# Lesson 2: Social Behavioral Theory and Its Roots *Companion-podcast transcript • Sarah & Kiffer* *Office Hours episode to listen to after working through the lesson* --- **Sarah:** In 2014, Falko Sniehotta and two colleagues published an editorial in the journal Health Psychology Review. The title was Time to retire the theory of planned behaviour. **Kiffer:** And Icek Ajzen, who proposed the theory, replied that it was alive and well, and that many of those failed interventions were poorly designed. **Sarah:** So one side says the theory failed its tests, and the other side says those weren't fair tests. **Kiffer:** Which means you'd first need to know what a fair test of a theory looks like. **Sarah:** Welcome back to Office Hours. I'm Sarah. **Kiffer:** And I'm Kiffer. We're on Lesson two, Social Behavioral Theory and Its Roots, and fair tests are exactly where Sarah wants to begin. **Sarah:** Question one. Suppose a program reports that dozens of participants quit smoking after its workshop, and calls its theory confirmed. Most people know Popper wouldn't accept that. I think fewer people can say why. **Kiffer:** Popper's reason is that confirming cases are easy to find for almost any claim. Some people quit smoking every month whether or not they've been to a workshop, so a list of quitters fits the workshop's theory, and it also fits the theory that people quit when they're ready. A test tells you something when the theory could have failed it. **Sarah:** Is there a historical case where the confirming evidence was real and still pointed the wrong way? **Kiffer:** Cholera gives us a good one. In the 1849 epidemic in London, William Farr, a leading medical statistician of his day, found that death rates were lower in districts that sat higher above the river. He read that as support for miasma, the idea that bad air hanging over low ground spread the disease. The pattern was real, and it fit the miasma account. **Sarah:** It would also fit the water account, if the low-lying districts were drinking dirtier water. **Kiffer:** Right. When two accounts predict the same pattern, seeing that pattern can't separate them. What separated them was a place where they predicted different things. In 1854, John Snow studied streets where two water companies' pipes ran side by side, so a single house often had a different supply from the houses on either side. The households breathed the same air and sat at the same height, and the deaths followed the water. **Sarah:** And Farr came around eventually, didn't he? **Kiffer:** He did. His report on an outbreak in 1866 tied the deaths to one water supply, and by then he had accepted the water explanation. The evidence for miasma had been genuine, and it still couldn't tell the two accounts apart. **Sarah:** Let's try making a theory testable, then. I've got a made-up case. A recreation centre in Prince George writes in a proposal that older adults don't use its indoor walking track in winter because they don't know it's free, and that advertising that it's free will bring them in. **Kiffer:** Good. Start by pulling out the constructs and the proposition, the way Kerlinger's definition asks you to. **Sarah:** The behaviour is use of the track. The construct the proposal names is awareness that the track is free. Underneath that is a belief about cost, that walking there costs money. The proposition connects them. Older adults who believe the track costs money stay away, and when they learn it's free, more of them will use it. **Kiffer:** And what hypothesis would you derive from that proposition? **Sarah:** After the ads run, more older adults in the area will know the track is free, and winter visits by older adults will rise compared with the same months last year. **Kiffer:** Here's the question for everyone. What observation would show this theory is wrong? Be precise about what you'd have to measure. Take a few seconds. *(Pause)* **Sarah:** You'd need two measurements. If a survey shows that most older adults in the area now know the track is free, and visits haven't changed, the proposition has failed. Knowing about the free track didn't bring people in. **Kiffer:** And what would you conclude if visits haven't changed and awareness hasn't changed either? **Sarah:** Then the theory was never tested. The ads didn't reach people, so the program failed to deliver, and we've learned nothing about whether cost was the barrier. **Kiffer:** That difference between a theory failing and a program failing to deliver is one of the most useful things to carry from Popper into practice. Plenty of evaluations measure only the behaviour, and when nothing changes they can't say which of the two happened. Whenever it's possible, I'd measure the construct a program is meant to move as well as the behaviour. **Sarah:** And this is where people rescue their theories. Visits didn't go up, but it was a very cold winter. **Kiffer:** That rescue is fair if it adds something you can check. You can look up the temperatures and compare that winter with others. Popper accepted that revising the assumptions around a theory is a normal part of science. What he objected to were revisions whose only purpose was to escape a failed test, like saying a program worked in ways no one can measure. **Sarah:** Does this argument come up with the theories we'll meet later? **Kiffer:** It does, and in the open. In 2014, Falko Sniehotta and two colleagues published an editorial in the journal Health Psychology Review titled Time to retire the theory of planned behaviour. One of their arguments was that interventions built on the theory had often failed to change behaviour. Icek Ajzen, who proposed the theory, replied that it was alive and well and that many of those failed interventions were poorly designed. Underneath, the two sides were arguing about whether the failed tests had been fair tests. **Sarah:** Question two. Wittgenstein's language games and Foucault's idea of discourse both say a theory has meaning and validity within a context. Students sometimes hear that as saying every account is as good as every other. Is that what it means? **Kiffer:** That reading goes further than the idea does. Both thinkers are telling you that you can't judge an account until you know which rules you're judging it by. Within one set of rules, some accounts do better than others. Miasma and Snow's water account were judged by the same empirical rules, and miasma lost. What the idea asks of you is to notice which rules you're using, and to remember that the people a program serves may judge it by different ones. **Sarah:** The fire and water myth from the lesson makes the same point from the other direction. You can derive hypotheses from it, which tells you that what separates a myth from a scientific theory is the set of rules used to judge it. The other mix-up I see is between empirical and numerical. People hear that scientific theory rests on empirical data and assume interviews and personal stories fall outside it. **Kiffer:** In the empiricist tradition of Locke and Hume, empirical means knowledge that comes through the senses. Listening to someone describe their illness and recording what they say is observation, so interview data are empirical. Phenomenology, from Dilthey and Husserl, raises a different issue. Every observer perceives through their own biology, experience, culture and history, so the observer is always part of the data. Phenomenologists still observe, and they pay close attention to whose eyes are doing the observing. **Sarah:** Which brings us to the part where we disagree. Here's a made-up case. A seniors' centre in a small town in British Columbia runs a weekly peer-support program for older adults who live alone. Participants say they feel they belong somewhere again. The funder wants a standard loneliness scale, measured before and after, and a comparison group. **Kiffer:** Tell us where you stand first, Sarah, and then I'll give you mine. **Sarah:** I think the participants' own accounts should carry the most weight here. Belonging is something a person experiences, and a scale written somewhere else carries its authors' idea of what loneliness is. And in a town that size, a comparison group may not exist. If funders require one, programs like this may never get funded at all. **Kiffer:** I agree the accounts are evidence, and I'd gather them carefully. My worry is that participants saying they belong is exactly the kind of confirming case Popper warned about. People who didn't feel welcome probably stopped coming, so the people you interview are mostly the ones it worked for. I'd want at least one result that could come out badly for the program, such as following up with people who left, or tracking whether new members keep coming. A full comparison group may be out of reach, and I wouldn't insist on one. **Sarah:** That's still the funder's rules deciding what counts, though. The people in the program might care most about things no scale measures, like having someone notice when they've missed a week. **Kiffer:** That's a fair point, and it's the phenomenological objection in practice. I'd ask participants at the start what would tell them the program was working, and I'd measure that. Where I hold my ground is that some result, chosen in advance, has to be able to come out badly. Otherwise the evaluation can only ever confirm what the program hoped to find. **Sarah:** I'd still give the participants' accounts more weight than you would. But I can live with that as a rule. **Kiffer:** So here's the rule we share. Before a program starts, agree with the people it serves on what would show it's working and what would show it isn't. Then gather their own accounts, and at least one result that could show the program falling short. **Sarah:** Question three. How do you tell which school of psychology a program is carrying? I'd like to be the one who tries this. **Kiffer:** Here's a made-up program. A hospital medical unit in Kamloops wants staff to clean their hands before every patient contact. It has four parts. First, whenever an auditor sees someone clean their hands at the right moment, that person's name goes into a monthly draw for a gift card. Second, the unit posts its audit results each month, showing that most of the team cleans their hands correctly. Third, a short session lays out how often contaminated hands lead to infections and what those infections cost patients. Fourth, the staff design part of the program themselves, around the kind of care they want to give. **Sarah:** Let's give everyone a moment to sort those four parts into behaviourist, cognitive, humanistic and social psychology before I try. Take a few seconds. *(Pause)* **Sarah:** Okay. The posted audit results are social psychology, because they work through norms, what people believe their colleagues do. The infection session is cognitive. It gives people risks and benefits to weigh. The staff designing the program around the care they want to give is humanistic, agency and growth. And the draw is cognitive too, since each person weighs the chance of winning a gift card and decides it's worth the bother. **Kiffer:** Three of those are right. Check the draw with the question a behaviourist would ask. Does the design need to know what anyone is thinking? **Sarah:** No, it doesn't. It works through what follows the behaviour. Clean your hands at the right moment and you might be rewarded, and the reward comes unpredictably, which is a reinforcement schedule. That's Skinner's operant conditioning, so it's behaviourist. I was describing a person's reasoning, and the design never asks for any. **Kiffer:** That's the check. A cognitive theorist could describe the draw in terms of expected outcomes, which is why it helps to ask which assumption is doing the most work. This part of the design works entirely by arranging rewards around the behaviour, and it leaves the black box closed. **Sarah:** Can I add one from social psychology? Imagine a nurse on that unit, also made up, who'd never given hand hygiene much thought. She starts cleaning her hands because of the draw. A few months later, she tells a colleague she's just someone who's careful about this. Is that dissonance or self-perception? Take a few seconds. *(Pause)* **Kiffer:** It's self-perception. She had no earlier attitude that her behaviour contradicted, so there was no discomfort to resolve. She watched what she'd been doing and inferred an attitude from it. Dissonance would be a colleague who believed hand hygiene mattered, skipped it on busy shifts, and gradually started saying the evidence for it is overblown. There the behaviour contradicted an attitude, and the attitude moved to fit. So the deciding question is whether an attitude was there first. **Sarah:** Question four. If a construct predicts a behaviour, will changing the construct change the behaviour? **Kiffer:** Take intention, the construct Fishbein and Ajzen placed between attitudes and behaviour. People with stronger intentions are more likely to act, so it's tempting to conclude that raising intentions will raise behaviour. A correlation can't settle that. Part of the link may come from other things, such as past behaviour, which can shape both what people intend and what they do. Someone who already walks every day intends to keep walking, and does. **Sarah:** So you'd need an experiment that changes people's intentions and then watches what they do. **Kiffer:** Those experiments have been run, and one pooled analysis of them, by Thomas Webb and Paschal Sheeran in 2006, found that changing intentions on purpose did change behaviour, though behaviour moved less than intentions did. There's a second reason the answer can disappoint. A program that moves one construct can move another one at the same time, sometimes in the wrong direction. That's easiest to see when a program mixes schools of psychology. **Sarah:** Then that sounds like our harder example. What's the program? **Kiffer:** It's invented. A community centre in Abbotsford runs an after-school program to get youth aged eleven to fourteen more active. It has three parts. Each day a young person comes and joins in, they earn a stamp, and ten stamps earn a prize, like a hoodie. A short workshop has each of them list what being active would do for their sleep, mood and concentration, and what it would cost them. And each of them picks their own activity, like skateboarding, dance or basketball, and sets a goal for what they want to get better at. The stamps run for ten weeks, and the drop-in stays open after that. **Sarah:** So the stamps are behaviourist, the workshop is cognitive and the self-chosen goals are humanistic. Let me state the stamps as a theory. The construct is reinforcement, a reward that follows the behaviour, and the behaviour is coming and joining in. The proposition is that behaviour followed by a reward becomes more frequent. The hypothesis is that attendance rises once the stamps start, compared with the centre's sign-in records from before. It's shown wrong if the stamps are handed out as planned and attendance doesn't rise. **Kiffer:** The workshop works the same way. Its construct is each young person's own tally of benefits and costs. The proposition is that the more the benefits outweigh the costs in their own eyes, the more active they'll be. The hypothesis is that youth whose tally shifts toward benefits come more often afterward. It's shown wrong if the tallies shift and attendance doesn't. **Sarah:** And the goals. The construct is agency, the young person's own choice of activity and goal, resting on the humanistic assumption that people want to grow. The proposition is that people keep at things they've chosen for themselves. The hypothesis is that youth who describe their goal as their own keep coming to the drop-in after the ten weeks. It's shown wrong if they drop away once the ten weeks are over. **Kiffer:** Now here's the trap. One of those three parts could weaken what another one is trying to build. Which two, and why? Take a few seconds. *(Pause)* **Sarah:** The stamps and the goals. The stamps assume behaviour is controlled from outside, by what follows it. The goals assume the drive comes from inside, from wanting to get better at something you chose. If the prize becomes the reason to come, it could crowd out the reason the goals were meant to build. A kid who loved skateboarding before the program might start to see it as something you do for stamps. And the hand-hygiene unit had the same pairing, the draw and the part the staff designed. **Kiffer:** That's the conflict. Humanistic psychology grew partly as a reaction to behaviourism's mechanical picture of people. Your skateboarder also brings in a third school, social psychology. In 1973, Mark Lepper and his colleagues built on self-perception, the idea we used for the nurse. If you see yourself doing something for a reward, you may conclude that the reward is why you do it. Our nurse's monthly draw was too weak a reason to explain her behaviour to herself, so she put it down to being careful. In Lepper's study, preschool children who already liked drawing with markers and were promised a certificate for drawing later drew less during free play than children who got no reward or a surprise one. **Sarah:** Has that held up beyond preschoolers drawing with markers? **Kiffer:** In 1999, Edward Deci and two colleagues pooled one hundred twenty-eight studies and found that expected tangible rewards generally lowered people's later free-choice interest in activities that were interesting to begin with. That's usually called an undermining effect. Other researchers have disputed how far the effect reaches, so I'd treat it as a real risk that's worth testing in any program that uses prizes. **Sarah:** So what observation would show that the stamps undercut the goals? I'd watch what happens after the stamps end. If attendance drops, that's the sign. **Kiffer:** The simple behaviourist picture expects a drop too, so a drop on its own won't decide it. In that picture, the attendance the reward added fades, and attendance falls back toward where it started. An undermining effect predicts something stronger for youth who were already keen. They can end up coming less often than they did before the stamps began. So look at the youth who already came regularly, and compare their drop-in visits after the ten weeks with their visits before the stamps started. **Sarah:** If they've fallen below their own starting level, that's a sign the stamps weakened their own reasons for coming. If they're back where they were, or above it, there's no sign of that. I'd also ask them why they come, before and after, because the self-perception account predicts their reasons will shift toward the prize. **Kiffer:** And the cold-winter rescue from the walking track applies here. Seasons move attendance on their own, so check the centre's records from earlier years for how much attendance usually shifts between those weeks, and see whether the keen youth fell further than that. If I were designing this program, I'd keep the stamps for youth who weren't coming yet, and give the ones already coming feedback on progress toward the goals they chose. **Sarah:** So each part of a program can carry a sensible hypothesis on its own, and two parts can still undercut each other once they're combined. **Kiffer:** Let's pull it together with three things to take away. **Sarah:** First, a theory earns its place by surviving a test it could have failed. Decide in advance what result would count as evidence that it's wrong, and measure the construct a program is meant to move as well as the behaviour, so you can tell a failed theory from a program that never delivered. **Kiffer:** Second, every judgment about a theory rests on ground rules. Knowing which rules you're using lets you compare accounts fairly within them, and it helps you see when the people a program serves are judging it by different ones. **Sarah:** And third, every program carries an assumption about where behaviour comes from, whether or not its designers say so. Name the school behind each part, state the hypothesis each part implies, and check whether two parts rest on assumptions that conflict before you combine them. **Kiffer:** If you'd like more practice, add a fourth part to the after-school program, a team leaderboard posted on the wall. Say which school it draws on, state its hypothesis the way we did for the other three, and ask whether it conflicts with any of them. **Sarah:** Next time, it's Lesson three, Individual Health Behavior Theories, where the Health Belief Model, the Theory of Planned Behavior, the Transtheoretical Model and the Precaution Adoption Process Model put the cognitive assumption to work, and you can ask of each one what would show it to be wrong. **Kiffer:** Take care, everyone. **Sarah:** See you in Lesson three.