Dr. Sophia Yin, DVM, MS
Download the Operant Conditioning poster.
(from How to Behave So Your Dog Behaves by Dr. Sophia Yin)
A reader who uses Gentle Leader head collars regularly in training recently asked me about the use of the head collar. States the reader, “I have gotten myself totally confused and need help sorting it out. What category of operant conditioning does the head collar use fall under? Negative punishment? Positive reinforcement? Negative Reinforcement? Positive punishment?”
You might think that being able to correctly determine the category of operant conditioning is just a scientific exercise; however, it’s much more than that. For instance, say you read an ad for a product that says it uses negative reinforcement to stop an unwanted behavior, when the product actually uses positive punishment. The fact that the product manufacturer can’t get the categories correct indicates that they did not have educated behaviorists guiding the development of their product.
While many people are immediately perplexed when they hear the word operant conditioning, it turns out that the principles and categories are actually straightforward. That’s because operant conditioning, or trial and error learning, is simply a description of how animals learn—a description that requires a few important definitions.
Reinforcement vs Punishment
The first two definitions to know are reinforcement and punishment. Reinforcement is anything that increases the likelihood that a behavior will occur again. For instance, if you call your dog and then give him a treat when he comes, he will be more likely to come the next time you call. Thus, by giving him a treat for coming, you reinforce his behavior of coming when called.
Punishment is anything that decreases the likelihood that a behavior will occur again. For instance, if you call your dog and then yell and scream at him when he comes, he will be less likely to come the next time you call. Thus, by yelling at him, you punish his behavior of coming when called. This second scenario may seem an unlikely event, but it happens to people every day. When owners call Rover five or six times before he comes running and then yell at him for taking his time, they are really punishing him for coming when called.
Positive vs Negative
The second set of terms to know are positive and negative. Positive and negative do not mean good or bad; instead, think of them as a plus sign or a minus sign. Positive means that you’re adding something, and negative means you’re subtracting something. Positive and negative can be applied to both reinforcement and punishment.
Combining the Terms
Now we can combine the terms into four categories—positive reinforcement, negative reinforcement, positive punishment, and negative punishment. Here’s what the categories are:
Positive and Negative Reinforcement
Reinforcement can be positive or negative. In either case, we are increasing the likelihood the behavior will occur again. Positive reinforcement means that by adding something the animal wants, you increase the likelihood that the behavior will occur again. For instance, if you teach your dog to come to you by giving him a treat when he comes, you’re using positive reinforcement. By giving him food, which he likes, you’re increasing the likelihood that he will come to you the next time too.
Negative reinforcement means that by removing something aversive, something Fido dislikes, you increase the likelihood the behavior will occur again. For example, you decide to teach Fido to come by putting him on a leash and choke chain. You pull on his leash until he takes a step forward, and as soon as he comes forward, you release the pressure. That is using negative reinforcement. By removing the pressure as soon as he starts coming, you increase the likelihood that he will come the next time in order to avoid the pulling.
Another trick for remembering negative reinforcement is to think of it as nagging. When I was a child and my mother wanted me clean my room, she often had to keep telling me until I cleaned it. I would finally clean my room in order to avoid her aversive nagging.
Positive and Negative Punishment
Punishment can be positive or negative, too. In either case we are decreasing the likelihood the behavior will occur again. It seems odd, but when we talk about punishment, we’re usually talking about positive punishment. Positive punishment just means that by adding something aversive, we decrease the likelihood that the behavior will occur again. For instance, your dog raids the garbage can when you’re not looking, so you booby-trap the garbage with mousetraps. The next time Spot sticks his nose in search of a snack, he gets a mousetrap surprise, which scares him. This booby trap decreases the likelihood that he will raid the garbage can again; thus, it is positive punishment.
Negative punishment means that by removing something the animal wants, we decrease the likelihood that the behavior will occur again. For instance, when dogs greet us by jumping, their goal is to get our attention. If we remove our attention every time Spot jumps by holding perfectly still and even looking away, eventually he will stop jumping. By removing the attention that he wanted, we decrease the likelihood that he will jump again.
How to Classify Techniques in a Methodical Manner
These terms seem straightforward so far, but often when you start classifying techniques in real life, things suddenly get confusing. The reason for the confusion is that some techniques may fall into more than one category, depending on how you describe the behavior and the technique. In order to avoid confusion, you have to approach this classification in a methodical way.
- Step 1: Define the behavior. For example, if the behavior you want to change is your dog’s overexuberant greeting behavior, you may say, “I hate the jumping behavior,” or “I prefer he greet by sitting.” So in general we usually have two behaviors from which to choose—one we want to increase and one we want to decrease.
- Step 2: Next decide whether you want to increase or decrease that behavior. If you want to increase the behavior, you will, by definition, use reinforcement. If you want to decrease the behavior, you will, by definition, use punishment. For instance, if you dislike the jumping behavior and want to stop it, you’ll be using punishment (positive or negative punishment). If you’re going to train him to perform the alternate appropriate behavior of sitting, you will be using reinforcement (positive or negative reinforcement).
- Step 3: Lastly, decide whether you’re adding something or subtracting something to determine if you’re using something positive or negative. If you yank the dog’s collar to make him stop jumping, what are you using? You’re adding an aversive to decrease the behavior, so you are using positive punishment (Note: I’m not advocating that you yank the dog’s collar; I’m just using this as an example). If, instead, you remove the attention that he wants in order to decrease the behavior, what type of punishment are you using? You’re using negative punishment, because you’re removing something he wants in order to decrease his jumping behavior. That is, you’re removing the reward for the undesirable behavior. If you wait until the dog sits and then increase this sitting behavior by giving him a treat that he wants, you are using positive reinforcement. If you hook Fido up to a leash and choke chain and step on it so that it tightens until he sits and then release the pressure immediately when he sits, you’re using negative reinforcement.
Now we’re ready to tackle the original reader question. What category of operant conditioning are you using when you use a Gentle Leader® head collar? It turns out that it depends on how you use the collar. For instance, say my dog sees food on the ground and makes a beeline for it. I react by holding completely still with the leash-holding hand firmly against my side (or more realistically, I’m wearing a hands free leash). When Fido gets to the end of the leash he pulls and pulls for 10 seconds, gets nowhere, and then realizes he can’t get to the food. Then he stops pulling and takes a step back so that the leash is hanging loose. Following the steps described above here’s how we would figure out the category of operant conditioning.
1. First decide what the behavior is. You may define the behavior as pulling on leash, the behavior you don’t like. Or some people may define the behavior as stepping closer to you so he’s on a loose leash (note that “stops pulling” or ‘not pulling” does not count as a good behavioral description). In the case we are describing here, the behavior we are really focused on is the pulling towards the food on the ground.
2. Next, determine if you want to increase or decrease the behavior of interest. In this case I want to stop the pulling. Consequently, the category of operant conditioning will be, by definition, punishment.
3. Third, figure out if you are adding something or removing something to decrease the pulling. The trick here is that you need to remember what the definitions are when you make this decision. With positive punishment, you are specifically adding something aversive (that the dog dislikes—because it’s uncomfortable or causes pain or is just annoying) in order to stop the behavior. With negative punishment, you are specifically removing something the dog wants or removing the ability to get to something the animal wants. In the particular case of the dog stops pulling because he realizes he can’t get to the food. So the category of operant conditioning we are using when we use the Gentle Leader in this manner is negative punishment. If on the other hand we stopped the pulling by having the dog wear a pinch collar and yanking really hard then we are adding an aversive to decrease the pulling behavior; we would be using positive punishment.
But this isn’t the only way that one can use a Gentle Leader® head collar. Let’s say that you have a dog who is staring at a cat and is about to take off in a chase. Before he has a chance, you guide his head towards you so that you can get him to focus on you. When he is looking towards you, you immediately release the guidance and pressure. Then you can engage him in a game such as targeting and reward him for that. Let’s go through the process again.
First we define the behavior. In this case I’m using the Gentle Leader® to increase the dog’s focus on me. So the behavior of interest for me is attention on me.
Next, determine if you want to increase or decrease the behavior. We’ve already stated that my goal is to increase the focus on me, so by definition I’m using reinforcement.
Lastly determine if you are adding something the dog dislikes or removing something he wants. Remember that positive reinforcement is adding something the dog wants in order to increase the behavior. Negative reinforcement involves removing something aversive in order to increase the behavior. And, with negative reinforcement, it’s essential to remove the aversive right as the dog performs the correct behavior. In this case, we are removing the aversive pressure in order to increase the behavior of focusing on us. So, in this case, the use of GL head collar is negative reinforcement.
One point to note is that negative reinforcement can be highly aversive or mildly aversive depending on what is used. For instance, someone could abruptly drag or yank a dog by its head collar in anger until the dog was looking and then release pressure. That could be scary and painful to the dog. Alternatively, one can guide quickly but in a gentle manner without yanking or dragging the dog such that the dog wasn’t scared or feeling pain. This is the way I prefer. So just because the category is negative reinforcement instead of positive punishment, doesn’t make it a friendlier technique. When I use the Gentle Leader head collar, I train owners how to guide the dog in a way that’s clear and not scary for the dog. For instance, we often start by luring with treats to help guide the dog. Then we quickly switch to rewarding the focus with food.
It’s that simple
So there you have it. A fool-proof method for categorizing methods into the proper category of operant conditioning. Now it’s your turn. Think of various techniques you use or interactions you have throughout the day and see if you can categorize them into the correct category of operant conditioning. You can find more practice questions in How To Behave So Your Dog Behaves.
Download the Operant Conditioning poster.