Is Removing Rewards (Negative Punishment) for Unwanted Behavior Mean?

7 | Posted:

When I think back to when I was a student in college majoring in Biochemistry I remember my professors constantly stressing that school was more than book learning. It was about getting an education so that we could see the world more broadly and understand how the information we were learning fit into real life.
I took their advice and looked for connections between my classes in everything I did. When I exercised, the glycolysis and Kreb cycle pathways ran through my head, watched TV commercials psychology and classical conditioning came to mind, and whenever I waited in a slow line, I couldn’t help but think zero order kinetics.

Because of this training I’m sometimes surprised when bright students get tunnel vision. For instance, recently I was lecturing about the history of learning theory and how Skinner’s research was on positive reinforcement (reinforcing desired behavior) and negative punishment/extinction (removing the rewards for unwanted behavior).  I was about to provide examples when I was interrupted by a slightly off-topic question; “I read a journal article that says that removing the reward for unwanted behavior (a.k.a. negative punishment) causes increased cortisol and stress,” said the student. Now, in the context of the talk, this question comes off as, “But negative punishment is stressful, so why are we using it? Can’t we be 100% positive only?” This may not be necessarily what the student meant, perhaps she just didn’t realize the implications of context on her question. Because the statement was out of context, it puzzled me.

My first thought was, “Reality check.  Think about it. Can it be stressful when YOU don’t get what you want?”  Let me explain with an example. When I see that crying kid begging for candy in the store, the answer is clearly, Yes! That tantruming kid is not getting what he wants is clearly anxious and stressed and letting everyone know. But does that mean the parent should give in because they don’t want their kids to be stressed (or realistically because they don’t want the stress themselves)? Will giving the child everything he wants lead to a non-stressful life for the child? Probably not since ultimately the rest of the world won’t give them exactly what they want when they want it. Without the skills to cope with what happens in the rest of the world and without developing impulse control, imagine what can happen…  adult tantrums! A.K.A reality T.V.

Now let’s got back thinking about the kid at the grocery store eyeing the candy at the checkout counter. The reason that child even tries the tantrum is that he is often eventually successful at getting the candy reward. That anxious crying, whining, screaming behavior has actually been shaped systematically by the parent giving in after trying to wait it out longer and longer and for variable time periods. Now imagine a research paper on this topic of grocery store crying phenomenon. The conclusions of such a study on the phenomenon would have to conclude that this naughty behavior occurs because the kids are rewarded (a.k.a positive reinforcement). Hence it could conclude that positive reinforcement can increase stress and cortisol levels and now a college student who read that paper might need to interrupt a lecture to ask the absurd question, “Does that mean we shouldn't use positive reinforcement?”  My answer to that would be, “No. It means we should avoid reinforcing naughty behavior and focus on rewarding the behaviors we want!”

So, back to the original question about negative punishment and stress. My general answer is that, the worse the training protocol, the higher the level of the stress. For instance, say you gave a dog lots of opportunities and hence rewards for chasing squirrels and then suddenly you tied the dog outside with a squirrel running around him, I’m guessing cortisol levels would be pretty high because you are not allowing him the reward of actually having the access to chase the squirrel. In fact one study shows that in this specific type of prey situation, just standing around and letting the dog pull does increase cortisol quite a bit. The implication is that, not only will this method of waiting for the dog to calm down before giving it access to the squirrels take a long time to work, it will also increasing stress.

On the other hand, say you wanted to train a dog to sit and greet you when you come home instead of jumping all over. You start in a low distraction environment with treats ready in your hands. The dog comes over and maybe tries to jump once or twice. But because you stand completely stationary with your hands against your belly button, the dog sits within 5 seconds. Once he’s seated you deliver at a speed of 0.2 seconds so there’s no question in the dog’s mind about when you are planning to deliver the treat—the treat is just suddenly in his mouth. Furthermore, it’s all the way in his mouth. You didn’t delivery it in a way that makes him accidentally need to jump or stand or sniff the ground because the treat dropped. And then you give him a sequence of additional treats for remaining seated so that he is now happily sitting and has not had a chance to get up. Then you walk away and repeat this exercise a handful of times until the dog immediately sits and remains seated even when you’ve increased the interval between treats.

Now you practice a few times in a more high excitement situation by walking out of the house and then returning shortly thereafter. And once you’re sure you have the routine down and so does your dog because you’ve practiced enough times, you move to the highest excitement situations where you come home from work. Most likely, the negative punishment in each of these increasingly exciting situations may lead to a few fleeting signs of stress and frustration. On the other hand if you start by training the dog in the highest excitement situation, deliver treats in a confusing manner, time your rewards poorly, or reward at a low reinforcement rate, the dog’s frustration and confusion level are more likely to be high.


The Take-home Message

The take-home message is that you have to read research papers carefully to get a clear picture of the methods used to know what the appropriate conclusion is. Also, you need to look at everything you read in the context of real life. Regardless of what the particular research paper the student was referring to specifically concluded, my general conclusion would be, that if you used negative punishment without first training a more appropriate replacement behavior, or in a situation where it is difficult for the animal to determine a more appropriate behavior that can get rewarded—because you needed to add shaping steps or improve the timing, delivery speed, treat placement or other technique—then you will cause increased stress. The solution is to make sure you are rewarding a more appropriate behavior and have good skill, because like with the crying kid, you can’t give dogs everything they want, whenever they want it and or you will end up rewarding unwanted behaviors and training impulsivity.

What it boils down to is that if you are unskilled, even when you’re attempting to use good techniques you can cause more stress than you should. However, if you improve your plans and skill at reinforcing desired behaviors and removing rewards for unwanted behaviors you develop a better understanding of your pet and your pet will have a better understanding of you. In other words better technique in training decreases stress and leads to a better bond between you and your pet!

Tags: , , , , ,

7 responses to “Is Removing Rewards (Negative Punishment) for Unwanted Behavior Mean?

  1. THANK YOU, for writing this and putting into words what I wanted to so eloquently. I have recently read blog entries that were *probably* fueled by these studies. A particular blogger suggested totally eliminating extinction because it causes stress in the dogs. I couldn’t agree more with what you have written on this!

  2. Camo Cross Dog Training: You’re welcome. And even if people wanted to eliminate extinctionl they couldn’t Because extinction isn’t a method of training. The definition of extinction is “extinction refers to the gradual weakening of a conditioned response that results in the behavior decreasing or disappearing.” Negative punishment; however is a process that leads to extinction! That’s another blog article topic!

  3. I’m slightly confused about something; isn’t the removal of a reward called negative reinforcement? Since you are removing (negative) the reinforcer (reward)? And isn’t extinction of unwanted behavior the desired outcome of using positive/negative reinforcement/punishment and not the same thing as negative punishment? Isn’t negative punishment the removal of a punishment since you are removing (negative) the punishment?

    Thank you

  4. Dr. Yin,

    I know that there is a lot of debate on this topic. I have seen some of the debates myself. I appreciate your stance on it, and I wholeheartedly agree with you. I prefer to help teach dogs what we want them to do by replacing the behavior, and if
    I see that removing rewards is helpful to the dog’s behavior, I will give the dog that information as well.

    Thank you for your time!

  5. I was the asker of this question, and my intent was not to interrupt the speaker with an off topic and tunnel visioned question, but to ask whether she was familiar with the study cited and whether she had an opinion on it. Personally I think that positive reinforcement and negative punishment both have important places in behavior modification, and I employ both of them when working with dogs. However, dogs with unknown genetic and behavioral histories and/or little prior training are often not equipped to deal with frustration during training. Some breeders select for dogs that handle mild stressors well, and advise their puppy buyers and training students how to give the dog tools in puppyhood to work through complex problems. These dogs likely respond well to negative punishment. A dog with no or little knowledge of how to learn or how his behavior alters consequences, particularly if working with a similarly inexperienced owner, *may* be being asked to handle an unfair level of stress when imperfect timing is combined with a predisposition to anxiety.

Leave a Reply

Your email address will not be published. Required fields are marked *