I just finished reading a methods paper and throughout they kept using the phrase “based on theory”. This is a phrase that I see all the time and not only in methods papers. As far as I can tell, what they meant was that many decisions you need to make in conducting a study or analyzing data should be based on a firm foundation, which is certainly true. But the phrase based on theory means little in the context where it is often used. To me it is more of a meme.
Why Based on Theory Means Little
In the organizational sciences, the word ‘theory’ is used in a very loose sense. It refers to conceptual thinking about a phenomenon. It might refer to an assumption, a supposition, a framework, a speculation, or just an observation. It does not usually refer to a specific theory. If I observe that two variables are related, I might speculate about a reason and propose a hypothesis that the two variables are related because of a third variable. Merely proposing one or more tentative explanations for a given observation is not a scientific theory.
The organizational sciences are dominated by model tests using complex statistical methods such as structural equation modeling. These models propose a sequence of events that occur over time and typically fit an environment-perception-outcome or E-P-O framework. Certain things occur in the environment, people perceive them, and they lead to outcomes. For example, poor leadership leads to job dissatisfaction that leads to employee misbehavior (i.e., counterproductive work behavior or CWB). The different paths in these models are justified by prior studies showing that they are correlated. We know that job satisfaction and CWB are correlated. The E-P-O framework justifies the causal order, that satisfaction leads to behavior. But that is not the only explanation. Cognitive dissonance and other cognitive consistency theories in psychology would say the opposite. People who engage in CWB will change their satisfaction to remain consistent between attitudes and behavior. The linkages in these models are based on theory only if by theory you mean an articulated idea or an argument that makes logical sense. You just need to convince peer reviewer that what you propose is novel and advances understanding.
Base Propositions on an Actual Theory
A scientific theory is an explanation of a phenomenon based on facts or data. As noted by Woodward and Ross in the Stanford Encyclopedia of Philosophy, a theoretical explanation is a statement that describes a phenomenon and then accounts for it with a sound argument that is based on established facts. In other words, theories are built on prior research that has collected systematic data on a phenomenon. This is different from theory salting where people just sprinkle the names of theories throughout a paper.
To base something on theory should mean that you first articulate the theory and then explain how that theory leads to whatever it is you are claiming. If you provide a model, the basis of that model should be an overarching theory that tells you which variables to include, and the order in time over which one leads to another. A model where each link between variables is “informed” by a different theory, or just someone else’s assumption or speculation is not basing that model on theory.
It is not easy to develop a true explanatory theory or a theory-based model, which is why we have so few examples. Unfortunately, the over-emphasis on theory has led to a situation where studies based on theory mean little.
Image generated with DALL-E 4.0. Prompt “Cartoonish picture of scientist looking puzzled with question marks over his head.”
SUBSCRIBE TO PAUL’S BLOG: Enter your e-mail and click SUBSCRIBE
Thank you!