14.9: Is it really necessary to draw theories?

Is it really necessary to draw theories? Theories are representations of our phenomena, and drawing theories forces us to make explicit all of our (or the theorist’s) decisions about the key constructs and their relationships. When “drawing,” we don't need to worry about our artistic skills—boxes and ovals, stick figures, and arrows are typically sufficient for depicting even complex theories and models. There is a clarity to drawing, or to moving around our pack of index cards (containing key constructs), that scaffolds clear thinking. We consider how constructs are potentially grouped according to individual and/or contextual attributes; we notice whether some constructs are embedded in other higher-order constructs. We register temporal and causal priority based on the convention that the time course moves from left to right, and that causes precede their effects. We indicate reciprocal relations by including both feedforward arrows and feedback arrows.

One of the things that we noticed about our students’ drawings is that the only context in which they have systematically been exposed to visual representations of their phenomena is in Chapter 3: Understanding a Theory 19 their statistics classes. And they have tended to adopt those depictions—based on multiple regressions, latent path models, hierarchical linear modeling, and so on. This can create a kind of “statistical creep” in which certain analytic techniques come to dominate our representations— and so shape our theories of phenomena. Instead of looking for analytic methods that allow us to test our genuine theoretical questions, we end up asking the theoretical questions that our statistical methods are designed to answer. So it was important to “defuse” conceptual form statistical representations, so that we could consider their fit in making decisions about design and analysis.

Finally, it was great fun to ask all our students to draw the same theory and then examine all the different ways that it could be represented. Some of the representations were “better” than others, in that they contained more information, were more consistent wit the theory’s propositions, or were easier to understand. And sometimes, some of the student representations were actually better in all those ways than the graphics offered by the theorists themselves! But often times, all the variety of representations were equally good—but just very different. Again, this allowed us to reflect on the multiple perspectives that can be brought to bear on any phenomena, and to seek out and appreciate ways of looking at them that are different from our own.