Which Is More Powerful: Reward or Punishment?

DUCKWORTH: Everyone get into your seats, right now!

You’re listening to No Stupid Questions, the podcast that explores the weird and occasionally wonderful ways in which humans behave. Here are your hosts: Stephen Dubner and Angela Duckworth.

*      *      *

DUBNER: Angela, I have today the question, I think, to end all questions.

DUCKWORTH: Great. We’ll be finished with the whole podcast.

DUBNER: I think we’re done after this.

DUCKWORTH: Drops mic.

DUBNER: This is elemental, and you’re the person to answer it.

DUCKWORTH: Okay. No pressure.

DUBNER: And I can’t believe it took me a year and a half of making the show to ask you.

DUCKWORTH: I’m on tenterhooks.

DUBNER: What are tenterhooks anyway? The things you hang a side of beef?

DUCKWORTH: Yeah. There’s a carcass involved or something.

DUBNER: The metaphor works because, why?

DUCKWORTH: Um, I guess if you’re on them, you really want to get off?

DUBNER: But doesn’t the phrase mean — when you’re on tenterhooks — like, “I can’t wait to —”

DUCKWORTH: Yeah. Yeah, because I want to get off these tenterhooks. It’s very uncomfortable.

DUBNER: That feels like a very halfway metaphor. It doesn’t connote what I think it really wants to connote. How about “bated breath”?

DUCKWORTH: How do you even spell “bated”?

DUBNER: Not with a worm “baited,” but B-A-T-E-D.

DUCKWORTH: But what the hell does B-A-T-E-D mean?

DUBNER: Maybe it’s short of “abated”? Like, your breath used to be abated, but now it’s only bated.

DUCKWORTH: Anyway, I can’t wait, Stephen. What’s the question?

DUBNER: Here’s the question.

DUCKWORTH: My breath is bated.

DUBNER: Which is more powerful: positive reinforcement or negative reinforcement? And, of course, I want to know upsides and downsides of each.

DUCKWORTH: Okay. And you don’t mean “negative reinforcement” in the sense of removing something that’s bad in order to reward. You just mean punishment. Right? Do you want to ask about reward and punishment? Or do you want to ask about positive and negative reinforcement? Because they’re different.

DUBNER: I guess I need you to tell me the difference.

DUCKWORTH: Let’s start with a little glossary. So, reward is when you pair something good with a behavior in order to get it to happen again. So, you provide points for going to the gym, and people can redeem those points for Amazon credit.

DUBNER: That sounds suspiciously like an actual reward that you and your colleagues have been working on.

DUCKWORTH: It came to mind because we — as you know, Stephen — recently did a big study on that. And bottom line is: That works. It’s an effective incentive. But, you know, we do this all the time. You praise people for doing things, like, “Hey, you know, I just want to thank the person who brought in bagels today! Props to you.” Or you say things like, “Oh, I’m really proud of you.” You certainly have, I’m sure, given Fifi treats to reward certain behaviors.

DUBNER: We don’t say that word. We spell it. That’s really the only word she knows in her life, but it’s an important word. 

DUCKWORTH: So, we know what rewards are, intuitively, I think. Punishments, like, “Bad dog,” or if you yell at somebody for doing something wrong. That’s punishment. And I think the reason why these things seem to be the two tools in the armory is, like, “Well, you can reward the good, or you could punish the bad.” So, that’s what rewards and punishments are. And negative reinforcement has a very non-intuitive meaning. So, I apologize on behalf of all psychologists. Negative reinforcement, first of all, sounds like punishment, but negative reinforcement actually means that you are rewarding a behavior — you are promoting it, you are incentivizing it — by removing a punishment. So, you are taking away something bad, and thereby making something more likely to happen. Drugs are one example of this. If you have back surgery, or you get your knee replaced, and your doctor prescribes an analgesic, that drug is negatively reinforcing, because it is taking away the pain.

DUBNER: How would you describe the relationship between negative reinforcement and punishment?

DUCKWORTH: So, negative reinforcement would be getting somebody to do this behavior more by taking away pain. Punishment would be getting somebody to do it less by creating pain.

DUBNER: Although the punishment could be the withholding of something that they want, also. Correct?

DUCKWORTH: Yes. That’s true. 

DUBNER: Do you feel that the universe generally dispenses more reward or punishment?

DUCKWORTH: I would say two things. One is that we must be rewarded more than we are punished in some absolute sense, because we’re alive. And it depends on what you consider to be a reward, but it’s like: you breathe, you go to sleep, and you wake up, and you’re still there — meaning that rewards are outweighing punishments, I guess, in some cosmic, universal, objective sense. In terms of our psychology, we’re much more sensitive to bad things in our life than good things. I know we were recently talking about Roy Baumeister, the great social psychologist, and one of his seminal papers is called “Bad Is Stronger Than Good,” which is that we are always scanning the horizon for threats. We are much less likely to be reflexively grateful. We more ruminate on the bad things.

DUBNER: Now doesn’t that, in and of itself, suggest that we, as a species, tend to respond pretty strongly toward punishment or negativity, generally? At least as opposed to positivity?

DUCKWORTH: Yeah, that might lead you — if that’s all you knew about this thing called a human — to say, like, “Well, you should definitely try to change their behavior through punishment. Because look at them. They are just really good at noticing and being sensitive to bad things.”

DUBNER: Whereas if I pat them on the head and say, “Job well done,” etc, etc —

DUCKWORTH: They might not even notice it.

DUBNER: Or they might stop working as hard as they worked to get to the place where they earned the reward in the first place.

DUCKWORTH: Yeah. I mean, fear, and even loathing— Like, these are very strong emotions, and they do change our behavior immediately and noticeably. So, if you shout at someone to stop doing something, you tend to get an immediate response. And so, this is, I think, one of the reasons why there’s a lot of punishment in leadership and in parenting. And I say that as somebody— You know, when I raised my kids, I would yell at them all the time. “Don’t do that! Stop it!” And that is all, in a way, punishing the negative, right? I wasn’t spanking them, but I was still dispensing verbal punishment.

DUBNER: You weren’t saying, “Wouldn’t it be nice, Lucy, if you stopped hitting Amanda in the face?”

DUCKWORTH: Well, okay. So, the person whose work I read at the time — my kids were younger — his name is Alan Kazdin. He’s at Yale. And he is a clinical psychologist who has this extremely long career. He’s seen a lot of different forms of bad behavior, from mild to extremely dysfunctional. And one of his key recommendations is that parents so often react with punishment to things that their kids are doing that are bad. And that is, again, understandable. You know, if you do yell at your kid to stop doing something, you may not get the optimal reaction, but your kid might stop what they’re doing for a second. His recommendation, though, is to reward the positive. He calls it a “positive opposite.” Say, for example, a younger kid might leave all their toys somewhere.

What Alan Kazdin would say is: You have to actually pause and think to yourself, “What is that opposite?” If your kid is leaving their stuff everywhere, the opposite behavior is for them to be cleaning up, putting things away. And now, you have something to look for. Every time your kid is cleaning up, you go out of your way to praise it. You’re like, “Hey, Stephen, good for you for putting those blocks away. That is such a big help to Mommy!”

DUBNER: Thanks, Mom. I really appreciate that. What about — rather than just sitting back and waiting for the positive thing to happen — what about facilitating it? What about removing the obstacles that prevent it?

DUCKWORTH: Yeah, you can kind of set them up for doing that. And that goes by the incredibly unhelpful term “scaffolding.” Scaffolding is when you make a hard problem easier for your kid. It’s kind of like rigging the system. And I think that’s a good ide,a too. What’s interesting is that I knew this research when I was parenting Lucy and Amanda, and I didn’t do a very good job of doing any of this — like, naming the positive opposite, looking for ways to praise the positive opposite. Like, “Oh, Lucy, I’m just so happy when you and your sister are getting along. It makes me so happy when you’re sharing your toys.” I would just yell at them. Even when you know it, in the way that a graduate student in psychology might know the right thing to do and yet not do it, these reactive strategies — this, like, “Oh, don’t do that!” — they take no forethought. They just happen after the fact, and this positive-opposite praise, that takes a lot of energy, and a lot of proactive, strategic thinking.

DUBNER: So, you’re saying you need a little bit more planning or thought to create the reward ahead of time, versus the punishment, which just comes out of your mouth naturally. But it’s not just the response that’s very quick to make, it also produces a really fast outcome. Whereas a reward — the outcome may be much more nuanced, right? It might be smaller. In other words, you don’t stop bashing someone on the head — you start becoming a slightly more mature and kind person — whatever. Also, there’s no guarantee that that process will continue unabated. And thirdly, it might take a really long time. So, the more I hear you talk, I think, “Wow, reward and punishment are not just flip sides of the same coin. They’re totally different.”

So, if that is true, then what we really want to know is: What’s the best way to think about problem-solving, on balance? We know that negativity is powerful, we know that punishment happens intuitively and quickly, and it often works quickly — whereas all the stuff on the reward side, is a little bit more costly, a little bit more nuanced. That sounds like it’s going to be hard to wait for. Would you encourage people to nevertheless cut down on the punishment and bulk up on the reward, even though the benefits may not manifest so quickly?  

DUCKWORTH: This one has a really clear answer. You should definitely cut down on punishment and increase the praising of the positive. This is a classroom management trick that I used to use, too. You walk into a room, and there’s a moderate level of mayhem. You got to class maybe, like, three minutes after the bell rang for some reason, and the kids are already, like— They’re off. And what do you do? Do you yell at everyone? Like, “Everyone stop talking! I said, everyone get into your seats, right now!”

DUBNER: “Mrs. Duckworth is so mean.”

DUCKWORTH: Mrs. D, as they used to call me.

DUBNER: I’m surprised you let them even do that.

DUCKWORTH: I thought it was, like, hip, at the time. But here’s what a more skilled teacher says: “Stephen, thank you so much for having your homework out. Julie, I see you’ve got a pencil. Excellent! You’re ready to go.”

DUBNER: So, you identify the positive behavior and pay attention only to that.

DUCKWORTH: Yeah. And I will tell you, you don’t have to wait, like, a day, or a week, for that to take effect. It actually works relatively quickly. And you can imagine that those little moments of praise, the kids don’t forget it. They clearly are sort of moved by it. I know bad is stronger than good in some ways, but there is something about that approval that you get that— One could argue that that is stronger than being criticized.

Still to come on No Stupid Questions: Stephen and Angela discuss how positive reinforcement and rewards could be used to encourage nation-wide prosocial behavior.

DUCKWORTH: Hey, I’ll pay you 10 bucks if you vote.

*      *      *

Before we return to Stephen and Angela’s conversation about incentives, I’d like to share some of your thoughts on the topic. We told listeners to let us know about their experiences with reward and punishment.

@SomniNaut shared the approach that they use with their son. “Overall we use rewards and punishments to get quick (and often temporary) results in critical situations, or to mark an instance of exceptionally good or bad behavior. In normal situations, we just explain a lot and try to convince him to do what we want, and he mostly complies.”

@Caleb_Mezzy writes, “Punishment is setting a standard, not just for the kid, but for those punishing. If X happens, Y follows. If a ‘bad word’ was said, the kid must say sorry and so should the parent when the kid tells Daddy he said that ‘bad word.’ Hold everyone accountable.”

@___Jeroen___DB says that his mother used an incentive for behavior change that didn’t involve reward OR punishment. “When I was a kid, my mother always told me there would appear a black cross on my forehead when I was telling a lie. Once she dipped her finger in the ashes of the fireplace (without me knowing it) and made the cross sign on my forehead and then said to me to look in the mirror. I didn’t dare to tell a lie for more than a year until she told me what she really did.”

If you’d like your thoughts to appear on an upcoming show, make sure to follow our Twitter account, @NSQShow. Now, back to Stephen and Angela’s conversation about how incentives affect behavior change.

 *      *      *

DUBNER: Now, what about the fear — since we’re talking about children — the fear that overpraise can create a sense of entitlement, endowment. You know, a kid ties their shoes and you’re over the moon about how brilliant they are. They’re going to think they’re the next Einstein. In fact, they don’t know anything about math at all.

DUCKWORTH: I think that you want to avoid two things. One is inauthentic praise. It really does become empty. I have to say, my late step-mother-in-law — God rest her soul — she used to say of all of her grandchildren that they were perfect. And she used to say it all the time. Like, “Oh Amanda, you’re just perfect. Lucy, my perfect grandchild.” And I was like, “You know, at some point, the word ‘perfect’ starts to lose its currency.” I do think that you should be looking for opportunities to praise, but it should still be authentic. If you really think your kid’s piano playing sounds like cats mating, you shouldn’t be like, “Oh my gosh, this is incredible!” You can praise something else.

DUBNER: “I think it’s great that you can sit through your own horrific piano playing.”

DUCKWORTH: Yeah! Exactly. So, I do think that authenticity matters. And people are pretty good, at any age, at rooting out the inauthentic from the authentic compliments. And the other thing is, there is this idea of intermittent reinforcement from animal experiments mostly, which is the following phenomenon: You can praise somebody — or in animal experiments, reward them with a little pellet of chow or some sugar water. You can do it every time an animal, say, presses a lever, or does something that we’re trying to teach the animal to do. You can do it 100 percent of the time, or you can do it, say, 80 percent of the time. And you might expect that rewarding a behavior 100 percent of the time — that’s the strongest form of reinforcement. That’s the best kind of reward.

But there are experiments that show that, in some ways, the 80 percent reward, you know, most of the time it comes out, but one out of five times it doesn’t, and you don’t know when that would be. That can actually be more lasting. So, some would argue that with kids, you shouldn’t praise a hundred percent of the time, not only because it’s inauthentic, but also because you want to get a little bit of this uncertainty into the system. That’s debated.

DUBNER: Okay, so avoid inauthentic praise. I want to consider intermittent reward. What about you, personally? Not as “dispenser,” because we already know that you love to punish. You understand the power of rewarding, but it sounds like you’re naturally inclined toward punishment.

DUCKWORTH: I wasn’t a very good parent!

DUBNER: We’ve got to get your kids on the show and get to the bottom of this. But let me ask you this: Not you as the dispenser of reward or punishment, but you as the recipient. Tell me about that. How do you respond best?

DUCKWORTH: I sometimes imagine what the world would be like if everybody got as much praise as I get daily, if not hourly. I am often praised. I get emails like, “Oh, I love your podcast.” Or I get students who write me. And, as a recipient of all that praise, I think about what it would be like if all people got this much praise. I think people would be so happy. And I also think they’d be really productive. I don’t think you would make people into complacent under-performers. It really makes me more excited to do better.

DUBNER: Now, it’s probably hard to establish causality there, though. In other words, does the praise make you more productive, or are you praiseworthy because you already are productive, and you’re just keeping on keeping on?

DUCKWORTH: I know economists like to ask the question of reverse causality. “Is X really driving Y, or could it be that Y drives X?”

DUBNER: Maybe it’s the umbrellas that are making the rain come down.

DUCKWORTH: Exactly. But you know what? I think when it comes to human interaction, almost always X drives Y, and Y drives X in a kind of virtuous — or vicious — cycle. Meaning, I get praised, and then I have a little more confidence, and I feel like, “Whew, when I do well, it’s rewarded.” That makes me try harder, which earns me more praise, et cetera. And I think if you ask the question about so many psychological dynamics, it’s both. And that’s why these cycles get entrenched.

DUBNER: It’s an estuary of praise — is what you’re describing.

DUCKWORTH: Not a limited reservoir of praise.

DUBNER: What about governments? Let me ask you whether you feel, in that realm, whether reward or punishment is more effective, because governments, typically, are on the punishment side of things.

DUCKWORTH: Yeah. Think about the whole judicial system — the penal system. You know, nobody’s praising the positive opposite of burglary and other more serious crimes. They’re all kind of after-the-fact, negative consequences, which are, in theory, supposed to have some dissuasive effect.

DUBNER: Well, our mutual friend, Richard Thaler, who’s an economist at the University of Chicago, he wrote an op-ed in the Times. This is some years back. It was called “Making Good Citizenship Fun.” So, we’re not talking about crime and punishment here. We’re talking about compliance, things like that. And he gives a couple of examples of how it’s a good idea to use what he calls “positive reinforcement” — you would call reward — versus punishment.

DUCKWORTH: I would also call it positive reinforcement. I just wouldn’t call punishment “negative reinforcement.”

DUBNER: Okay, okay. So, he gives a bunch of examples of companies and governments around the world that have implemented fun tools to help people do the, quote, “right things.” This is all very much in keeping with the Thaler brand of what he calls “nudging.” Here’s an example, though. New Taipei City in Taiwan recently initiated a lottery as an inducement for dog owners to clean up after their pets. Somehow it always comes back to dog poop on this show.

DUCKWORTH: There’s a theme here — it’s a leitmotif.

DUBNER: “Owners who deposited dog waste into a special depository were made eligible for a lottery to win gold ingots. Thus, literally turning dog waste into gold. The top prize was worth about $2,000. The city reports that it halved” — cut in half — “the fecal pollution in its streets during the initiative.” So, that’s a nice, happy-ending story. There are also probably a whole lot of other cases where something like that is tried, and it doesn’t work. So, would you suggest that governments try to take a little bit more from the positive playbook than they typically do? And, if so, how would you suggest that they ensure that it does work?

DUCKWORTH: I love the idea that we should try some things. I think the Taiwanese dog-poop lottery is a completely plausible story. I do wonder about how that could get us into other problems.

DUBNER: So, you’re immediately thinking a lot of counterfeit dog poop, and you’re stuffing the ballot box with fake dog poop to get the gold ingots.

DUCKWORTH: Well, there’s moral hazard, but there’s also— You know, we can reward people, not just through dollars and ingots of gold, but, like, think about raising children. I don’t think many parents would say that a monetary system of financial incentives and fines is the best way to parent. I think we also want our children to want to earn our admiration, or our respect. So, if the government could find ways of dispensing incentives that aren’t just financial — and the reason is not just a dollars-and-cents budget costs thing. It’s also that there can be some problems with extrinsic incentives. Like, “Hey, if you do this, I’ll pay you five bucks. I’ll pay you, like, 10 bucks if you vote.” That also sends the message: that’s the reason why you would do those things. And I think there are other reasons that you would do those things. Like, it’s the right thing. It’s good for everyone. So, you have to be careful about how you reward things, but I think we should try. And it is interesting, now that you say it: almost everything governments are set up to do is about punishing the bad, and almost nothing is about rewarding the good.

DUBNER: The notion of using a positive reinforcement or reward to prevent someone from doing the bad thing — that’s really, to me, the hardest part of that matrix. And people have talked about this over the years. You know: give a bounty to potential criminals, rather than committing the crime, rather than dealing drugs, which hurts society overall, wouldn’t it be nice to incentivize people to do the pro-social thing? But not only is that a slippery slope morally, as you put it — if you pay someone 10 dollars to vote this year, do you expect 20 next year? And then what if you don’t have 30 the next year? 

DUCKWORTH: And also, haven’t you undermined the institution of democracy by saying that this is something that people would only do because we paid you?

DUBNER: But I have to say, when I look around the world, generally — all the world — the governments, the institutions, firms, families, and so on.

DUCKWORTH: When you look around the whole world. Yes, Stephen, go on.

DUBNER: It feels like, at least in most of the Western cultures that I’m more familiar with, the standard tool we reach for is the negative, is the punishment.

DUCKWORTH: Is the fine.

DUBNER: It’s the prohibition. I think back to something I learned a couple of years ago — actually, at the conference that you put together, the first Behavior Change for Good conference, where you had Danny Kahneman come and talk. He was interviewed by Max Bazerman from Harvard. And Danny talked about his primary influence, Kurt Lewin. And how Kurt Lewin’s realization — this goes back probably a hundred years now — is: If you want to get someone to do something, remove the barriers that prevent them from doing it. And similarly, I think if you flip that, if you want to get someone to not do something, well, our standard response is find a way to lock it dow,n or to threaten them. And I’m just curious whether all this learning that you and your peers have been doing for the past hundred years or so might provide us with some clever opportunity to think, even in a tiny way, about flipping that script a little bit, and using some kind of reward, or positive reinforcement, to prevent bad behaviors, antisocial behaviors, a little bit more than we’re doing currently.

DUCKWORTH: I remember when Danny said there was a smart way and there’s a foolish way to change behavior. That’s the same thing that we’re talking about here. I mean, there’s differences between rewards and punishments, and the particular principles he was talking about with Lewin. But what they have in common is that sometimes the smart way is really the non-obvious way. Like, what do you do when your kid is throwing their stuff everywhere? Tell them not to. That’s the obvious thing to do. But the smart thing is not obvious. You know, this is another thing that Alan Kazdin said: If you have a really, really disruptive kid — like, they are hardly doing anything that you can praise — and they’re certainly not doing the positive opposite. He was like, “Just praise anything. Literally just praise them when they’re just sitting there. You don’t even have to praise extraordinarily great behavior, just praise them when they’re not doing something terrible.” And so I do think this non-obvious route is very often the smart route. And it’s maybe for lack of wherewithal, or just that we are kind of frazzled, we do the more foolish, reactive, but less effective, thing. 

DUBNER: Well, Angela, I have to say, I think — parenthesis, (authentic praise) — you handled this question really well today. I also think — (intermittent praise) — you almost always do that. And therefore, I think, parenthetically, (as a friend), that we should probably do this again next week, and it’s probably time to go.

DUCKWORTH: Parentheses, (smile), end parentheses. Thank you, Stephen.

*      *      *

No Stupid Questions is produced by me, Rebecca Lee Douglas. And now here is a fact-check of today’s conversation.

In the first half of the show, Stephen and Angela conclude that tenterhooks are tools used to hang slabs of meat. This is incorrect. A tenter is a frame on which fabric is stretched to prevent shrinkage while it dries. Tenterhooks attach the material to the tenter. The phrase “to be on tenterhooks” originated from this method of drying, and was used to mean being in a state of suspense or anxiety.

Angela wonders about the definition of “bated” — as in “bated breath.” And Stephen says that it might just be a shortened version of “abated,” an adjective meaning to lessen in intensity or amount. This is correct. The phrase “bated breath” — meaning to hold one’s breath in suspense — was first used in Shakespeare’s play The Merchant of Venice. Today, “bated” is rarely used outside of that one specific phrase.

Later, Stephen references a 2012 New York Times op-ed by Nobel Prize-winning economist Richard Thaler, and he reads the following passage about dog owners in New Taipei City: “Owners who deposited dog waste into a special depository were made eligible for a lottery to win gold ingots. Thus, literally turning dog waste into gold.” Sorry, Professor Thaler, but I have to fact-check you here. They “figuratively” turned dog waste into gold. Not literally. A literal transformation would mean a bizarre form of alchemy that would certainly result in humans exuberantly collecting and investing their pets’ defecation, rather than depositing it into a special bin and waiting for the results of a lottery.

Finally, Stephen thinks that in most Western cultures, punishment is the standard tool. According to the World Prison Brief, the United States is the most punitive country in the world — with 629 prisoners per 100,000 people. But some non-Western countries have high incarceration rates as well. Rwanda comes in second on the list with 580 prisoners per 100,000 people and Turkmenistan is third with 576.

That’s it for the fact-check.

Coming up next week on No Stupid Questions: What happens when something you’ve believed for a long time turns out to be wrong?

DUBNER: And he said, “Yeah, that guy, he’s the “epit-toam” of stupidity.”

That’s next week on No Stupid Questions.

*      *      *

No Stupid Questions is produced by Stitcher and Renbud Radio and is a part of the Freakonomics Radio Network, which also includes Freakonomics Radio, People I (Mostly) Admire, and Freakonomics, M.D. This show was mixed by Eleanor Osborne. Our staff also includes Alison Craiglow, Greg Rippin, Morgan Levey, Zack Lapinski, Mary Diduch, Ryan Kelley, Jasmin Klinger, Emma Tyrell, Lyric Bowdich, and Jacob Clemente. Our theme song is “And She Was” by Talking Heads — special thanks to David Byrne and Warner Chappell Music. If you’d like to listen to the show ad-free, subscribe to Stitcher Premium. You can follow us on Twitter at NSQ_Show and on Facebook @NSQShow. If you have a question for a future episode, please email it to nsq@freakonomics.com. To learn more, visit Freakonomics.com/NSQ. Thanks for listening!

DUCKWORTH: We were real there.

DUBNER: We were real as [bleep]. That’s what the kids say.