# Lesson 8: Communications Campaigns *Companion-podcast transcript • Sarah & Kiffer* *Office Hours episode to listen to after working through the lesson* --- **Sarah:** Think about the antidrug campaign from the lesson. Recall among youth was relatively high, so plenty of young people saw it and remembered it. **Kiffer:** And the evaluators reported that more exposure was followed by a stronger belief that other young people were using marijuana. **Sarah:** So the more teenagers saw the ads, the more they believed lots of other kids were using it? **Kiffer:** That was Robert Hornik's hypothesis, and he was lead author of the article that reported the campaign's effects on youth. **Sarah:** Welcome back to Office Hours. I'm Sarah. **Kiffer:** And I'm Kiffer. We're on Lesson eight, Communications Campaigns. Sarah, take us into the first question. **Sarah:** Question one. When a campaign doesn't change anything, most people's first guess is that the message was weak. The ad wasn't clever enough, or the tone was off. **Kiffer:** That's the instinct the exposure argument pushes on. Robert Hornik, the communication scholar at the University of Pennsylvania whose model types appear in this lesson, has argued since at least 2002 that a campaign's success is better predicted by the exposure it achieves than by the quality of its messages. At a National Academies workshop in 2014, he said that campaigns with promising messages often fail because they have no effective strategy for getting the exposure they need, and that most people won't go looking for a health message, so it has to reach them as part of their everyday routines. **Sarah:** So let's make that concrete with a made-up case. A health unit wants to reach young adults aged nineteen to twenty-four in a city that has thirty thousand of them. Plan A is a polished video on one platform. The budget puts it in front of twenty percent of the segment often enough to notice, and half of those people remember it. Plan B is a plainer message on three channels. It reaches seventy percent of the segment, and a quarter of those people remember it. Which plan leaves more young adults remembering the message? Take a few seconds. *(Pause)* **Kiffer:** Plan A reaches twenty percent of thirty thousand, which is six thousand people, and half of them remember it, so three thousand. Plan B reaches seventy percent of thirty thousand, which is twenty-one thousand people, and a quarter of them remember it, which comes to five thousand two hundred fifty. The plainer message comes out ahead by more than two thousand people. **Sarah:** Even though each person who saw Plan A was twice as likely to remember it. **Kiffer:** Right. The quality of the message multiplies whatever exposure you've bought, so it can't make up for an audience that never sees it. There's also a trap in the way exposure gets reported. Suppose the vendor for Plan A sends back a report saying the video got one hundred twenty thousand impressions. **Sarah:** That's four for every young adult in the city, so it sounds as if everybody saw it. **Kiffer:** It could mean that. An impression is one showing, so the same total could come from all thirty thousand people seeing it four times each, or from six thousand people seeing it twenty times each. Advertisers often summarize exposure with gross rating points, which are the percentage of the audience reached times the average number of times each person reached saw the ad. One hundred percent times four is four hundred, and twenty percent times twenty is also four hundred. **Sarah:** So the same impressions and the same gross rating points can describe two campaigns with completely different reach. You'd have to ask for reach and frequency separately. **Kiffer:** Yes, and you'd want recall from a survey of the segment, which is the first link in the evaluation chain from the lesson. Shares on social media raise the same issue. At that 2014 workshop, Hornik put it this way: "Virality comes and goes. It is not equal to useful effects." **Sarah:** Here's my pushback, though. The antidrug campaign had good exposure. Recall among youth was relatively high, and it still didn't move them. **Kiffer:** That's the right counterexample, and there's some history behind it. Hornik was lead author of the 2008 journal article that reported the campaign's effects on youth, and he had used the same routes of effect, individual, social and institutional, to design the evaluation. The evaluators reported that more exposure was followed by a stronger belief that other young people were using marijuana. Hornik's hypothesis was that the more teenagers saw the ads, the more they came to believe that lots of other kids were using it. **Sarah:** That's the norms reading of the boomerang effect that the lesson offers as a hypothesis. So the order of questions is, did they see it and remember it, and then, given that they did, did it fit them? **Kiffer:** That's the order I'd use. Check exposure first, because it's the most common reason campaigns fall short. Once you know the audience saw and remembered the message, the question becomes whether it fit them. **Sarah:** Question two. If formative research finds that a segment is low on one construct, should the message aim straight at that construct? **Kiffer:** Let's work through it with a group that's easy to overlook, people who work outdoors. CAREX Canada, a research team based in the Faculty of Health Sciences at SFU, estimates that about 1.6 million Canadians are exposed to the sun's ultraviolet radiation at work, and some of the largest groups are construction labourers and farmers. **Sarah:** And researchers with the Occupational Cancer Research Centre estimated that in 2011, about six percent of non-melanoma skin cancers in Canada, more than four thousand five hundred cases, were due to sun exposure on the job, mostly among men in agriculture and construction. So the data point to a segment. **Kiffer:** They do. Now imagine, and this part is made up, that a health authority in the Fraser Valley interviews farm and construction workers to plan a sun safety campaign. Here are four things people said. One: "I've worked outside for thirty years and never had a problem. Skin cancer's for people who lie on beaches." Two: "Our foreman brings it up at the morning safety talk when the UV index is high." Three: "I'd put sunscreen on at lunch if somebody left a bottle by the water cooler." Four: "I wouldn't know how much to put on or how often." Sarah, can you map those onto constructs? **Sarah:** Sure. The first is low perceived susceptibility, from the Health Belief Model. He sees skin cancer as a risk for other people. The second is a cue to action, the foreman's reminder on high UV days. The third is also a cue to action, the bottle by the water cooler reminding him at lunch. And the fourth sounds like low self-efficacy, or maybe behavioural capability from Social Cognitive Theory, because he doesn't know what to do. **Kiffer:** On the fourth, I'd lean toward behavioural capability, because he's describing knowledge he doesn't have, and self-efficacy is about confidence that he could act. Now check the third one by asking what the campaign would do about it if your label were right. **Sarah:** If it's a cue, the campaign sends a reminder at lunchtime. But a reminder won't help if there's no sunscreen on the site. He already says he'd use it. What's missing is the sunscreen itself. So it's a perceived barrier, access, and probably a real one too. I think the bottle in the sentence pulled me toward cue, because a bottle in plain sight sounds like a reminder. **Kiffer:** That's an easy slip, because the same object can do both jobs, and the test you just used sorts them out. A cue to action is a trigger, like a reminder or a symptom, that prompts someone to act now, and it works when their beliefs already favour acting and acting is possible. A barrier is a cost or obstacle between the decision and the action, and it's often a fact about the world. Getting the label right changes the plan, because a message can't put sunscreen on a work site. Either the program supplies it, or the communications objective becomes getting employers to supply it. **Sarah:** Which is why the communications plan separates the health objective from the communications objective. It keeps a campaign honest about what messages alone can do. **Kiffer:** Now the harder one, which is the first statement. Here's the question for everyone listening. This group shows low perceived susceptibility, and the lesson describes a brick wall when campaigns try to raise susceptibility among adolescents. Does the brick wall tell you what to do with experienced outdoor workers? Take a few seconds. *(Pause)* **Sarah:** I'd say it doesn't settle it. The brick wall is a claim about the developmental stage of adolescents. For these workers we'd need our own formative research, guided by the broader principle that theory selection is tied to population and circumstance. **Kiffer:** And listen to what he actually said. He tied skin cancer to lying on beaches, which is a specific belief about where the risk comes from, and he treats thirty years without a problem as evidence. A message that simply says you're at risk runs into his own experience. A message that takes on the specific belief, that it's beach holidays that matter, may have a better chance, because it gives him new information. **Sarah:** So you'd attach the risk to the job itself. The estimate we just heard, more than four thousand five hundred non-melanoma skin cancers in a year from sun exposure at work, is the information he's missing. **Kiffer:** That's a reasonable hypothesis to pretest. A second option works through the crew. The foreman is already a cue to action, and he's also a model in the Social Cognitive Theory sense. A campaign could make covering up and taking shade breaks something crews expect of each other. **Sarah:** And that second option moves us from an individual belief to a norm, which brings us to question three. How do you decide which of Hornik's model types a campaign belongs to? **Kiffer:** By the mechanism, meaning the route from the message to the behaviour. Individual models work through a person's awareness, knowledge and skills. Social diffusion models work through public norms that change behaviour across groups. Institutional diffusion works through elite opinion, then policy, then individuals. The channel doesn't decide the type, and neither does the label on the audience. **Sarah:** Let's test that on the sun safety campaign. Suppose it has three parts. Part one is radio spots and a card handed out with pay stubs, explaining how to use sunscreen and cover up. Part two is a push through crew leaders, unions and farm associations, with short videos of crews who take shade breaks together, also shared on social media. Part three is a coalition that gives local reporters stories of outdoor workers who've had skin cancer and asks city council to provide shade and sunscreen at all city work sites and to require them in city contracts. Which model type is each part? Take a few seconds. *(Pause)* **Kiffer:** Part one is an individual model. It works on knowledge and skills. Part two is social diffusion. It works by making shade breaks a normal thing for crews to do, and running some of it on social media doesn't change that. At most it adds a virtual dissemination piece within the social diffusion route. Part three is institutional diffusion, carried out through media advocacy. **Sarah:** And in part three the target is council. The residents who read the stories matter because they put pressure on council. It's also the route with the longest chain. If council agrees, every city crew and contractor gets shade and sunscreen, including workers who never saw a single message. If council says no, the coalition may have little to show for its budget. **Kiffer:** That's the trade-off drawn in the lesson's figure of the three routes. The longer the chain, the more people it can eventually reach, and the more links can break along the way. **Sarah:** Question four. What happens when a message that's right for one segment reaches another? **Kiffer:** This is the harder example, and it builds on the Surrey nicotine pouch case from the lesson. In August 2024, Health Canada limited the sale of nicotine pouches sold as quit aids to pharmacists, behind the counter, and banned advertising that could appeal to youth. Now the made-up part. A school district's survey finds that one in five students in grades nine to twelve has tried a nicotine pouch, and four in five never have. The planning group drafts a message for parents: one in five of our high school students has tried nicotine pouches, so learn what they look like and ask your teen. **Sarah:** For parents, that looks well designed. It raises perceived susceptibility, since my kid could be the one in five, it adds knowledge, and it gives a specific action, a conversation. **Kiffer:** Now read it as a fifteen-year-old who's never tried one. **Sarah:** It tells me a lot of people my age use them, probably more than I'd have guessed. That's the norms problem from the antidrug campaign. A message about how common a behaviour is can move perceived norms the wrong way. **Kiffer:** Right. What most people do is called a descriptive norm, and what they approve of is an injunctive norm. Robert Cialdini and colleagues have shown how descriptive messages can backfire. At Petrified Forest National Park in Arizona, a sign saying that many past visitors had removed petrified wood was followed by a theft rate of about eight percent, compared with under two percent for a sign that simply asked visitors not to remove any. And in a 2007 study, telling California households their neighbourhood's average energy use led those below the average to use more, while households that also got a smiling face signalling approval kept their use low. **Sarah:** So the two segments need opposite things from the same survey. Parents are moved by the one in five. Students are better served by the other side of the same result, that four in five students here have never tried one. **Kiffer:** Both statements are true, so neither segment is misled. Now the exposure plan, where the trap is. With made-up figures, the district's email newsletter reaches seventy percent of the six thousand parents, about three times over the month, and almost no students. A boosted video on the district's social media account reaches twenty-five percent of parents and sixty percent of the five thousand students, about four times each. The group wants the parent message on both channels. Assume the video reaches students who have and haven't tried pouches equally. On the video alone, how many parents would see the parent message, and how many students who've never tried a pouch? Take a few seconds. *(Pause)* **Sarah:** Twenty-five percent of six thousand parents is fifteen hundred. Sixty percent of five thousand students is three thousand, and four in five of them have never tried a pouch, so two thousand four hundred. The video would show the one-in-five message to more students who've never used a pouch than parents. **Kiffer:** That's the trap. The plan was judged by how many parents it reached, and nobody counted who else the channel reaches. And look at what the video adds. If the two channels reach parents independently, seventy percent of the video's fifteen hundred parents already get the newsletter, so the video adds only about four hundred fifty new parents. **Sarah:** And in gross rating points, the newsletter gives parents seventy times three, which is two hundred ten. The video gives students who've never tried a pouch sixty times four, which is two hundred forty. Divide gross rating points by one hundred to get the average number of showings per person. That's 2.4 showings for the average never-user, compared with 2.1 for the average parent from the newsletter. **Kiffer:** So the fix is in the channel plan. The one-in-five message stays in parent channels, the newsletter, parent advisory council meetings and family doctors. The video carries the four-in-five message, ideally with creators the students already follow declining pouches, as in the lesson's case. Parents who see the video still learn something true. And because a newsletter can be screenshotted and posted within minutes, I'd pretest the parent message with a few students too, and add the injunctive side if the survey supports it, for example that most students disapprove of pouch use. **Sarah:** Okay. Now the part where we disagree, which is fear appeals. I think they have more going for them than this lesson's treatment suggests. A 2015 meta-analysis by Tannenbaum and colleagues pooled one hundred twenty-seven articles with more than twenty-seven thousand participants. On average, fear appeals had a positive effect on attitudes, intentions and behaviours, and the authors reported that they found no circumstances in which fear appeals backfired. **Kiffer:** That's a serious body of evidence. My hesitation is about the conditions. The same meta-analysis found that fear appeals worked better when the message included efficacy statements, meaning statements that the recommended action works and that the audience can carry it out. They also worked better for one-time behaviours, like getting screened, than for behaviours that have to be repeated. Eating and physical activity are repeated behaviours, so the obesity campaign is aiming at the kind of behaviour where fear appeals did less well. **Sarah:** But Canada's picture warnings on cigarette packages are a fear appeal, and since 2012 they've covered three quarters of the front and back of the package. You wouldn't drop those. **Kiffer:** I wouldn't. They reach adults at every purchase, and since that same change they've pointed people to a toll-free quitline and a website, which is the efficacy part. My bigger worry is campaigns about weight. In 2012, Rebecca Puhl and colleagues at Yale showed about a thousand American adults a set of real obesity campaign messages. The ones already criticized for shaming people were rated most negatively, and people said they were least willing to act on them. The messages rated most favourably focused on specific healthy behaviours and didn't mention obesity at all. **Sarah:** That's a case for avoiding shame, though. A fear appeal can be about a disease without being about a group of people. **Kiffer:** In principle, yes. With weight, I find that line hard to hold in practice, because the feared outcome tends to get attached to people with higher body weight. And for older youth, the formative research in the lesson already tells us they read scare messages as adults trying to push them around. **Sarah:** And the same formative research suggests that younger youth may respond to them. So I'd keep fear available for the right segment, with a doable action attached. **Kiffer:** I'm more cautious than Sarah, but I think we can agree on a practical rule. Use a threat only when formative research with that segment supports it, pair it with a specific action people believe they can take, keep the threat on the behaviour or the disease and away from any group of people, and pretest it with the people it's meant for. **Sarah:** I'm more open to fear appeals than Kiffer is, but that's the rule I'd use too. And I'd add one test from the meta-analysis. If a campaign can't tell people about a specific action that works and that they're able to take, I'd leave the fear out. **Kiffer:** Let's pull it together with three things to take away. **Sarah:** First, check exposure before you judge the message, and ask for reach and frequency separately, because impressions and gross rating points can hide how many people a campaign actually reached. **Kiffer:** Second, label each formative research finding by what the campaign would have to do about it. A cue needs a well-timed prompt, a barrier needs something removed, and some barriers need a program or a policy that no message can supply. **Sarah:** And third, when segments need different messages, count where each channel's exposure lands, segment by segment. A message that serves one segment can backfire with another, especially when it tells people how common a risky behaviour is. **Kiffer:** If you'd like more practice, go back to the sun safety campaign from this episode. Take the four things the outdoor workers said, invent four more of your own, and map each one onto a construct, using the check of asking what the campaign would have to do about it. **Sarah:** Next time, it's Lesson nine, Global Health Applications, where these theories are put to work in health programs around the world. **Kiffer:** Take care, everyone. **Sarah:** See you in Lesson nine.