Search the Site

Episode Transcript

Hey, Happy New Year! The episode you are about to hear, which comes from the Freakonomics Radio archives, is pretty much the perfect New Year’s episode. Because if you’re like most people, you make a New Year’s resolution and you’re pretty good about it for a week or three and then you think, “Eh, I think I’ll stop now.” But wouldn’t it be nice to find a way to keep going? This episode will tell you how. It’s called “When Willpower Isn’t Enough.” I hope it helps.

KATHERINE MILKMAN: So, I struggle at the end of a long day to get myself to the gym even though I know I should go. And at the end of a long day, I also struggle with the desire to watch my favorite TV shows instead of getting work done. And so I actually realized that those two temptations, those two struggles I face, could be combined to solve both problems.

That’s Katy Milkman.

MILKMAN: And I’m an assistant professor at the Wharton School, where I study behavioral economics and how people make choices.

STEPHEN DUBNER: OK, which might lead one to think that you are an economist by training, which would make one wrong, correct?

MILKMAN: That’s right. My background is actually that I have a Ph.D. in computer science and business, a joint Ph.D. those two fields. But I spend a lot of time hanging around economists and got really interested in this new field.

This “new field” that Milkman mentioned, behavioral economics, is something we talk about a lot on this program. It is essentially a marriage of psychology and economics. It is inhabited by people who wish to blend the economist’s view of incentives with the psychologist’s view that most people don’t respond to incentives as rationally as economic theory would predict. It was behavioral economics that years ago got me interested in economics in the first place. Behavioral economics may not seem as consequential as macroeconomics or labor economics — but, I would argue, it can offer tremendous leverage when properly applied. One well-placed nudge can go a very long way. It is also a field that appreciates simple, clever solutions. Like something that Katy Milkman’s been working on. She calls it “temptation bundling.”

MILKMAN: I invented that phrase, although I’m sure many other people had invented the same solution that I came up with and used that phrase to describe.

Here’s how it works in Milkman’s own case:

MILKMAN: What I realized is that if I only allowed myself to watch my favorite TV shows while exercising at the gym, then I’d stop wasting time at home on useless television, and I’d start craving trips to the gym at the end of a long day because I’d want to find out what happens next in my show. And not only that, I’d actually enjoy my workout and my show more, combined. I wouldn’t feel guilty watching TV, and time would fly while I was at the gym. So when I talk about temptation bundling, I mean combining a temptation — something like a TV show, a guilty pleasure, something that will pull you into engaging in a behavior — with something you know you should do but might struggle to do.

It’s a nice idea, isn’t it? As is often the case, ideas that seem new in academia have in fact been running wild in the real world for quite some time. As we learned when we solicited your examples of temptation bundling:


VOICE: My temptation bundle is to listen to Freakonomics podcast[s] while I’m running. I’m doing it right now.

VOICE: What I like to do is skip an afternoon of work and go [to] the movies after my annual pap smear.

VOICE: I really wish my temptation bundle was acceptable, but it would be drinking at work.  

*     *     *

DUBNER: So, I would love to know how you describe the unifying view of your research, which we’re going to talk about today. And so here’s, here’s the way I would describe it, maybe, is that there are a lot of activities that most of us consider “should” activities, and then there are a lot of ways that we avoid those “should” activities. And you try to identify the gap and look for ways to close the gap that are not too intrusive or costly or painful. That was my lame attempt to present your worldview. Let me hear it from you.

MILKMAN: I love your attempt to present my worldview. I think it’s much more cohesive. Actually one of the things that I think is always funny is, whenever I’m at a cocktail party, and I’m with my husband, and someone asks me what I do, I say, “Well, actually, he’ll describe it better than I will.” Because I get so focused on the little details that it’s hard to actually step back and see the big picture. But I think your description is beautiful. So I think, in short, I struggle a lot with willpower. And I find it difficult at the end of a long day to get to the gym. I find it difficult to stick to my diet. I find it difficult to stick to my goals more generally. And one of the things I’ve found curious is why and what I can do to solve those problems for myself and for others. And that’s where a lot of my research focuses.

DUBNER: OK, so let’s start with temptation bundling.

MILKMAN: So, what if you only let yourself get a pedicure while catching up on overdue e-mails for work. Or what if you only let yourself listen to your favorite CDs while catching up on household chores. Or only let yourself go to your very favorite restaurant, whose hamburgers you crave, while spending time with a difficult relative who you should see more of. Those would all be examples of temptation bundling.

DUBNER: Gotcha. OK, so the key is that the activity that you want to do is constrained, you can only do it when you’re doing the other activity. So that sounds to some degree like a commitment device, which we’ve talked about on the show before. And our examples were — so I’ll give you a couple of our strange examples of commitment devices. One was this guy who wanted to get healthy. He wanted to lose weight but he wanted to get healthy — he had a new kid. So he drew up a list of prohibitions, everything from alcohol to junk food to pornography, on and on and on, and if he violated them, then he would force himself to write a check to Oprah Winfrey, who he hated more than anything in the world for whatever reason. Steve Levitt had a diet idea also which was for people who wanted to lose weight:

STEVE LEVITT: One thing I know would work is just take a little can, like, say, a baby food jar, and fill it with vomit, okay? And wear it around your neck. And every time you decide that you’re hungry, just open the jar and take a little sniff. And I guarantee you, you will lose weight. Guaranteed.

DUBNER: So those are commitment devices of some sort. I’m curious to know whether you see temptation bundling as a commitment device, or it’s a, more of a first cousin somehow?

MILKMAN: I see temptation bundling as a new type of commitment device with some distinct features from standard commitment devices. So, a standard commitment device typically provides some consequence if you fail to engage in the intended behavior. And so this is a little different. What we’re doing here is basically combining two commitments with each other, and they sort of fit like puzzle pieces. So you’re using something that’s instantly gratifying to create a pull to provide the motivation you need to do something that’s unpleasurable at the moment of engagement. And then the other component that’s different is that you can actually have complementarities, which is an econ-speak term for peanut butter and jelly, two things that would go better together and are more enjoyable together than they would be separately. And so one of the neat things about, for instance, only allowing yourself to watch your favorite TV show while you’re at the gym is the fact that you might actually enjoy your workout more and you might enjoy the TV show more when you do them together, whereas a traditional commitment device just penalizes some behavior.

DUBNER: Great. Okay, so in theory, I’m sure everybody hearing this says, “Oh, man, Katy, that’s brilliant. That is such a good way to approach problem-solving.” So, that’s the theory at least, but now you need to find out if in practice it works. So, talk about what you did to find out if temptation bundling could actually improve outcomes.

MILKMAN: Absolutely. So, we ran an experiment actually at the gym, at the University of Pennsylvania, with a bunch of participants who told us that they wanted to exercise more. And we randomly assigned—

DUBNER: How did you find these subjects in the first place?

MILKMAN: We posted flyers all over the campus; we sent out email blasts to every listserv we could find; and the people who came told us they desperately wanted to exercise more, and they also had to have an iPod to be in our study, and they had to belong to the University of Pennsylvania gym so that they could enter and exit as they pleased.

DUBNER: Okay, good.

MILKMAN: So, we took these participants who wanted to exercise more and we randomly assigned them through a coin toss to one of several different groups. The first experimental group was a treatment group. And these participants came in and they had their iPod with them, and we told them that we were actually going to give them a new iPod in addition to the one that they already had that would be preloaded with four tempting audio novels of their choice. So we had a list of 82 books that had been pre-rated as extremely tempting and difficult to put down once you get engaged. So these are books like The Da Vinci Code and The Hunger Games. Participants choose the four that appeal most to them, load them on this loaned audio novel device, an iPod, and then they take them to the gym and they do a 30-minute workout while listening to the first audio novel.

DUBNER: I have to say that it’s so flattering, as a writer, that you are treating hearing a book as the treat that people get for having to work out. That’s so awesome.

MILKMAN: It is awesome. I love that it works, too. I do suspect that TV would have been even more tempting and enticing, but I think there are some novels that really grab you and, in fact, actually almost half of the participants in our study at the time when we ran it chose The Hunger Games. It was really popular right then. It was a fabulous book for this study, I think, because it’s so addictive once you get into it. It’s just impossible to put the book down because you desperately want to know what happens.

DUBNER: You only have access to the audio version of the book when you’re at the gym, correct?

MILKMAN: That’s right. So the participants in our study, they listen to the first 30 minutes and then they’re told — and this is while exercising — they’re told if they want to hear what happens next, they’ll have to come back to the gym where we’re going to hold this iPod we’ve loaned them in a locked, monitored locker that they can only check out when exercising. So that’s the first experimental group.


MILKMAN: The second experimental group has a very similar experience. They also get tempting audio novels. They also listen to the first 30 minutes. But this time those tempting audio novels are loaded onto their personal iPods and they also do the workout for 30 minutes while listening to the novel, but they are told, “Why don’t you try to only listen while exercising? But we won’t enforce it.” So we want to see if people can actually self-impose this rule successfully, which would be great, because if we can just educate people about this strategy and that’s all we need to do, and that can have a benefit, that means it’s a really easy job to roll out temptation bundling to the world.

DUBNER: Great, okay.

MILKMAN: Finally, we have our third group, which is actually a control group. And the control group, instead of receiving a tempting audio novel, they received an equally valued gift certificate to Barnes & Noble, and they also completed a thirty minute workout at the beginning of the study, but not while listening to a novel. They could do whatever they wanted. They had their iPods, so maybe they listened to music, maybe they listened to a novel, they could do whatever they wanted, but we didn’t control that. And we simply encouraged them to exercise more.

DUBNER: OK. So you’ve got the three groups. The ones who can only listen to the novel that you get them addicted to while at the gym.


DUBNER: The second group who, you give them the addictive material but they don’t have to come to the gym to listen to it. And then the third group who don’t even get the addictive material, unless they go out and buy it themselves with the money you gave them, right? Those are the three groups?

MILKMAN: Those are the three groups. I’d add to the middle group that gets the addictive material and doesn’t have to listen to it at the gym, that we strongly encourage them to only allow themselves to listen when they’re at the gym. So it’s framed as, “Try to only use this to lure yourself to the gym.” As opposed to, “Here’s just some material.” So they’re given the tool, in a sense. They’re just not given the restriction. They have to self-impose it.

DUBNER: Okay, so three groups. The data you collected for — people went for a minimum of nine weeks, I believe you said, yes?

MILKMAN: That’s right. That’s our follow-on period. So, basically what we’re doing is, we have this period at the end of that they’re paid for being in the study, dismissed, they fill out some survey measures, and we thank them. So our critical period is whether or not they exercise more during those 9 weeks while we’re imposing these different rules on them and giving them access to temptation bundling.


MILKMAN: And so what we see is that for the first seven weeks of our study, the full treatment group exercises significantly more than the control. And the group that’s given access to the audio novels and asked to just self-impose the rules, right in between the two. So that’s good news in a sense. They actually look a bit better than the control initially. And so that seems like maybe people can actually do this for themselves, although they’re not doing as well as the people who have the full, binding commitment that we’ve provided. And then we see a little bit of wear-off. And then over seven weeks this is working. Participants then on the seventh week go home — it’s Thanksgiving break, the gym is closed. When they come back, we look at the data and we see that the effect is completely eliminated. So when they come back, they seem to have forgotten about their novels, perhaps, and they are no longer motivated by temptation bundling. So on the one hand, this is actually kind of neat because it suggests that wanting to know what happens next in your novel may really be what’s driving our effects as we’d hoped, because forced separation is the one thing we know reduces cravings. But it also means we have more work to do to figure out how to re-engage people with temptation bundling after time away from the gym.

DUBNER: For the people in the first group, the treatment group, who can only hear the material at the gym, do they actually have to be working out at the gym to hear it? Or can they just come and get the iPod out of the locker and listen to the novel while sitting and drinking a milkshake or something?

MILKMAN: That’s a great question. When we did the study, we had no way of forcing them to workout while they were listening to the novel. We did actually have — there are security folks who sit right by the locker where these things are being held, and we asked them whether or not they saw people coming and sitting, hanging out. They said the people who were coming to the locker were coming to work out and they were in their workout clothes. But that’s anecdotal rather than empirical. What would be really cool is if we had Fitbits for instance on everyone and could actually track how much exercise they’re doing. And I’m doing new research right now, now that Fitbits are much more widely distributed than they were when I ran this study, where we are trying to do things like that, where we actually track physical activity, not just — in this case, all we could track is whether or not they enter and exit the gym.

DUBNER: Awesome. And I’m guessing there are so many places you can go with that because you could see pace and intensity and all different kinds of things that work out depending on material and inspiration, you know.

MILKMAN: Exactly, exactly, yeah. There are so many neat things you can see with Fitbit data.

There are also a lot of neat things that Milkman can envision doing with temptation bundling more generally.

MILKMAN: One thing you could do is imagine building this into an app, for instance, that let you create certain geo locations, like all the gyms you belong to or the only gym you belong to, where you can access tempting content on your iPod and then set restrictions. So you could say, “I can only access this content at those tempting locations.”

And then there’s what she calls “GymFlix.”

MILKMAN: Gymflix is one of my favorite suggested products. So, imagine that you took a company like Netflix and you called it Gymflix, and you let people set aside certain TV shows for gym-only access. My research suggests that a product like that might be very attractive to people.

And what Katy Milkman calls the “fresh-start effect.”

MILKMAN: So, the work I’ve done on the “fresh-start effect” was actually motivated by giving a talk out at Google a couple of years ago and telling them about some of the different tools I had done research on that can be helpful to getting people to do things like exercise and get flu shots and follow through on various other goals that they intend to follow through on, but sometimes fail to. And one of the HR folks at Google asked me this great question while I was out there. He said, “You know, we see that there are all these tools available and we’d love to roll them out and try to promote them to our employees to encourage more long-term focused behaviors, but do you have a sense of when it would be best to deploy these tools? What is the best time?” And we just hadn’t looked at that in my research group. So that’s actually what motivated the work I’ve done on the “fresh-start effect.”

DUBNER: Right. And what kind of tools were they trying to roll out?

MILKMAN: So, I had actually talked about the idea of temptation bundling, so that was one of the things I had presented. I had also presented some of my research showing that if you prompt people to form concrete plans about exactly when they intend to follow through on a given behavior — in this case it was getting a flu shot or getting a colonoscopy — even in the privacy of their own home, if they’re just prompted to think through privately the date and time when they intend to do that, that actually increases follow-through quite significantly and substantially. So those are the tools that I had just presented, and they said, “When do we roll these out?” And I sat back and I scratched my head and I thought, you know, well, we know about the New Year’s effect.

We do know about the New Year’s effect — that a lot of people make resolutions that they at least try to keep. Here are a few samples from people we spoke with:


VOICE: So, my resolution for New Year’s was to smile more.

VOICE: I make the same one every year: eat less, [BEEP] more.

VOICE: I mean, it was New Year’s last year, and I did start doing yoga, and now I do go three times a week, and it’s only to stop the inevitable decline of my person, because I’m 33, but I think it’s the best thing I’ve done for myself in a long time, and it was New Year’s.

VOICE: Is it a resolution if your wife decides that you can’t wear tube socks anymore? That was a resolution because they were thrown away!


MILKMAN: We create these resolutions and there’s a lot of social support for that kind of thing, so it’s not clear that that is anything besides a social norm or a social ritual, but I wondered if there might be something more to it. And in particular one of the things that struck me is that at the start of a new year, we feel like we have a clean slate. It’s the “fresh start effect,” in essence, that I feel at the beginning of a new year is driven by the fact that last year’s behind me, all of my past failures are from last year and I can think, “Those are not me. That’s old me. That’s not new me. New me isn’t going to make these mistakes.” And so I thought, if that’s true, if that’s part of what’s going on at the beginning of a new year, then there should be lots of other cycles that show that pattern as well. So we should see the same kinds of dissociation from our past failings and the same motivation to do better on this cycle at the beginning of a new week, at the beginning of a new month, following birthdays when we feel like we’re beginning a new cycle in our own lives, following holidays which may stand apart from other dates, and create the start of a new cycle for us personally as well.

DUBNER: So, talk about the evidence you have and where the evidence comes from —what you go looking for, where are you getting the data to do that?

MILKMAN: Yeah. The very first data set we explored, actually, was maybe very natural, given the idea originated at Google. The first thing we did is we downloaded searches for the term “diet” on Google. And diet, by the way, is the most popular New Year’s resolution. It’s the thing that we most struggle with, and that applies to me as well. So we downloaded daily search volume for this term over the course of eight years, which you can get from Google. And what we did is we looked to see whether or not there were any particular moments when it bounced up, and whether or not, in fact, it was consistent with our hypothesis, which was that people would search more for the term “diet” at the beginning of a new week, month, year, and following holidays. And we found, indeed, that that’s exactly the case in that data set. So we see bigger effects for dates that feel like fresh starts. And if we look at placebo terms, something like weather, news, that people often search for, we don’t see those patterns. So it’s something distinct happening when we look at the term diet. So that was our very first exploration and we got all excited about what we saw, and we said, “Well, let’s see. That’s just information search, information gathering. What about real behavior?” So we actually went to this data set on all of the visits to the gym by undergraduates at the University of Pennsylvania, which we had at our fingertips, and said, “Maybe we’ll see it in gym visits, too.” And when we looked at that data, what we saw is that people visit the gym, the very same undergraduate, in fact, we can look within student and see that the very same undergraduate is more likely to visit the gym at a start of a new week, month, year, following holidays and following birthdays — with the notable exception of 21st birthdays. And I’ll let you speculate as to why that might be different.

DUBNER: I’m guessing alcohol may be involved.

MILKMAN: I’m letting you speculate. I don’t have, I don’t have any data on exactly why. But, you know, my speculation may or may not have gone in the same direction.

DUBNER: Judging by the submissions you sent when we asked for your examples of temptation bundling, I’d say that, yes, alcohol may be involved.


VOICE: I like to drink Scotch while I fold laundry.

VOICE: If I don’t work out, I don’t drink beer. That’s my tradeoff.

VOICE: My name is Philip Nielson and I live in Sweden and my temptation bundle is actually cleaning my apartment before going out for drinks with my friends.


DUBNER: Okay, so I’m thinking of the different ways that one might look at this “fresh start effect.” So, so one would be that, you know, the beginning of a new year or maybe a new school term, or some new, you know, some demarcation of time that feels like your past is behind you and your future is ahead of you, and then you embark on this thing. And theoretically that thing you embark on is better for you. But then, talk about the degree to which people fall off. And the implication is that it’s bad to fall off, but I guess I’m thinking, well maybe this activity they were trying to change or add to wasn’t that important to them.

MILKMAN: That’s interesting. Yeah, I mean there’s a lot of fall-off with New Year’s resolutions, or with any goal that we’re pursuing with anything that requires will power.


VOICE: New Year’s, I always say, “I need to stop smoking cigarettes.” And I always revert back to smoking a couple days after.

VOICE: You know, I find that resolutions are tough, like I try to meditate more in the new year and it doesn’t go well.

VOICE: For the new year, yeah, you say you’re going to lose weight or do something like that — then by February, fell off.


MILKMAN: Some people say, “That means why even bother in the first place? Why even form New Year’s resolutions if you’re going to fall off?”

DUBNER: Right.

MILKMAN: I think there’s a few different responses to that. One is, you sort of can’t hit a home run if you don’t swing. So it’s never going to work if you don’t start, and I think that that suggests that we want fresh-start moments to motivate people, and hopefully eventually they’ll actually hit the home run, even if they fall off a bunch of times. So that’s one answer. Another answer is that sometimes, actually, the behavior just takes a tiny bit of effort. So, there are some things we can do that are just a one-time decision and they make a huge impact on the rest of our lives. For instance, if you just sign up for a 401(k) plan just once, then you’re going to be saving forever after. Or if you just get your flu shot, you only have to do that once a year and then it has benefits for the remainder of the year. A lot, a lot of different medical procedures have that characteristic, where if you could just get people to do it once every so often, even at low frequency, like a colonoscopy once every ten years could be enough to save many lives. So I think we can also think about the “fresh start effect” as having a lot of benefits for those kind of one-time behaviors with huge, huge downstream consequences that are positive.

DUBNER: Okay, so considering that you’ve identified this and found it, and found evidence for it in the data, what can you do about it? How can you use this knowledge to help people make more pro-social choices?

MILKMAN: Yeah, that’s a great question. Well, I think there’s a bunch of different implications. So, one thing we’ve tried is just reminding people that a given day is a fresh start. So, for instance, we have one experiment where we reminded people that a certain day was the first day of spring. And we experimentally compared people who we reminded a certain day was the first day of spring with another group that we didn’t. And the group that got that first day of spring reminder was more motivated to pursue their goals and receive a reminder about their goals specifically on the first day of spring, when it was labeled as such. And so you can think about just reframing a given day, reminding someone that it is an opportunity for a fresh start, is one intervention that might increase engagement in fresh start behaviors. You could also think about just asking people to do things that are good for them on fresh-start dates. So, you might try to roll out, for instance, a planning prompt campaign or offer people an opportunity to sign up for a commitment device or for a temptation bundling device on a fresh start date when we know their natural inclination and their motivation to do things like exercise and diet — and, by the way, we also found the “fresh start effect” with non-health related goals, so looking at things like financial goals and educational goals. So, whatever it is that they’re striving to do, if we provide them with the tools they need at a moment when they’re feeling fresh, they may be more likely to take advantage of those tools and start a good, new habit.

[MUSIC: Ruby Velle & The Soulphonics, “Looking For A Better Thing” (from It’s About Time)]

DUBNER: All right, so, Katy, let me play devil’s advocate for a minute. It’s more like jerk’s advocate, not so much devil’s advocate. But why should smart, motivated people like you have to work so hard to nudge and trick people into taking better care of themselves, for instance? Why not just let Darwinian forces do their thing and weed out the people who are too lazy to go to the gym or to get a vaccine?

MILKMAN: There are a lot of different answers. One could be that you simply want to see better outcomes for people who maybe didn’t realize what they needed to do or were too busy to figure out what they needed to do. So this is one way of helping people who maybe really would do this if they had the time and energy and education. Another answer would be that there’s huge social benefits, right? So even selfishly, we all pay the costs when other people fail to exert the willpower needed to take care of themselves. And so you can even think of a selfish reason for engaging in this kind of research and providing these kinds of nudges, which is that society is going to pay the cost when bad things happen to people who didn’t take care of themselves. But I like to think of it as a more altruistic motive.

DUBNER: Can I just say for the record that I don’t really agree with the jerk’s advocate argument, I just wanted to hear what you’d say.

MILKMAN: I know.

DUBNER: I’m on your side.

Maybe Katy Milkman’s research has inspired you to harness the fresh-start effect. But maybe you need one more nudge, a bit of further inspiration to follow. If that’s the case, we will leave you with some more of your examples that we gathered along the way.


VOICE: I’m not a particularly religious person, but during Yom Kippur, I think about some of the things that I haven’t done particularly well that year and some of the things I want to improve upon the next year.

VOICE: My grandmother, she died from smoking. So, I remember, I used to smoke cigarettes with my friends all the time in high school. As soon as she died, it was like it’s over. I was like: “Why?” Like, “That’s it. I’m not smoking cigarettes ever again.” But now I just smoke weed. But, yeah, that’s a better solution.

VOICE: Last August was my birthday and I turned 25, and something that was a long-term goal for me was to apply to grad school. So after I celebrated my birthday, I started with the application process for all the grad schools.

VOICE: Leaving a job is probably the perfect, sort of, fresh start moment.

VOICE: So, generally, in the morning I make better decisions than I do in the evening, let’s put it that way.

VOICE: I think that every day, honestly, you can start fresh. So, let’s say I went out to dinner the night before and I ate a lot of sweets and carbs, the next day, I wake up and instead of beating myself up about it, I try to eat a little bit better that day.

*     *     *

Freakonomics Radio is produced by WNYC Studios and Dubner Productions. This episode was produced by Greg Rosalsky, Suzie Lechtenberg, Christopher WerthMerritt Jacob and Caroline English, with help from Daniel Dzula, Paul Schneider, Anna Hiatt, David Herman and Rick Kwan. Our staff also includes: Arwa GunjaJay CowitAlison Hockenberry and Kasia Mychajlowycz. And why don’t you start your new year by finding us on Twitter and Facebook. And please subscribe to this podcast on iTunes, which just included Freakonomics Radio in its Best of 2015 roundup. Thanks.

Read full Transcript




  • Watch Katherine Milkman explain temptation bundling