The Management Theory Toolbox

Episode 14: Mastering Extinction: How to Stop Reinforcing Bad Behavior at Work with Dr. Michael Domjan

• Season 2 • Episode 14

Share episode

Episode Highlights:

  • Introduction to the concept of extinction in behavioral psychology.
  • Explanation of operant conditioning and its application in the workplace.
  • Examples from "The Big Bang Theory" illustrating positive reinforcement and extinction.
  • Detailed discussion with Dr. Michael Domjan on the process and effects of extinction.
  • Real-world examples of applying extinction to undesirable workplace behaviors.
  • The importance of careful implementation to avoid frustration and aggression.

Key Topics:

  • Operant Conditioning: Modifying behavior through reinforcement.
  • Extinction: Reducing behaviors by removing positive reinforcers.
  • Paradoxical Reward Effects: How intermittent reinforcement can lead to greater persistence.
  • Practical Tips: Gradual phasing out of reinforcers and managing expectations.

Guest Expert: Dr. Michael Domjan

  • Professor of Psychology at the University of Texas at Austin.
  • Expert in conditioning and learning with over 50 years of teaching experience.
  • Author and co-author of numerous publications (over 200), including "The Essentials of Conditioning and Learning."

Related Resources:

  1. Dr. Michael Domjan's YouTube Channel: Learning and Behavior: Key Concepts by M. Domjan
  2. Book: The Essentials of Conditioning and Learning (5th Edition)
  3. Research Article: R. Hinkin and C. A. Schreisheim, "If You Don't Hear from Me You Know You Are Doing Fine: The Effects of Management Nonresponse to Employee Performance," Cornell Hotel and Restaurant Administration Quarterly 45 (2004): 362-373.

Connect with Dr. Michael Domjan:

Upcoming Episodes:

  • Stay tuned for our next episode where we dive into schedules of reinforcement and their applications in management.

Join the Conversation:

Subscribe and Review:

  • Don't forget to subscribe to The Management Theory Toolbox on your favorite podcast platform.
  • Leave us a review to help others discover the show and improve our content.
Speaker 1:

Just watch what people do when they put their money in a Coke machine and it doesn't work. Just watch the people's interaction with that machine. If you try this in the workplace, be very careful.

Speaker 2:

Welcome back to the Management Theory Toolbox, your top destination for uncovering the why behind management and business concepts. If you're an executive manager, consultant or business student and you want to dive into a scientifically rooted discussion of management theory, then you're in the right place. We're continuing to learn about well learning in organizations and we've been covering this topic called operant conditioning, which is just a fancy psychology term for modifying people's behavior using positive and negative reinforcement. A good example of this is in the Big Bang Theory show, where Sheldon Cooper tries to change Penny's behavior using positive reinforcement. Every time she does something he likes, for example, by not talking too loud or following Sheldon's house rules, sheldon offers her a piece of chocolate. He goes so far as to start rewarding her when she speaks in a lower-toned voice, because Sheldon finds her natural high pitches annoying. Eventually, Penny's boyfriend Leonard catches on.

Speaker 3:

I know what you're doing Really. Yes, you're using chocolates as positive reinforcement for what you're doing Really. Yes, you're using chocolates as positive reinforcement for what you consider correct behavior. Very good Chocolate. No, I don't want any chocolate. Sheldon, you can't train my girlfriend like a lab rat. Actually, it turns out I can Well, you shouldn't. There's just no pleasing you is there, leonard? You weren't happy with my previous approach to dealing with her, so I decided to employ operant conditioning techniques, building on the works of Thorndyke and BF Skinner. Yet by this time next week, I believe, I can have her jumping out of a pool bouncing a beach ball on her nose. No, this has to stop now. Are you saying that I am forbidden from applying a harmless, scientifically valid protocol that will make our lives better? Yes, you're forbidden.

Speaker 2:

We've so far covered the type of behavior modification that Sheldon employs in episodes 11, 12, and 13. But here's a question what happens when Sheldon stops giving Penny the chocolates? Will she immediately return to her previous annoying behaviors or will it be a gradual transition? Or perhaps some of those behaviors will never or rarely return, meaning that Sheldon's behavior modification program could have a lasting impact? It's a situation you might encounter at work. After learning about operant conditioning from this podcast, maybe you overzealously start implementing some positive and negative reinforcement regimes in your team. But what happens if you have to stop those reinforcements? Or what happens if you find out that something you're doing is reinforcing an undesirable behavior? Clearly you want to stop reinforcing that behavior.

Speaker 2:

In psychology and organizational behavior, the process in which a behavior is followed by the absence of a previously encountered positive consequence, thereby reducing the likelihood that the behavior will be repeated, is called extinction. Reducing the likelihood that the behavior will be repeated is called extinction. You can think of the positive reinforcer as a candle flame adding energy to and heating up some object. When the flame is extinguished, what happens? How quickly does the object cool down? Was some of the object melted or otherwise permanently changed by the application of that flame. This process of extinction is our topic for today, but this topic has several angles to explore and we're going to need some help from an expert in behavioral principles. We're fortunate enough to have with us today Dr Michael Daumian of University of Texas at Austin. Hi, dr Daumian, and welcome to the show. Thank you, glad to be here. Before we get started, go ahead and introduce yourself and tell us a bit about your background and your work.

Speaker 1:

I'm a professor of psychology at the University of Texas at Austin, where I've been teaching conditioning and learning for the past 50 years. If you can believe that, I supervised a learning lab for about 35 years studying basic processes of Pavlovlovian conditioning. I no longer have a lab, but I concentrate on scholarly writing and disseminating knowledge about conditioning and learning through books and a YouTube channel that has about 50 mini lectures about various topics in conditioning and has attracted viewers from 30 different countries around the world. I recently collaborated with Andrew Delamater in writing the fifth edition of the book the Essentials of Conditioning and Learning, which was published by the American Psychological Association in the summer of 2023.

Speaker 2:

Excellent, thank you, and 50 years of teaching at U of T, that's quite something. We're really happy to be able to get your expertise today. In today's episode, we're talking about the idea of extinction in the context of behavior conditioning. Could you tell us what extinction is?

Speaker 1:

So there are basically two types of conditioning procedures. One of these is called Pavlovian conditioning and was originated by Pavlov. A lot of us are familiar with that. In instrumental conditioning the emphasis is on the individual performing a response in order to obtain a desirable object. Instrumental conditioning, sometimes also called operant conditioning, was popularized by BF Skinner, and a lot of the procedures and issues that he discussed fall under the general rubric of instrumental conditioning.

Speaker 1:

Extinction is a learning procedure that you cannot introduce from the beginning. That is, you first have to have an earlier training phase, and the training can either involve Pavlovian conditioning or the training can involve instrumental conditioning. Once behavior has been trained with either of those techniques, then you can introduce extinction, which involves a modification of the procedure such that the source of motivation for the learning is no longer provided. In a conditioning procedure, be it Pavlovian conditioning or instrumental conditioning, you have what's called a reinforcer or unconditioned stimulus. So in Pavlov's experiments this was food. In Skinner's experiments with rats and pigeons he also provided a little bit of food. In an extinction procedure you no longer provide that bit of food. Originally the food was what motivated learning. In an extinction procedure you no longer provide the food and so you remove an important source of motivation for the original learning.

Speaker 2:

So here's an example I'm familiar with from my work as an engineering manager.

Speaker 2:

We'd commonly train associates and try to give them positive feedback on their work, to reinforce when they're following procedures or situations where they went really deep into a design. We want to reinforce that and provide that positive feedback. But maybe a decade or two down the road, when they're a lead engineer or a principal engineer, it's not really expected that we should have to keep giving them positive feedback on all these tiny details in order to continue to motivate them to do it. We should, of course, give them positive feedback on their good performance on the bigger tasks that are part of their more senior job description. But if we have to keep telling our senior engineer how great it is that they solved the problem in a thorough manner just to keep that behavior from lapsing, then something's probably wrong and maybe they're not ready for that more senior role. So, if I'm understanding this correctly, this is an example of extinction, where it's expected that the behavior should still continue even without the positive reinforcement that was previously given.

Speaker 1:

In the context of providing feedback and encouragement to get someone to perform correctly or perform at a high level. Extinction would be no longer providing that feedback, no longer providing that encouragement.

Speaker 2:

So how do we go about doing that properly? Is there a correct procedure for extinction?

Speaker 1:

Well, in Pavlovian conditioning we have a signal, or what's technically referred to as a conditioned stimulus, that during original training is paired with an unconditioned stimulus. Extinction involves presenting the conditioned stimulus by itself, without food, without the unconditioned stimulus.

Speaker 2:

Okay, so it's as simple as that. Simple as that. Yeah, Okay, just stop the positive reinforcement. But does that actually work? Are there any challenges with that?

Speaker 1:

Of course, when you say does it work? It depends on how you want things to work. The typical outcome of an extinction procedure is that the learned response declines. You get a reduction in conditioned responding once you no longer provide the unconditioned stimulus or reinforce the source of encouragement for the individual performing.

Speaker 2:

Now, in chapter 10 of your book, the Essentials of Conditioning and Learning, which I'll link to in the show notes for our listeners. You talk about some of the quote paradoxical reward effects in extinction. Tell us a bit about these paradoxical effects.

Speaker 1:

Paradoxical reward effects are very interesting and they fit into the general characterization is that you get more behavior in extinction or having provided less reinforcement or less encouragement during training. One of the first paradoxical reward effects was the so-called partial reinforcement, extinction effect. In this phenomenon two groups of individuals are compared. Extinction effect In this phenomenon two groups of individuals are compared. One group gets rewarded every time they do something and you would think that if you reward somebody every time they do something, they would be more likely to continue doing that when extinction is introduced. The second group, during original training, is given intermittent reinforcement. So they are given the reward or encouragement only some of the instances where they performed a behavior, and you would think that that would lead to weaker behavior and the individual would be more likely to abandon doing that when their extinction is introduced. But it doesn't work that way. It works just the opposite. Providing encouragement or reward less frequently during original training makes the individual continue to do that behavior for longer stretches of time in the face of extinction or in the face of no longer receiving the reward or encouragement. You see this clearly if you compare how long individuals continue to operate slot machines as opposed to how often they continue to operate a vending machine when extinction is introduced.

Speaker 1:

Extinction in a slot machine consists of the machine no longer working, so it never pays out If you create a malfunction, so it never pays out. Individuals who have been playing a slot machine are going to continue to operate the machine for long stretches of time even though the machine is no longer working properly. That's persistence of behavior following partial reinforcement. If you operate a vending machine, you put your money in. You expect to get the can of Coke every time you put your money in. So that's continuous reinforcement. If the machine now no longer functions, that means that you're on extinction. You don't persist on continuing to put money in the machine. The first time you fail to get your can of Coke, you probably swear at the machine, maybe kick it, try to shake it, put money in it again, hoping it's going to work the second time. If it doesn't work the first time, you get mad at it and you walk away. So extinction following continuous reinforcement results in a decrease in behavior very rapidly. So that's the paradoxical partial reinforcement, extinction.

Speaker 1:

Another paradoxical effect is that we expect that if we provide a larger reward for behavior, pay somebody a higher bonus for doing something that that is going to make them perform that activity more often in the face of not getting paid. Not getting paid is the extinction part. Well, the larger the reward that you provide during training. If you introduce extinction, you stop providing rewards. The individuals that received the larger reward quit a lot faster. If you want somebody to continue to perform the desired response in the face of no longer getting paid for it, you're better off having paid them a little bit and intermittently than having paid them a lot consistently.

Speaker 1:

It's paradoxical in that smaller rewards actually lead to greater persistence in behavior. You see this in the persistence of responding among artists. You know what is the probability that a poet is going to get their next poem published? It's next to zero. It's very difficult to get a poem published. But if they do get one published, that will encourage them to write more poetry and keep trying to get it published. So intermittent reward leads to quite strong persistence in the face of long strings of failure. Of course you see it in gambling.

Speaker 2:

Somebody can be losing bets most of the time, but the occasional time they win is going to trap them into continuing to bet in the face of no reward in the patterns of extinction are really related to the details of the schedule of reinforcement which, for our listeners, is actually going to be the topic of our next episode episode.

Speaker 1:

Absolutely, you're absolutely right. How quickly you stop responding in the face of consistent failure or consistent lack of pay on depends very much on the schedule of reinforcement that was in effect before the extinction was introduced.

Speaker 2:

Are there ways to perform the extinction procedure which takes into account those details of the schedule of reinforcement, for example? Maybe you know that just eliminating this positive reinforcer is very likely going to completely eliminate a desired behavior, but what if you were to gradually phase it out? Does the end result still reach zero?

Speaker 1:

The behavior doesn't always disappear completely. You get a decline in the behavior and depending on the circumstances you may not reach zero and even behavior that has declined sometimes will recur. So extinction is not a foolproof method of eliminating behavior. But you're absolutely correct that if you take a schedule of reinforcement or a program of reward or a program of payouts and you gradually thin them out so that you're introducing more trials where the payoff doesn't occur, that will tend to encourage the subject to continue to respond once you completely turn off any rewards or payouts. So you can engineer what the subject's reaction is going to be to an extinction procedure by appropriately adjusting the schedule of rewards and reinforcement before you introduce extinction.

Speaker 2:

So far we've been talking about the situation where a positive reinforcement is being used to produce a desired behavior, but for some reason or another we need to extinguish that reinforcer. So what happens to that desired behavior? But if we flip that around and look at situations where we're maybe reinforcing an undesirable behavior For example, an employee might be showing up late to work consistently and in doing so they're getting a reward by avoiding the morning rush hour. Maybe they get a sleep in later. Those are both rewarding consequences of this undesirable behavior. Now in this case I understand that probably it's difficult for a manager to use extinction because they don't really have control over those rewards. But in general, is the extinction process any different when we're talking about reducing the likelihood of undesirable behavior?

Speaker 1:

Well, the example that you presented is a complex one, and unfortunately, the real world is a lot more complex than the world of the laboratory, in which a lot of these principles have been developed. I mean, if you're obtaining a benefit for going to work late, a benefit being less traffic and not having to get up as early, those are benefits. Once you get to work, there may be penalties you missed a meeting, you didn't get a job done submitted on time, and maybe you're going to be docked some leave time or vacation time. While there are benefits to turning up late, there may be disadvantages. There are many consequences of turning up to work late, and it makes the analysis of the situation a lot more complex.

Speaker 1:

Obviously, this becomes a matter of deciding what's most important to you. Is it more important to sleep later or facing the displeasure of your boss? So it becomes a complex decision, and all of these factors could enter into it. As individuals make these decisions, they often consider the immediate consequences to be more salient than remote ones. A remote consequence is the next time raises are calculated, you may be at a disadvantage, but that won't happen for the next six months, and so you may risk being late a few times because the loss in a raise is too far away for you to worry about. So these decisions become complicated.

Speaker 2:

Yes, yes. And then from the manager's perspective, the manager is tasked with dealing out some of those penalties or punishments that you described. But in this example I don't think a manager could use extinction to remove positive reinforcers of showing up late. You can't, as a manager, control the traffic or the extra benefit of sleep or whatever. So let's look at another example where extinction might work. Suppose an employee has developed a habit of regularly visiting the manager's office to complain about their coworkers and most of the complaints are just trivial. The manager's kind of tired of this because it interferes with their job and wants to use extinction. So in this case it seems possible for the manager to have control over the stimulus or the reinforcer, specifically the attention the manager gives. The manager could remove that attention and stop giving this employee their attention when they're complaining about trivial things.

Speaker 1:

Attention is a huge reinforcer among people. Everybody loves attention. Everybody works hard to get attention. If the managers attending to these complaints actually reinforces the frequency of the complaints, that would be an excellent place to put extinction into effect. That is, don't provide attention for these things. You can do that by just telling them I don't want to talk about this, or you could say it's inappropriate. The simplest thing to say is I don't want to talk about this. Another way to do it is to create a suggestion box. You have a problem, put it in the suggestion box. That creates a delay and it also allows you to not ignore it. The best way to control inappropriate efforts to get attention is to not provide attention, and that is extinction, and that would be an appropriate and reasonably effective technique. Now, one of the things you have to be careful of is that you want to make sure that you do provide attention for appropriate kind of inquiries, so you don't want to put all behaviors on extinction.

Speaker 1:

This is a common issue in parenting. Actually, this attempt to get attention is a frequent event in parent-child interactions. Children are constantly trying to get attention and they get your attention by doing things like being too loud or throwing a baseball through the living room window. The mistake that parents make is that they don't provide attention when the child is doing constructive and useful things. It's like parents are busy, and this is also true in classrooms and daycare centers. Parents in charge are very busy having to look at a number of kids and so forth. So good behavior often gets ignored and bad behavior is the behavior that gets paid attention to. In order to solve that problem, you have to pay attention to good behavior. That way, the kid gets the attention that he or she desires, without resorting to extreme, unpleasant behaviors that are impossible to ignore.

Speaker 2:

What are some of your encouragements or cautions that you want to offer our listeners who might be eager to go try this concept of extinction in their workplace?

Speaker 1:

My recommendation, if you try this in the workplace, is to be very careful. And the reason I say that is that extinction is unpleasant and, depending on the circumstances of extinction, you get frustration, anger and even aggression. I mean, just watch what people do when they put their money in a Coke machine and it doesn't work. Just watch the people's interaction with that machine. All of a sudden they start swearing at it, they kick the thing, they shake the thing, they get very annoyed and then they stomp away. That will happen in the workplace. It'll happen in personal relationships.

Speaker 1:

If you are living with someone and you get in the habit of making coffee for them every morning, that is a perfectly predictable source of reward. One day you just don't feel like it and you don't make the coffee. If you do something regularly and then you introduce extinction, the individuals you interact with are going to be pretty upset with you. That's why I say if you're going to use extinction, be careful. Particularly if a behavior has been consistently reinforced, then the sudden introduction of extinction is going to be very unpleasant and it could result in unpleasant consequences.

Speaker 1:

So in a case of someone turning up to work late, if you introduce an extinction procedure, announce it ahead of time, you might discuss it, so it's not a surprise, everybody understands it. You might even engage your employees in designing that extinction procedure. Like we talked about the example of an employee that frequently comes to bother the manager about trivial issues. You could have a group meeting in which the manager explains that these kinds of interruptions make it difficult for the manager to do their work and we need to institute some different kind of procedures in the workplace and people working there could be recruited to make suggestions. You could have a sign that during this period the manager is unavailable for complaints or you could have a designated 15 minutes in the morning, 15 minutes in the afternoon. If a problem that's not urgent, drop in at those times. So you can introduce extinction. But be careful about how you do it so that you don't trigger the very intense and unpleasant consequences of extinction.

Speaker 2:

Well, that's all the time that we have. Thank you so much for joining us. This was such a fascinating subject and you're definitely an expert in this area, so we're grateful to learn from your expertise. Before we sign off, can you tell our listeners how they can find you and your work?

Speaker 1:

I'm a professor and so everything I do is out there in public. You can easily find research articles that I've written through Google, scholar or ResearchGate. I have an extensive faculty webpage at the University of Texas at Austin and, as I mentioned earlier, I've got this extensive YouTube channel on principles of learning, so you can search for that on YouTube, but also find my books on Amazon and other book retailers, so it's easy to find me. In fact, it would be harder for me at this point to try to hide. Anyway, appreciate your interest and thanks very much, travis, for having me on.

Speaker 2:

Yes, and thank you, my pleasure, that was really interesting. I think my biggest takeaway is the caution regarding using extinction. It's so easy to think, in the example of the employee who's making trivial complaints, that it'd be best to not give the attention to those and they'll stop. It's kind of like the scene in Parks and Recreation where a Pawnee citizen complains to Ron Swanson about the city's sprinkler system. To avoid engaging with the complaint, ron continuously swivels his chair away from the citizen, effectively ignoring her.

Speaker 3:

Excuse me. There's a sign at Rampset Park that says do not drink the sprinkler water. So I made some tea with it and now I have an infection. Sir, Sir, are you listening to me? Sir, Sir, I'm talking to you. Sir, Sir, are you aware that there is waste in your water system?

Speaker 2:

In that example, Ron's tactics elicited a negative response, but his use of extinction seems to more or less work and the complaints eventually go away. But wait, is this really extinction? Ron Swanson wasn't exactly known for attending to people's complaints, so he hasn't really conditioned the people around him to get attention as a reward. This, in fact, isn't exactly a clear-cut case of extinction, except in the general sense that people have been conditioned to usually get some kind of attention when they talk to someone directly, hence the irritation when he doesn't listen. But usually by the time we realize that we've been inadvertently reinforcing an undesirable behavior, it's probably too late by then. The employee is well-trained to seek the reinforcement and extinction becomes tricky.

Speaker 2:

This week, take some time to think about if there are any behaviors in your workplace that are undesirable. If so, it's probably a pretty safe bet that people aren't doing them just out of spite. Instead, there's probably some underlying reward that comes from the behavior. Is there a way to carefully extinguish that reward? Or maybe a safer option, is there another positive reinforcement that you can use to encourage a better behavior, over and above the rewards they might get from the undesirable one? Either way, be careful and be smart and, as always, revisit the theory and think deeply about how it applies to your individual situation and context. So with that, thank you for joining me on another episode of the Management Theory Toolbox. Stay tuned for our next episode, where we talk about schedules of reinforcement. In the meantime, keep learning, keep growing and keep adding to your management theory toolbox. Thank you.

People on this episode

Head Shot

Travis Mallett, ALM

Host
Head Shot

Michael P. Domjan, Ph.D.

Guest