I’ve been a scientist most of my adult life. When I hear the word “theory”, I have a  particular meaning associated with it. To me, a theory is an explanation about how something works. A theory that has been proven to be true beyond all reasonable doubt can be considered to be a fact. When a theory about something is proposed, it is assumed that it will be tested in a rigorous fashion, the result of which is either to verify or falsify it. A theory that has no supporting evidence is really little more than an idea – it is not until sufficient supporting evidence has been gathered that the theory begins to be taken seriously. Often, experiments in science are designed to try and disprove a theory – much of what happens in science basically involves trying to break things. I’m quite comfortable with the notion that any theory I propose must be subjected to considered skepticism and careful scrutiny.

I have considerable difficulty getting my head around the notion of a theory the way it is typically understood in the social sciences and humanities. In art and philosophy, a theory is not measurable. Here, a “theory” can be quite loosely developed and poorly tested (see NIH p.4), and yet still gain acceptance and even legitimacy. The NIH manual “Theory at a Glance” admits that theories are abstract, not highly developed, and may not have been rigorously tested, yet it still talks about how important theories are as tools for evaluating and designing health promotion behaviours. It even claims that these loosely developed and poorly designed theories are important for planning effective programs. This is quite at odds with how theories are considered in science.

Much of the first part of the NIH manual reads likes a sales brochure – it glosses over the testing and verification portion of theory development and jumps right into the advantages that theory can provide in planning and development (over what?). It claims that adherence to a theory frees us, implying that the alternative is to follow a paint-by-numbers approach. Is it not possible that there are a other ways to find solutions to problems that do not fall into either category?

Somewhat in contrast, the Briggs paper does talk more about testing and verification. Here we learn that a theory is a causal model, that the phenomenon of interest is the effect and never the cause, and that all models are not created equal. Perhaps part of the message here is that theory is not as simple to use as the NIH introduction implies.