Leventhal's Common-Sense Model and Two Minds Theory

Leventhal, Diefenbach, and Leventhal's (1992) "common sense model" of self-regulation.

My 2018 paper describing Two Minds Theory (TMT) cites work by my colleague and coauthor Dr. Paula Meek, who conducted studies of patients experiencing the symptom of breathlessness due to chronic obstructive pulmonary disorder (COPD). Paula's research used a model by Howard and Elaine Leventhal (with Michael Diefenbach) that was an early iteration of the dual-process approach also used in TMT. She found that people who focused their attention on different aspects of the feeling of breathlessness then in turn had different interpretations of what that symptom meant for them, and that those interpretations changed their perception of the symptom's intensity. This example illustrates a back-and-forth between perceptions and thoughts, which is characteristic of Leventhal's model.

Leventhal's dual-process model, sometimes called the "common sense model" of self-regulatory behavior, is shown above. (It is also sometimes called the "illness perception model," although I will attempt to show below why that's not a great label). Leventhal starts from the observation that people have body sensations that we call "symptoms," and then they also have thoughts about those symptoms. Leventhal's first paper on this model notes that these concrete thoughts about what a symptom means can be particularly impactful in guiding behavior: "Indeed, the concrete, symptomatic aspect of the representation, the headache attributed to hypertension and the sick feeling arising after using insulin, may be more persuasive guides to action than the abstract knowledge a person has about a specific illness," (p. 144). The idea of an illness representation, then, is a cognitive or Narrative-mind construct. However, it's a lot more personal and embodied than high-level beliefs about what disease a person has, how that disease affects the body, or how things might be expected to play out over the long term. An illness perception is an even more specific thought about one's symptoms, which together add up to the internal model that someone creates as an illness representation.

Subsequent research identified many different sub-types of illness perceptions (note that I would describe all of these as "beliefs" or "narratives" -- statements about facts as the person sees them):

• Identity - what is the illness, and what are its possible effects on the body?
• Cause - why do some people have this illness, or what factors caused it to occur?
• Consequences - how will having this illness affect different areas of someone's life?
• Timeline - how long will the symptoms go on (duration), and how are they expected to change, improve, or worsen over time (progression)?
• Coherence - how does this illness make sense, or fit with the person's other experiences?
• Control - to what extent does the patient have control over what happens with the disease (personal control), or over the process of treating the disease (treatment control)?

These are all great questions that can help us to develop a rich description of a person's internal narrative about their illness. And it doesn't seem surprising that the answers to these questions, describing the patient's beliefs about their own personal situation, will be more predictive of the patient's behavior than their broad beliefs about the objective facts of the disease and its treatment for people in general. One strength of research based on Leventhal's model is that all of these sub-categories of beliefs can be assessed using a survey instrument, the Illness Perception Questionnaire (IPQ). The ease of data collection with the IPQ, and the number of potentially interesting subsets of illness perceptions that can be assessed, may explain why the vast majority of studies that claim to be "using Leventhal's dual-process model" are in fact simply studies that use the IPQ.

So what's the problem with the IPQ? There's nothing wrong with the tool per se, or with the idea of illness perceptions as far as they go. The problem is that so far, we have only talked about this top half of Leventhal's model: 

If you are a regular reader of this blog, you will recognize that this top half of the model only describes activities of the Narrative Mind. The bottom half, corresponding to the Intuitive Mind, is missing. In the top half of Leventhal's model we see illness representations (made up of illness perceptions), which are narratives or beliefs. Those supposedly lead to coping behaviors, and then the illness representations and coping responses go through an appraisal process before becoming part of the next step in the cycle and informing the person's subsequent symptom perceptions. But when the IPQ is the focus of research, the lower half of Leventhal's original model (emotional reactions, the Intuitive Mind) is missing.

TMT shows a similar but subtly different cycle, like this:

A key difference between these diagrams is that the intuitive response comes first in TMT. It leads directly to a behavior (no narrative "representation" step yet) and that behavior has consequences. Then comes the appraisal step, labeled as "narrative response" in the TMT cycle diagram. The first time the Narrative Mind has a chance to be involved is after the Intuitive Mind's initial response has already occurred. An Intuitive-Mind response might involve reactions that seem to imply a particular perception of what an illness is, what a symptom means, or how much a person can control the outcome. But crucially, that Intuitive-Mind response happens very fast and outside of conscious awareness. It is a blink-and-you'll-miss-it type of response, not a deliberative process with formal propositional beliefs, which is what is implied by the IPQ. 

Here's an analogy from mathematics: If given the equation x + 2 = 4, you can do a little mental algebra and solve for x. That's the Narrative Mind. But if I ask you "what is 2 plus 2," for most of us who have moved beyond the count-on-your-fingers stage, there's no logic involved. Two and two just is four -- you see it or you don't. If you are forced to justify how you know that two and two is four, you mostly have to go back to the counting approach. And even then, I can count to two twice or else I can count to four: If the person to whom I am giving this little performance stubbornly refuses to believe that those two procedures produce the same result, there is little that I can say to convince them.

So the timing of illness representations is one important difference between TMT and Leventhal's model (it's potentially something we could design a study to test, in order to find out which version of dual-process thinking is correct). The other important consideration is that Intuitive-mind processes are not the same as thoughts. Studies on Leventhal's model that only use the IPQ are missing half the picture: 

This is the Intuitive-Mind portion of Leventhal's model, and it was a great innovation when the model was first introduced. People are not just thinking machines, we also have feelings! And those feelings affect our behavior!  Leventhal's model was part of the 1990s "emotion revolution" in cognitive-behavioral therapy, when the idea of emotion-driven perceptions and reactions first found it way back into cognitive models of psychotherapy. The lower half of Leventhal's model is a much better match for TMT: First the stimulus generates an emotional reaction (emotions are one aspect of the Intuitive Mind), then the person has a behavioral response (coping procedure), and then the Narrative Mind gets involved to provide an appraisal of what just happened. The whole thing feeds back in to the next round of perception and reaction. You can see the same cycle at work in my TMT cycle diagram above.

One problem with how this part of Leventhal's model has been operationalized in research is that the word "representation" was used again in reference to the person's emotional reaction. Because of that, some researchers have included "emotional representation of symptoms" as just one more subscale on the Narrative-Mind-focused IPQ. For example, it might be represented by a propositional belief like "How much does your illness affect you emotionally? (e.g. does it make you angry, scared, upset or depressed?)" But any child can tell you that the feeling of being afraid is much different from the cognitive label "I am afraid. The latter is in fact a mindfulness intervention that creates distance between the embodied Intuitive-Mind experience of fear and the rational Narrative-Mind commentary on that experience, so that perhaps one can better cope with the fear. Sometimes CBT-trained therapists will attempt to reduce fear to a Narrative statement in answer to the question "what are you afraid of?" But fear is by its very nature non-rational. It simply is, it intrudes into a person's experience of the world. When we attempt to reduce Intuitive-level experiences to Narrative-mind statements, we lose sight of their power and disconnect them from people's behaviors.

In Leventhal's model, then, we have a model with two pathways -- Narrative Mind on top, Intuitive Mind on the bottom -- and some interactions between them. We have a cycle where one sequence of experience and appraisal feeds forward into future cycles. And in the example of Dr. Meek's study, we have some evidence that focusing attention on one or another aspect of experience can also change behavior. All of this bears a great deal of similarity to the original (non-cycle) version of the TMT diagram, which I suggest is just a different illustration of exactly the same theory:

Compare this to the diagram at the top of the blog: You have a stimulus on the left, which activates two distinct pathways -- Narrative on top, Intuitive on the bottom. There's a spot in the middle where the two pathways can interact (via attention). And there's a feedback look described in the text even if not shown in the diagram -- picture a second version of this diagram, just to the right of the first one, that takes into account everything that happened the first time around.

There's a perfectly good reason for these parallels. TMT is a descendant of Leventhal's model, which is cited (by way of Dr. Meek's work) in my original paper about this theory. (In fact, it's so clear to me that this was a source for TMT that I was surprised to find I had never written about Leventhal's work here before!). Where TMT differs from Leventhal's model is in asserting that emotions are not cognitions (so the bottom half of the diagram cannot be assessed using the same cognition-focused measures as the top half), and in insisting that the bottom half of the diagram is in fact the more important one, because it contains the only pathway that goes all the way through to behavior. The outcome of the top half is just narratives, the things we say about our experiences and behaviors. Narratives are an interesting result of illness perceptions, but they can only indirectly affect behavior, by becoming part of the background noise that affects our perception of stimuli the next time we encounter them.