The Behavior Change Podcast by Lirio explores the various ways humans can leverage behavioral science to personalize our messaging, engage our audience, and drive better behavior at scale.

Guest: Julie O’Brien (JO), Principal and Behavioral Scientist at the Center for Advanced Hindsight at Duke University

Host: Greg Stielstra (GS), Senior Director of Behavioral Science at Lirio

Summary: Julie O’Brien shares her findings with Lirio. Listen for a brief description of social proof and learn how behavioral science can be applied in real life to help people make healthier decisions—from vaccinations to dieting and beyond.


Greg Stielstra:    
Hello and welcome to The Behavior Change Podcast by Lirio, the program where we explore the marvels of behavioral science and ways of applying it to make a better world. I’m your host, Greg Stielstra. On today’s show, we’ll talk with Behavioral Scientist, Julie O’Brien, from the Center for Advanced Hindsight at Duke University, about how behavioral science can be applied to healthcare challenges.

But as always, we’ll begin with a Bias Brief. Lirio Bias Brief #96: Social Proof. If your friends jumped off a bridge, would you? Probably. In uncertain situations, we imitate other people’s actions as we try to do the right thing. It’s called social proof, and if you’ve ever joined the standing ovation for a mediocre performance, then you’ve experienced its power.

A powerful example comes from the 1962 episode of Candid Camera called “Face the Rear.” In the episode, an unsuspecting person boards an elevator, turns to face the doors, just as most people would, however the next three people to enter the elevator, all Candid Camera staff, faced the rear. Suddenly the first rider was uncertain of their choice, and little by little the camera captured them gradually turning to copy the example set by the others. Ironically, a sign instructing riders to face the rear may have had little influence on their behavior, but an example set by similar others was irresistible.

Social proof draws its power from uncertainty. The more unfamiliar the situation, the more we rely on others’ behavior to guide our own. Consider applying it to your marketing and engagement efforts when you’re trying to convince people to try something new or when the difference between alternatives is technical or complicated. Five rules apply to social proof. You can read them on the written version of the Bias Brief at

Our guest today is Julie O’Brien. She is Principal and Behavioral Scientist at the Center for Advanced Hindsight at Duke University. She has a PhD in social psychology and a background in product research. She loves behavioral science and believes it has the potential to solve lots of the world’s problems, especially when paired with technology and bold ideas.

Julie leads health and research at the Center for Advanced Hindsight, where she and her team focus on medical decision making, health management, daily activities like diet and exercise, and sexual misconduct. In her spare time, she attempts to use behavioral science to get her kids to eat more vegetables, brush their teeth, and go to bed.

So Julie, we can tackle the easy stuff later, but let’s start with the important question. How are things going with the kids? Do they now beg for Brussels sprouts, brush their teeth, and go to bed on time?

Julie O’Brien:      
Sadly and unfortunately, they do not and it is a constant struggle. And it turns out that little humans are not like big humans in all cases, but I do try to use some tricks like helping them remember their most recent instances of good behavior to kind of promote a consistency bias, which has been working okay recently.

But it’s much harder when it comes to your own kids than designing interventions for the larger population.

Do they ever suspect that mommy is using behavioral science on them again?

They do, and they’re four and seven, and their interpretation of behavioral science is learning how to manipulate people, so they know that I’m probably often trying to manipulate them.

Well, maybe that’s a phase they’ll grow out of.


Behavioral science is growing in popularity, but you’ve been at it for a while. You might say that you were doing behavioral science before behavioral science was cool. How did you first become interested?

Well, I guess like many people in college, I was trying to figure out what to do with my life and what things I actually really enjoyed. I ended up taking a social psychology class: Intro to Social Psychology. And it basically was one of these moments where you realize—this is the thing I’ve been looking for all my life. It was probably one of the first times I was exposed to randomized control trials and the idea of an experiment as a tool for being able to understand with confidence what is causing certain outcomes that you’re observing.

And I had also been at that point in my life trying to really understand certain behaviors that I would witness among my friends. So, the thing that was really the most interesting thing that got me hooked on this was what is, now I realize, one of these very complex behaviors that feels sort of irrational or doesn’t quite add up objectively.

I had all these friends who were really kind of academically talented and smart and competent, and yet they would read these unbelievably trashy magazines. I would look at the content of these magazines and think like, “This just can’t be interesting to you. There’s no way that learning what shade of lipstick is going to look better on your skin type is really that interesting when we’re having all of these other really deep and engaging conversations about real and important issues.”

So, that was one of the first observations that I had around how people kind of receive a lot of value from doing things that are normative or that are kind of behavioral scripts and how we can sort of balance different types of our identities. Sometimes our academic and social justice identity is salient. Sometimes our being a young girl or a young woman and wanting to be attractive and kind of fit into social norms and expectations can be salient.

And even if things feel like they’re irrational, there’s actually a lot of a lot of psychological processes and mechanisms that can explain this type of behavior. So, that was the original thing that got me interested. And then once I learned how to run an experiment, I was hooked and never considered doing anything else.

Behavioral science allowed you to retain your quirky friends despite their reading habit.

I did retain my quirky friends, and they tease me a lot about my inability to understand these things that are really simple and straightforward to them.

Now, you have applied your career in behavioral science to many things, but a special emphasis on the energy sector and also healthcare. And Lirio, as a company, focuses on those two verticals, and so we have special interest there. I’m wondering what you would say about why behavioral science is important to healthcare especially and how a behaviorally informed approach differs from maybe the way organizations have gone about healthcare engagement in the past.

I think in healthcare and energy, in both cases we have really, really major issues to try and solve. On the one hand, we basically just need many more perspectives trying to address these problems. It’s insufficient to take a standard approach of providing information because that’s what we’ve been doing for the last however many years and we are basically getting worse.

So, these are big problems that need multiple perspectives, multiple viewpoints, and more innovation to try and address them. I think the other thing that is really important and a really good fit for behavioral science is that the benefit and the payoff for both health and energy is in the future, but the temptation and the reward for the unhealthy or kind of bad behavior is right now.

And what we know is that people disproportionately value things that are happening right now relative to things in the future. Unfortunately we have not designed our world in a way that would help us actually make a good tradeoff between benefit in the future versus temptations right now. So, if you walk down the street, you are tempted with all kinds of unhealthy foods and a bar and maybe even smoking, but you’re not really tempted with a Paleo diet or …

I mean maybe in some places now there’s more temptations around Yoga or other kinds of activities that are healthy for people, but for the most part the things that we’re exposed to in our daily lives are often things that feel really, really good in the moment but are not so good for us in the long term. It’s the same with energy: it feels really great to be warm in the winter, but being warm in the winter means that we’re using more energy, and that has a lot of consequences.

For behavioral science, we are trying to solve problems where people pretty much understand what the actions are that they should take, so it’s not a question of not understanding that energy usage is associated with climate change or not understanding that unhealthy eating is associated with obesity. People get that. It’s actually a fairly basic concept, but it’s really hard to implement changes in the moment when you’re faced with things that are really appealing and can give you this immediate reward that you really, really want.

So, this trade off between the way it feels right now and how badly I want it right now and some future consequence is just really hard to make, and so as a result we kind of give into these momentary temptations. We all want to feel good, and then we end up doing all sorts of things that are really bad for our long-term health outcomes or our long-term environmental world.

Behavioral science is great because we can both understand what is happening psychologically, so we’re able to identify the mechanisms that are responsible for these irrational types of judgements that people are making, but also by understanding these mechanisms, we’re able to design better interventions that tap into the way people are actually thinking and behaving to help people make better choices in the moment.

We can use our rational self to arrange a context that is less tempting to our present self, our irrational self.

Yes. Ideally that is what we would do, yeah.

Good. Speaking of the difference between rational and irrational, you seem to have dipped into a little bit of dual-process theory there, so I wonder if you could tell our listeners about systems one and two and how they operate and what we can do as behavior-change agents as a result.

Yeah. Dual process is this idea that there are ways of thinking that are very conscious and deliberate, but then there’s also lots and lots of stuff that’s happening beneath the surface. I like to use the metaphor of an iceberg. I say the brain is like an iceberg. And with icebergs, you have a little bit of it that’s above the surface, and that’s like the part of your brain that is responsible for conscious, deliberate choice. In psychology, we talk about this in terms of working memory.

It’s the part of your brain that is active and that can engage with the content that you’re focusing on. It can weigh pros and cons. It can be very deliberate and make decisions and perform calculations, but that’s actually a small portion of your overall brain functioning. So much more of what’s happening is kind of beneath the surface, where you do things on autopilot or you have habits or things are happening maybe outside of your awareness.

So, you might have some clue that you’re doing something, but you’re not really fully paying attention to what you’re doing. And humans have evolved to have a brain that functions this way because it allows us to operate efficiently in our environments. If we were taking in every single piece of stimuli that we were ever exposed to and thinking about it deliberately, we would never be able to do anything because there’s just too much in our environments.

We would never get anything done, and we would never know what to respond to, what not to respond to, so humans evolved to have rules and heuristics that simplify our worlds. And this means we have things like stereotypes; we can make judgements about people very quickly. And sometimes our stereotypes are based on some kernel of truth. Often they’re not, but the way that they have evolved is to help us make decisions that are generally good enough.

And the problem is that now, in modern life we have so many temptations and also so many social issues where our stereotypes or our quick judgments lead us down the wrong path. We make the wrong judgments. We’re not really relying enough on that conscious, deliberate processing when we should. We become overloaded very quickly. The capacity of your working memory or that part of your brain that’s above the surface is about what it would take you to memorize a seven-digit number, so this is really, really small and it can become maxed our very quickly.

So, if you’re working on something with a lot of focus at work or you just have a lot of stress in your life or kids are screaming at you and you need to do things for them, your ability to focus on other things in your environment becomes minimized very, very quickly. And what that means is that you rely on your defaults or you do things on autopilot. And if your defaults or your autopilot isn’t already set up in the right way, you’re going to end up doing things that are unhealthy or suboptimal or you end up making bad choices, relying on too many heuristics.

So the busier we get, the more we rely on our subconscious self.

Yes, sadly.

Does this help to explain why we’re so bad at getting what we want? I mean most smokers say they want to quit. Most people who are overweight say they want to diet. And yet very few of those people manage to quit or lose any weight.

Yeah. I mean I think this is part of it. We know that, from the research on self-regulation and cognitive depletion, when your working memory becomes depleted, so if you are busy or stressed, your ability to make a decision defers to your automatic processing. And if you’ve lived your life … Let’s say you’re 40 years old and your habits have always been to smoke or always been to eat fried foods on Friday nights and now all of a sudden you have this new goal where you want to do something differently. When you are in these moments …

We actually have a paper coming out soon around this idea of breaking points, and we think about these moments in life where everything gets too much and our ability to really make a conscious, deliberate decision is sort of minimized. And at these points you really just go back to your default and you say, “Okay. I’m just going to reward myself with the thing that feels good, that I know, that I understand, that is immediate, and that is right now.” And you end up doing something that you probably would end up regretting later.

So, part of the problem is that we don’t do so well when we’re in these moments of breaking points or stress or cognitive resource depletion, but I think there’s some other things that are problems for us as well. We take on really lofty goals, and then we don’t actually know how to reach those goals. One of the behavioral tricks that we rely on in a lot of our studies is called implementation intentions. And what we know is that if you can get people to think through their plan for taking an action very concretely, you remove the guesswork for carrying out that action in the moment.

So, if I decide I’m going to start a new diet and I say my goal is to lose five pounds and I’m going to be reducing my calorie consumption by 500 calories a day, I need to plan very specifically what I’m going to do for my meals so that I don’t have any room for my conscious resources to become depleted, and then my default, which in that case would be my plan that I’ve made. And this has been shown in many, many different contexts. It’s a really effective way of helping people reach their goals because you’re removing that choice in the moment, which is likely to fail.

So, if the prescription for eating the elephant is one bite at a time, implementation intentions tell us where to start nibbling.

Yeah, exactly.

You’ve done some interesting work recently, publishing some things about tricking ourselves when it comes to overeating and, also, why it’s so hard to control diabetes. What are some of the most interesting behavioral healthcare findings you’ve encountered?

The one about overeating I’m really interested in. A lot of the work in social psychology, and also in behavioral science more broadly, has focused on how to help people avoid temptations in the moment. And this is obviously really, really important—but it’s extremely difficult, and it’s extremely difficult over the long term. So, what I’ve been thinking about lately is more, “How can we help people balance across days and across weeks, if we understand that there will be temptations that people give into?”

So, there’s some findings in psychology around something called the abstinence violation effect or some people call this the “What the hell?” effect. Let’s say I’m on a diet and I’ve decided I’m going to eat 500 calories fewer each day, but then I go to a party, and my friends have made really delicious cookies, and I can’t resist them. I have some cookies. But now I’ve failed at my rule that I had set for myself. I then feel like, “Well, what’s the point? My diet is sort of done with. I’ve ruined it all. I’m just going to keep eating cookies.”

And then what happens is you sort of … Maybe if you’re lucky you reset the next day, but you never really make up for those cookies that you ate on that day. And so if we assume that A: eating cookies can be really enjoyable and we should not take cookies away from people forever, but we do want people to compensate for it afterwards. Then, we need to figure out what are the things that are preventing people from making up for it after they have their cookies or after they eat in an unhealthy way.

So, the paper that we published recently was trying to describe this bias that makes it hard for people to compensate. And what we found was that there’s this distortion that happens. If I overeat by a small amount, I’m kind of okay at making sure that I make up for it later. And we know that from some naturalistic studies in the field that people do some natural compensations.

If I know I’m going to have a large meal later today, I might eat a little bit less for lunch, but what happens is we don’t compensate enough over longer periods of time and especially after larger meals. What our paper found was that when people are eating these larger meals compared to the smaller meals, they actually trick themselves into thinking that their body is going to deal with those calories without as much work.

So, if you looked at the amount of effort it would take to burn off one calorie of over-consumption, people end up thinking that if I consume a really large meal compared to a small meal, the calories in that large meal are going to disappear more quickly and more easily. And obviously that’s not how it works. And so, what ends up happening is people, at least from our research around what people think is how the body works, people think they don’t have to do as much work after these large meals. They think their body is going to do some of the work for them.

And we know that this is not because people can’t estimate the size of the large meals. We know that people are bad at estimating calories in meals, but people are even worse at compensating, so even if we account for this poor estimation pattern, where we don’t really estimate enough, or we don’t really know that we’ve eaten quite as much as we have, we still compensate even less than what we think we ate.

We just think that our bodies are really hardy. We give ourselves the benefit of the doubt. We tend to think that we are even more special than other people are. This is this better-than-average effect. “My body is going to work off calories even faster than everybody else’s body is, and so therefore, I don’t have to do as much work.” And what this means is that over time, as we have these larger meals, we just don’t compensate enough for it and then we end up gaining weight.

What I think is really interesting about this is we’ve identified this mechanism that is this behavioral bias. And now we know that people have this tendency to under-compensate, we can design interventions that make it easier to compensate a more appropriate amount. So, what we’re working on right now is doing some followup studies, where we look at this in the field and we give people simple rules and heuristics that make it easier to compensate the right amount after these bouts of over-consumption.

That is an interesting study that I was involved in. I think probably one of my other favorite studies in the health domain comes from a really good colleague and friend of mine, Zack Zenko. He did a study where he wanted to try and change the exercise experience to make people enjoy exercise more. And I think this is really great because everybody is told all the time that they need to exercise, and exercise is so important for all of these benefits: Alzheimer’s and weight and all these other things. But exercise is really, really miserable, especially if you’re not exercising regularly already.

But we also know that you can change the intensity over the course of an exercise routine and as the intensity changes, people’s enjoyment will also change. So, he wanted to understand, “Can we play around with the sequence of different intensities to get people to remember their experiences as more positive?” He did an experiment where he applied the peak and end rule. This is a principle that basically suggests that we form our memory of an event based on the worst part of it and the last part of it.

If we take this principle and we think about the way exercise is typically structured: You go to do exercise, you start with a warmup where you’re doing lower intensity, gradually you increase, you increase, you increase, until you’re at your max intensity, and then you’re done. If we apply the peak and end rule to this, we see that the maximum intensity and the end are both the worst, so you average those together and your memory of the experience was really, really miserable because you’re just thinking about the most intense part of your workout.

What he did was he tested the standard exercise routine of going from low intensity to high intensity against a different routine, where people went from high intensity to low intensity. And he made sure that the actual exertion across the workout session was exactly the same, so everybody is getting the exact same amount of exercise, but it’s a question of “Do you start with the worst and end with the best or do you start with the best and end with the worst?”

And what he found was that when you start with the worst, the high intensity, and then end with the best, the low intensity, you actually remember your exercise session as being much less unpleasant and more pleasant. And this is because what you remember is the worst of it and the last of it, so the peak and the end. And in the case where you start with the worst and end with the best, the average is much better overall than if you started with the best and ended with the worst.

I have a friend who is an athletic trainer, and I’m giving him that advice tomorrow.

Yes. Good.

You recently published a paper on vaccinations, and vaccinations are interesting because they’re proven effective at preventing nasty diseases, and so you’d think that people would get them, but quite often they don’t. And your paper explored some reasons why that might be and some possible solutions.

Can you talk about that?

Yeah, so vaccinations are a really interesting problem, and I think it’s a nice example of when people take a sort of traditional approach to trying to increase vaccination, they focus on trying to change attitudes and persuade people, and there’s sort of this assumption that many people are not getting vaccinated because they don’t like vaccines. And it’s true that there are a certain population that are kind of anti-vaxxers and they really distrust the pharmaceutical industry and they think that vaccines are really dangerous, but that’s actually a really, really small percentage of the population.

And for pretty much everyone else, there’s a lot of other explanations that can account for why people are not getting vaccinated. Probably one of the biggest explanations is context and friction. When we think about “What is it that’s explaining behaviors?” especially if we think about how people are not really thinking consciously, they’re not taking time to make a deliberate choice about what to do, then it means that they’re relying on cues in their environment.

And if their environment is set up in a way that it’s really hard to do the right thing, then people won’t do the right thing. And if you think about vaccination, when we look at childhood vaccination rates in the US, it’s actually pretty good. And that’s because every parent has to take their kid for a well-baby visit, or else they feel like they’re a bad parent. And when they go into the well-baby visit, it’s a really strong default you get a vaccine.

Of course, there are states where you can opt out, but the default is everybody getting vaccinated, and for the most part that works pretty well. But if you think about other vaccines like the flu shot or HPV vaccine, we don’t have those same defaults, and so, as a result, we have much lower rates of vaccination.

And what we know is that when we train doctors, for example, to have a presumptive recommendation where we say, “Everybody is getting vaccinated. It’s your time for the vaccine.” Of course, the person can still opt out, they can choose not to do it, but when we make it the default, people are just much more likely to do it.

I think one of the first reasons why people don’t get vaccinated is just friction, and this is something that we can all think of examples for. I, luckily, yesterday got my flu shot because they were offering flu shots in my office building, and so I went downstairs, and I got my flu shot. But nobody else in my family has done this yet because it’s just not easy. Last year I got my flu shot at Target, and it took me about an hour of waiting, and I had the kids, and it was really an unpleasant experience.

So, if it takes people an hour of waiting at the pharmacy to get their vaccine, that’s not an appealing process. There’s too many barriers; people are not going to follow through. So, one reason people don’t get vaccinated is really just about the context and the environment and friction. And then we have some things that are more about people’s beliefs about vaccines and the way that they think about vaccines in their social contexts.

When we think about the way we talk to people about vaccines … Like the flu shot is a great example. We say, “Get a flu shot. You’re going to be healthier this year. You’re going to avoid getting the flu,” or, “If you get it, you’re not going to be as sick.” But actually, the benefit of the flu shot is that you’re not going to spread it to somebody else who might have a higher chance of getting sick.

So, one of the things that we’re really interested in is: Can we take advantage of this social obligation that we have to other people around us and remind people that we do these things for other people, not just for ourselves? And when we do things for other people, then we have this sense of shame and we feel really badly if we didn’t do the right thing. Sometimes the context is the problem, but sometimes the social norms are the problem.

We don’t really think about vaccination as something that you do for other people, and if we thought about it as something that we did for other people, it would be much more visible because you would always want to know who’s vaccinated and who’s not vaccinated because the unvaccinated person has the potential of infecting everyone else.

And then I think you have this component that’s really about kind of the risks and the benefits.

We tend to do a really bad job of weighing risks versus benefits in general. Math is hard. It takes time. It takes that working memory. And we tend to really, really overestimate the chances that something bad is going to happen. And part of this is because when we go in and get our shot, the nurse says, “Let me know if you’re not feeling good and come back if you have any symptoms.”

So, we’re kind of priming people to think that there’s a good possibility that they’ll react badly to the vaccine. But actually, it’s extremely unlikely that anything bad is going to happen, so we overemphasize the risks. We think, “Oh, okay. Something bad could actually happen. Maybe I don’t want to take the risk.” And then when we think about the benefit, the benefit is avoiding diseases that we haven’t had for many, many years.

So, if the benefit is not getting measles, I can’t even imagine what measles would do, so it’s hard for me to imagine, “What is the chance that I would get measles?” if I don’t even know what it is. So, I’m overemphasizing the risks and I’m underemphasizing the benefits, which means then I’m not really making a very accurate cost/benefit analysis around whether I should get it.

Does omission bias play a role too?

Yeah, absolutely. The omission bias is this idea that we think that more harm is done when we take an action than when we don’t take an action. The really classic example of this is the trolley problem, and it’s sort of a thought experiment. Imagine that you are the conductor of a trolley and you’re going down a track and there’s five people, or some group of people, in front of you on the track.

You have the option of veering off onto another track where there’s only one person, so of course it’s much better to veer off onto the other track where you would only kill one person instead of five people. But, we tend to think that if we are the ones who take the action to move that trolley from the track with five people to the track with one person, we’re going to end up doing more harm, so we tend to just do nothing, and then we feel like we’re not morally responsible for any of the outcomes.

We end up actually doing more harm because then five people would be killed instead of one person, but because it wasn’t an action that we took or an explicit action that we took, we feel that it’s more morally acceptable. And of course, this is this nice example of where, really, we should all agree that it’s better to have one person harmed than five people harmed, but we misinterpret what defines harm and how our actions have a role.

With vaccines, if you imagine somebody who is maybe a bit hesitant and they think, “What if my child has a reaction that’s bad? If my child has a reaction that’s caused by the vaccine that I got for them, then it’s my fault, and that’s worse than if my child gets sick because I didn’t give them a vaccine.” So, it’s the same kind of thing where we think if we take an action that causes something bad to happen, then it’s our fault.

And I think there’s one other thing that has been kind of understudied within the realm of vaccines, which is also really important, and it’s basically just the affective or emotional experience that you have in the moment. A year ago, I took my daughter to get her flu shot, and the nurse kind of grabbed her arm, and it was a surprise, and she jabbed her in the arm, and then my daughter was hysterical for about an hour. And of course, I’m a researcher who is studying increasing vaccination.

I’m watching this, and I’m thinking, “This is such a terrible experience. I never want to go through this again. I can’t calm her down. Something must be wrong. This is just so, so awful.” And of course, I know rationally that vaccines are great, and I really want my kids to be protected, but it’s really hard to balance that against this hugely emotional and negative experience that I had in that moment that made me just never want to repeat that ever again.

We’ve already established that you routinely experiment with behavioral science on your children. I’m wondering whether you experiment on yourself. Do you ever apply behavioral science to yourself to trick yourself into getting what’s right for you?

I do. And I should confess that many of my research ideas come from my own behavioral failings.

Well, that makes perfect sense to me. Hey, thanks so much for being a part of the podcast. I really do appreciate it, and you have been a wealth of information. If someone wanted to contact you, learn more about the Center for Advanced Hindsight, what’s a good way to do that?

They can follow me on Twitter @jdpobrien, or you can check out our website or send me an email at [email protected].

Thanks again. I noticed that you worked for Opower. Maybe we can connect again in the future and talk about behavioral science and electricity.

Yeah. Sounds good.

You’ve been listening to The Behavior Change Podcast by Lirio. Lirio provides an email-based behavioral engagement solution that uses machine learning, persona-based messaging, and behavioral science to help organizations motivate the people they serve to achieve better outcomes. On the web at or follow us on Twitter @lirio_llc.

The thoughts and opinions expressed in this podcast are solely those of the person speaking.The opinions expressed are as of the date of this podcast and may change as subsequent conditions vary. There is no guarantee that any forecasts made will come to pass. Reliance upon information in this podcast is at the sole discretion of the listener.

© Lirio, LLC


Reach Out

Want to learn more about how Lirio’s behavioral engagement solution utilizes behavioral science and machine learning to help organizations motivate the people they serve to achieve better outcomes?

12 + 15 =