This post is dedicated to my friend and mentor Dr. Paula Meek, who passed away on July 29th. Dr. Meek was crucial to the development of the model described below, and it was she who suggested the "shape shifter" metaphor as its name. Dr. Meek was also a coauthor of my 2018 paper describing Two Minds Theory, and had a major impact on my thinking during her tenure at CU Nursing. I join her many students and trainees in saying that Paula will be deeply missed.
Does your mental state cause your behavior, or does your current physiology (influenced by behaviors like sleep, diet, and exercise) cause your mental state? The answer to this question is of course "yes." Or it's "both/and," as we Lutherans like to say about many of life's mysteries. My colleagues and I addressed questions of biology and behavior in a 2012 paper describing the "shape shifter" model for biobehavioral research. Although originally intended as a tool specifically for understanding symptoms in nursing theory, the shape-shifter model can be used more generally to understand any complex interaction between mind and body.
The example above (from Cook et al., 2016) shows a cycle in which people who feel tired ("fatigue" on the left-hand side of the diagram) are also more likely to experience a range of negative feelings (worse mood, more stress, perceived lack of control, lack of motivation for completing tasks like medical treatments). Besides the emotions, they also could have more passive coping behaviors, experience more stigma in their everyday interactions, and actually receive less social support from others -- e.g., because they stay home and interact less. All of those experiences could then lead them to feel even more tired ("fatigue" on the right-hand side of the diagram), a classic vicious cycle.
In behavioral research, it's often helpful to think of something you are studying as either a "phenomenon" -- the object under study -- or a "determinant" -- the cause behind that phenomenon. Notice that in this diagram, fatigue is both phenomenon and determinant but at different times. The variables in the middle of the diagram have an intermediate status, as both phemonena caused by fatigue, and determinants that cause later fatigue. This type of cyclical model often confounds behavioral scientists, because our statistical models require us to specify some variables as "independent" or "predictors" (i.e., determinants), and others as "dependent" or "criterion" variables (i.e., phenomena caused by other things). A variable quite obviously cannot be its own cause -- that is a tautology. The shape-shifter model gets around this limitation by specifying a time element to the process. Fatigue on the left-hand side of the diagram above is on one day (e.g., a Thursday), and fatigue on the right-hand side is on the subsequent day (e.g., Friday). Adding the time element allowed us to test the model above in our 2016 paper, and to confirm many of its predictions.
Of course, adding a time element helps only partly. People often want to know the "real cause" of some phenomenon. In fact, that desire almost derailed our efforts to write the original shape-shifters paper in 2012: Some of my colleagues had a more biological research orientation, and thought that of course someone's fatigue is caused their physiological state, depending directly on things like how they slept the previous night. Others had a more behavioral research focus, and thought that of course the subjective experience of fatigue was able to affect a person's behavior and thereby their biology -- for instance, if they felt more tired they might be less active, which would in turn lead them to sleep less well. When some of our coauthors wanted the causal arrow to point only one way, or argued that the "true" measure of fatigue was in fact sleep quality or duration, we found ourselves at an impasse. My colleague Paula Meek and I agreed that symptoms are only accessible via patients' self-report -- objectively measured sleep quality is not the same as feeling well-rested, which is a subjective experience that is knowable only by asking someone. My colleague Betsy Corwin, a physiological researcher, argued that of course the subjective experience was caused by the physiological state. My colleague Nancy Lowe solved this conundrum by proposing that what we faced was in fact a measurement issue: Even if a researcher defined the "real" phenomenon of interest as sleep, one could point their measurement tools at the subjective experience of fatigue. And even if one was primarily interested in participants' subjective reports, one could still use measurement tools like sensor devices to gather physiological data that relate to participant report in a greater or lesser degree.
This might seem like mere scholastic argumentation (how many angels can dance on the head of a pin?), but our assumptions about the "real" nature of phenomena can obscure important details. In the example above, if we assume that mood is the "true" determinant of all the other variables, we could miss important relationships between fatigue and social support -- for instance, we might think that depressed people just feel unsupported by others, without considering changes in their behavior that lead them to actually receive less social support. If we gather ongoing data using multiple measurement strategies, we can test some of these potential patterns directly. In its recommendations to study phenomena using multiple tools (e.g., sensors and surveys) and to conduct longitudinal within-person research (to account for bidirectional causation over time), you can see how the shape-shifter model prefigures key research recommendations related to Two Minds Theory.
To answer questions about causation, philosopher David Hume (1711-1776) recommends three criteria:
1. Two phenomena A and B occur together, so that when A occurs, B usually also occurs. This is the modern definition of correlation, and as any introductory student of the social sciences knows, correlation does not equal causation. (This principle wasn't commonly accepted until after Hume, though -- it seems to have been first stated by Francis Galton in 1889).
2. There is temporal precedence, such that A always comes before B in time. Hume seems to be the first person who added a formal time element to causation, and adding time to the analysis was our group's solution to the backward-causation problems created by the model at the top of this page. Recently, physics has suggested that true backward causation is in fact possible, with mathematical models of the universe running just as well backwards as they do forwards. But for the past 3 centuries, backward causation (or teleology, an intention-focused mechanism that Aristotle called an event's "final cause" as opposed to its Humean "efficient cause") has been seen as a logical fallacy. For the most part, scientists still believe that causes can't follow effects.
3. Finally, Hume demanded that there be a logical connection between cause and effect. This criterion has been frequently cited over the decades, but also frequently abused. Why, in the final analysis, should it matter to nature whether we can understand it or not? Einstein ridiculed quantum physics as "spooky action at a distance" because it didn't make sense to him, but it nevertheless is our best explanation of causation in some subatomic scenarios. Still, causal models are most acceptable if people can understand them in terms of their established views of the world.
The "logical connection" requirement is where people can get stuck in thinking about biobehavioral phenomena and determinants. The philosophical pull of material reductionism leads many scientists to assume that when a phenomenon is measured biologically, then biology must also be its ultimate cause. But the shape-shifter model illustrates how Hume's other two criteria, correlation and temporal precedence, can suggest alternative causal orderings. It behooves researchers to at least try to wrap their minds about the behavior-causes-biology direction of the causal arrow, because the true underlying nature of reality is often stranger than we think.
Comments
Post a Comment