# Lesson 6: Multilevel Theories and the Ecological Perspective *Companion-podcast transcript • Sarah & Kiffer* *Office Hours episode to listen to after working through the lesson* --- **Sarah:** Rodney's story in the lesson ends with two opposite possibilities, drug trafficking or a technical school program. **Kiffer:** And the theory fits both endings, whichever way Rodney's life goes. **Sarah:** That's what bothers me. If every outcome fits, what is the theory actually risking? **Kiffer:** It's a fair challenge, and I'd answer it in two parts. First, though, it helps to watch one of these models at work. **Sarah:** Welcome back to Office Hours. I'm Sarah. **Kiffer:** And I'm Kiffer. We're on Lesson six, Multilevel Theories and the Ecological Perspective, and we'll start with Bronfenbrenner's systems. I'd like Sarah to do the sorting for this one. **Sarah:** Okay. Give me a case to work with. **Kiffer:** Here's a made-up one. Jonah is nine. He lives about a kilometre from his school, and his parents drive him there every day. His dad worries about the four-lane road he'd have to cross. His mom's shift at the hospital starts at seven, so she drops him off on her way. The city's transportation department turned down a request for a crosswalk at that intersection. At a parent-teacher meeting, his teacher and his parents agreed to try a walking group with a few other families. And there's a broad cultural belief that kids aren't safe outside on their own. **Sarah:** Okay. Dad's worry is part of family life, so that's the microsystem. The meeting links two settings Jonah takes part in, his family and his school, so it's the mesosystem, even though Jonah wasn't in the room. Mom's hospital is a setting he never enters, but its shift schedule reaches him, so that's the exosystem. The belief that kids aren't safe outside is cultural, so that's the macrosystem. And the city's decision about the crosswalk is the government making a rule about roads, so I'd put that in the macrosystem too. **Kiffer:** You've placed most of them well. Let me ask about the last one. Is the crosswalk decision a cultural value or a law that shapes every setting in Jonah's life, or is it a decision made by one particular office? **Sarah:** It's one particular office. A transportation department in one city, with its own staff and budget, made a decision about one intersection. Jonah never sets foot in that office, but its decision reaches him every morning. So that's the exosystem, the same as his mom's workplace. I think I put it in the macrosystem because it sounded official. **Kiffer:** It's an easy slip to make, because anything with a government label can feel like it belongs in the macrosystem. The macrosystem is the broad culture, values and laws that all the other settings sit inside. A specific institution making a specific decision that reaches a child, whether it's an employer, a school board or a city department, usually belongs in the exosystem. **Sarah:** And the chronosystem? I didn't place anything there. **Kiffer:** You could add one. Say Jonah's parents walked to school on their own when they were his age. The change between their childhood and his is the chronosystem. So is Jonah getting older. A walking group that makes sense at nine might look very different when he's twelve and wants to go on his own. **Sarah:** Here's one for everyone listening. For Jonah, his mom's hospital is in the exosystem. Now suppose a program focuses on his mom, say a workplace change that lets parents shift their start times. For her, which of Bronfenbrenner's systems is the hospital in? Take a few seconds. *(Pause)* **Kiffer:** For her, it's the microsystem. She's there every shift, in direct contact with the people and the rules of that setting. The systems are defined relative to the developing person at the centre. A microsystem is a setting that contains the person, and an exosystem is one that doesn't contain them but still affects them. Change the person at the centre, and the same hospital moves to a different ring. **Sarah:** So before I place any factor, I should say whose rings I'm drawing. **Kiffer:** Yes, and that connects to the critique about placement from the lesson. The Theory of Triadic Influence puts social competence at a causal level above social skills, and it isn't obvious why, since you could argue that skills come first. Placement matters because it carries a causal claim. When you put a factor upstream, you're saying it works through the factors below it, and that tells a planner where along the chain an intervention could act. **Sarah:** Question two. The Theory of Triadic Influence says proximal causes predict most behaviour, and distal and ultimate causes explain it. If the proximal factors are the best predictors, why wouldn't every program aim straight at them? **Kiffer:** Prediction and explanation answer different questions. Proximal factors such as intentions, self-efficacy and normative beliefs tell you which students in a school are likely to vape next month. Distal and ultimate causes tell you why one school's rate is double another's. The epidemiologist Geoffrey Rose made a related point in a lecture he gave in Vancouver in 1984, published the next year as "Sick individuals and sick populations." He argued that the causes of individual cases can differ from the causes of a population's rate. **Sarah:** Let me turn that into a made-up example. Two schools run the same survey. Within each school, students who believe most of their friends vape are about three times as likely to vape themselves. School A's vaping rate is twice School B's. A board member says the gap must come from normative beliefs, so the answer is a norms correction program in School A. What's missing from that reasoning? Take a few seconds. *(Pause)* **Kiffer:** The survey finding is about who vapes within each school. On its own, it can't tell you why the two schools differ. For that, you'd look upstream, at how many shops near each school sell vaping products, what families and older siblings do, and how the price compares with what students have to spend. Those upstream conditions probably shape the normative beliefs as well. **Sarah:** And if more students in School A really do vape, then believing that most of your friends vape is partly accurate. **Kiffer:** That's the part people often miss. Norms correction works on the gap between what students think their peers do and what their peers actually do. Where the belief is close to accurate, there's less of a gap to correct, and the upstream supply keeps the rate where it is. **Sarah:** There's a second problem in the board member's reasoning, too. The survey shows that normative beliefs and vaping go together. Whether changing the belief would change the behaviour is a separate question, and it takes an intervention study to answer it. **Kiffer:** Right. A correlation between a construct and a behaviour is a good reason to test a program built on that construct, and the test is what tells you whether the program works. **Sarah:** Does the field actually act on all this? Almost everything I read endorses the ecological model. **Kiffer:** The published record suggests the field leans toward the individual end. A 2012 review by Shelley Golden and Jo Anne Earp coded one hundred fifty-seven intervention articles from twenty years of the journal Health Education and Behavior. Interventions were more likely to focus on individual and interpersonal characteristics than on institutional, community or policy factors. **Sarah:** Why do you think that is? **Kiffer:** I think it's mostly practical. A classroom program is something a school or a health unit can run, measure and report on within a funding cycle. Changing a retail environment or a provincial policy takes partners, time and advocacy, and it's harder to evaluate. I'd call the tilt understandable. The risk comes when the components that are easiest to run become the whole program. **Sarah:** Question three. In a PRECEDE diagnosis, how do you tell a phase three environmental risk factor from a phase four enabling factor, and does a factor's changeability depend on who's acting on it? I'd like a real problem for this one. **Kiffer:** Let's use extreme heat. In late June 2021, a heat dome settled over British Columbia. The BC Coroners Service later identified six hundred nineteen deaths as heat-related. Sixty-seven percent of the people who died were seventy or older, fifty-six percent lived alone, and ninety-eight percent died indoors. Most were in homes without adequate cooling, such as air conditioners or fans. **Sarah:** So the phase two epidemiological assessment more or less hands us a target group, older adults who live alone. **Kiffer:** It does. Now let's carry it forward with a made-up health unit in a made-up city in the southern Interior, call it Fernvale, planning for the next heat event. Sarah, would you take phase three? **Sarah:** Sure. Behavioural risk factors are what people do. Staying in an overheated apartment through the day and night, and not going somewhere cooler, would be behavioural. Environmental risk factors are conditions around people. A top-floor unit with no cooling that stays hot all night would be environmental. **Kiffer:** Good. Phase four goes back one more causal level. Predisposing factors in Fernvale might include the belief "I've been through hot summers before, I'll be fine," and not recognizing the early signs that heat is becoming dangerous. Enabling factors might include the cost of an air conditioner, building rules about installing one, a cooling centre that closes at five, and no ride to get there. A reinforcing factor might be a daughter who's relieved, and says so, when her father spends a hot afternoon at the library. **Sarah:** Here's the question for everyone. A Fernvale resident has no air conditioner. Is that a phase three environmental risk factor or a phase four enabling factor? Take a few seconds. *(Pause)* **Kiffer:** It can be both, at two different steps in the chain. The hot apartment is the phase three environmental risk factor, because the heat itself is what harms the person. The reasons the apartment has no cooling, such as the cost and the building's rules, are phase four enabling factors. They're the barriers that keep the person from cooling their home. **Sarah:** That's how the lesson treats retail access in the vaping case. It shows up in phase three as an environmental risk factor and again in phase four as an enabling factor. **Kiffer:** It's the same logic. Now for changeability. For the Fernvale health unit acting alone, air conditioning rates high on importance and low on changeability, because buying appliances for residents usually sits outside a health unit's budget. Partners can change that rating. In 2023, the province gave BC Hydro ten million dollars to provide about eight thousand free portable air conditioners over three years to people with low incomes who are medically vulnerable. **Sarah:** So the same factor moves from the innovate-and-evaluate quadrant toward program focus once the province and a utility are involved. **Kiffer:** That's the lesson's point about partners, with real numbers attached. Changeability depends on the factor and on who is acting on it. It's why the phase five administrative and policy assessment can change ratings you made earlier in the plan. **Sarah:** And there's something within the health unit's own reach. In 2022, the province set up a heat alert and response system with two levels, heat warnings and extreme heat emergencies. The health unit could organize volunteers to phone or visit people who live alone whenever a warning goes out. **Kiffer:** I'd rate the factor behind it, people living alone with nobody checking on them in the heat, fairly high on both. The coroners' data point straight to it, and a health unit can organize check-ins with community partners within a season. It seems a reasonable first battle to pick, with the air conditioning work running alongside it. **Sarah:** Question four, and this is the harder worked example. When you plan with the risk and protective factors model, how do you choose which factors to act on first? **Kiffer:** Here's a made-up case. Tamarack is a small mining town, and a youth agency there wants to reduce early school leaving. Using the Hawkins and Catalano model, its assessment finds four risk factors. Many students believe a diploma doesn't matter, because the mine has hired people without one. A close group of older teens who left school earn good money at the mine, and younger students look up to them. Many parents work two-week rotations at the mine, so there's less supervision and less help with homework. And many students don't know what they need to graduate. **Sarah:** Let me place those quickly. The belief about the diploma and the gap in knowledge both sit in the individual domain, as attitudes and knowledge. The older group is the peer domain. The rotations are the family domain, under family management. Did the assessment find anything protective? **Kiffer:** It found one thing. Most younger students play in the town's hockey program, and the coach requires passing marks to play. Now the agency's committee rates each risk factor from one to five. The diploma belief gets four for importance and three for changeability. The older peer group gets five for importance and one for changeability. The rotations get three and two. The knowledge gap gets one for importance, because most students who left already knew the requirements, and five for changeability, because an information session would close it. **Sarah:** And then someone on the committee adds the two ratings together to rank them. **Kiffer:** That's what happened. The diploma belief totals seven. The peer group and the knowledge gap both total six, and the rotations total five. The information session is the cheaper of the two sixes, so the committee puts it second on the list. Here's the question for everyone. The knowledge gap and the older peer group both total six. What does the importance and changeability grid tell you about those two factors that the total hides? Take a few seconds. *(Pause)* **Sarah:** They sit in opposite corners of the grid. The knowledge gap is changeable but unimportant, which the grid treats as a low priority, unless a quick visible win matters politically. The peer group is important but hard to change, which calls for an innovative component with evaluation built in. Adding the ratings turns two very different situations into the same number. **Kiffer:** And once cost breaks the tie, the most changeable factor that matters least ends up near the top. The agency would run a well-attended information session, report good process numbers, and see very little change in who leaves school, because the knowledge gap was barely linked to leaving in the first place. Keeping the two ratings separate is what stops that from happening. **Sarah:** So what would a better plan look like? **Kiffer:** It would start with the diploma belief, which is important and moderately changeable. It would treat the older peer group as an innovate-and-evaluate component. For example, the agency could recruit young mine apprentices who finished school to mentor younger students, and measure whether that changes who the younger students look up to. And it would build on the protective factor. **Sarah:** That's where I'd be careful, because it's the second trap. The committee's first idea for the protective factor was to start more teams and clubs to build belonging. But attachment shows up on both sides of this model. Attachment to positive, prosocial groups is the social bonding protective domain. Attachment to specific peer groups is part of the peer risk domain, and younger students are attached to the older group that left school, too. **Kiffer:** Right. Belonging on its own doesn't tell you which way it pushes. What makes the hockey program protective is who the bond is with and what they expect. The coach is a prosocial adult, and the passing-marks rule is the healthy beliefs and clear standards domain at work. A new club with no expectations about school could end up strengthening ties to the older group, if that's who it draws in. **Sarah:** So when the agency builds on the hockey program, it should copy the standard along with the team. And the mentoring idea works for the same reason. It gives younger students a different older group to be attached to, one that finished school. **Kiffer:** Yes. That fits the shift toward building protective factors from the lesson, with one condition attached. Before you count a bond as protective, check which group it ties someone to and what that group expects. **Sarah:** Okay. Now the part where we don't quite agree. My question is whether a multilevel theory can ever be shown to be wrong. Rodney's story ends with two opposite possibilities, drug trafficking or a technical school program, and the theory fits both. If every outcome fits, I'm not sure what the theory is risking. **Kiffer:** That's a fair challenge, and I'd answer it in two parts. Bronfenbrenner's model works mostly as a framework. It makes sure you look at every level before you decide anything, and a framework earns its place by showing you things you'd otherwise miss. The Theory of Triadic Influence is more specific, and it makes predictions that can fail. In the Washington State smoking study, it predicted that thrill-seeking, an ultimate-level trait, would matter less than parents' and friends' smoking. For first trying a cigarette, thrill-seeking mattered more. **Sarah:** And the overall verdict was still mixed but supportive. That's my worry. When some predictions fail and others hold, a theory this flexible can call the result partial support and carry on. **Kiffer:** I agree that's a real risk, and it's close to the complexity critique of multilevel theories. Where I'd push back is on setting the perspective aside because of it. The CDC review from the lesson, on macroeconomic factors and youth violence, found that family and school programs were often undercut by community conditions. A theory confined to one level wouldn't have told you to look for that. **Sarah:** That finding is useful. It also came from reviewing research and expert input after the fact. I'd still say the general theories are better at organizing what we already know than at predicting what will happen next. **Kiffer:** That's a defensible reading. I think the ecological perspective earns its place in assessment and planning, and I'd hold any specific multilevel claim to the same standard as any other theory. If a claim can't say in advance what result would show it was wrong, it's working as a framework, and we should describe it that way. **Sarah:** I'm more skeptical than Kiffer about how often these theories get tested, but we agree on the practical rule. Use the ecological model to look at every level during assessment. Then write down in advance which factor, at which level, you expect to change which behaviour, and by when, so the evaluation can show you were wrong. **Kiffer:** Let's pull it together with three things to take away. **Sarah:** First, a factor's place in Bronfenbrenner's systems depends on whose rings you're drawing. Use the participation test, and remember that a specific institution's decision usually sits in the exosystem, even when that institution is a government office. **Kiffer:** Second, proximal causes predict who will act, and distal and ultimate causes explain why rates differ between places. A program aimed only at the predictors leaves in place the upstream conditions that keep producing them. **Sarah:** And third, keep importance and changeability as two separate ratings, because a total can hide the difference between a low priority and a factor that needs an innovative component. And before you count a bond as protective, check which group it ties someone to. **Kiffer:** If you'd like more practice, work through the Fernvale heat example on your own. Add two factors of your own, place each one in phase three or phase four, and rate its importance and changeability for the health unit alone, then again with one partner added. **Sarah:** Next time, it's Lesson seven, Communities, Populations, and Settings, where we look at what makes a program community-based and how programs work in schools and worksites. **Kiffer:** Take care, everyone. **Sarah:** See you in Lesson seven.