The other day, I was looking at a model of transition to motherhood class of nursing PhD students. The authors had used grounded theory to identify stages in the process by which women entered the role of being a mother. As we were puzzling over the diagram — a group of labeled, overlapping circles — one of the students remarked that “this is just like the transition to being a PhD student.” We laughed, and indeed, all of the categories fit perfectly. First you felt unready and lost, then you figured it out, and finally you were comfortable; transition complete.
The categories of the model, it turned out, fit any transition whatsoever. They captured “transition to pizza delivery drver” just as well as “transition to motherhood.” The problem is that the categories were vague, even metaphorical, and that the relationships among the stages was not articulated. One stage was said to “cause” another, but there was no description of what this cause amounted to.
So, here’s the lesson: if there is an absurd application of your model, something is wrong. If your model can be parodied by making it easily fit pizza delivery or window washing, when it is supposed to be about a specific social process, then model is under-specified. A model that applies indiscriminately says nothing.
Tags: Theory evaluation