding of Popper and Kuhn regarding the nature of scientific investigation and theoretical change was that Popper critiqued positivism for failing to understand that facts were contingent on values, and that scientific knowledge goes through stages: tentative theories, subject to tests of falsifiability rather than proof, give rise to new tentative theories. And Kuhn’s work built upon Popper’s – especially in respect of its critique of positivist objectivity – but he is most widely cited for his approach to understanding how theories change.
But while I appreciate the post-positivist implications of Popper and Kuhn, there’s always been something about the Popper mystique that has bothered me. Mintzberg revels in the irony of the cult of Popper and is quite humorous in his scathing critique of the Propper Way (a term he uses because a secretary once mistyped it as such):
[W]hy, when he devoted the rest of his book to “the deductive method of testing” (p. 30), did Popper title his book “The Logic of Scientific Discovery”? What discovery is there in deduction? … And why have untold numbers of researchers-in-training been given this book to read as if it is science, and research, when it is only one side of these, and the side wholly dependent on the other (i.e., induction), which is dismissed with a few words at the beginning? … Popper devoted his book to deductive research for the purposes of falsifying theories. But … falsification by itself adds nothing; only when it is followed by the creation of new theories … do we get the necessary insights.
2. Speaking of Propper – there is an incentive in academia – and pressure on graduate students – to do proper research: quantitative, technically proficient, non-speculative. The trade-off that this bias leads to, says Mintzberg, is a preference for rigour over relevance; academics are forced to be methodologically correct rather than socially meaningful. Again, that I can spend years as a graduate student and be surprised that theory creation was a viable path is shocking in itself.
Can I pause for a moment and consider this question more precisely, which is another fundamental problem I am grappling with. Are we able to accept a dissertation that doesn’t involve any original research other than reading and thinking? Just what exactly qualifies as a dissertation?
In step 2 of Mintzberg’s 20 steps (see below), he speaks of needing to be stimulated by a body of rich inputs – tangible data, thick descriptions, stories and anecdotal data. However, he says this need not be “data” per se. He calls The Structure of Organizations his favourite book, written almost entirely on existing literature: the theories, research findings and descriptions of others. Is this something he was allowed to do because he did the structured observation dissertation first?
3. The issue of technical prowess leads to another bias in academia: for quantitative research over qualitative research. This distinction also gets manifest as an equating of quantitative with deductive research and qualitative with inductive research. I know where I got this from: Creswell’s book “Research Design: Qualitative and Quantitative Approaches” is fairly emphatic that quantitative = deductive and qualitative = inductive. Again, Mintzberg:
“Theories can be assessed without numbers … just as numbers can be used to induce theories …. This mix-up leaves the impression that ‘quantitative’ research is somehow proper (or Propper) – i.e., ‘scientific’ – even if it contributes no insight, while qualitative research is something to be tolerated at best, and then only when exemplary.”
After having vented for a bit, Mintzberg does accede to the request of the editors of the book to describe how he creates a theory – or, at least, the steps he goes through in the process of theory development. Just a few of his 20 steps to theory development include:
- Start with an interesting question, not a fancy hypothesis. Let yourself be pulled by big questions: What do managers do? What do policy analysts do?
- Bootstrap an outline: the “ultimate problem”. He doesn’t say how to do this, however, other than to just do it. Very frustrating.
- Diagram: express inter-relationships amongst concepts graphically. “I have been puzzled to find that some people are puzzled by my diagrams.”
- Keep the research method simple, direct and straightforward. Recall the simplicity of “structured observation”: go and look.
- Everything depends on the creative leap (which can be trivial): Keep yourself open to surprises, ask why trivial observations might be happening. Creativity isn’t such a rare talent – it just requires seeing what’s in front of you.