The Use of Rewards and Punishment in Dog Training
Dog Training: Rewards and Punishment
Animals are Blessed With Their Own Uniqueness
There is no shadow of doubt that rewards and punishments are subjective in dogs. As a dog owner, trainer and behavior consultant, and most importantly, as an animal lover, I value the life of animals, big and small. When one of my chickens is sick, people often remark, "It's just a chicken, aren't you doing too much?" Yet, I know in my heart that all living creatures in this world have the right to live and be the unique beings they deserve to be. On the other hand, I am often surprised when people dedicate so much love and care for a car, jewelry and a collection of glasses. These are manufactured by machines. If you are a collector of glasses, most likely, should you break a glass, you can easily replace it. You will, therefore, find an identical specimen of the same color, same material and same pattern. Same goes with a car, set of earrings, or fine china. But a living animal? You can never ultimately replace a living animal, since, just as people, they are completely unique in their own little ways.
Interestingly, training different species of animals (I have a hobby farm with 2 dogs, 2 cats and 15 chicken and my own training center) has brought me to further feel this uniqueness. Being a force free trainer, I have noticed that what is rewarding to one animal, may be punishment to another and vice-versa. Indeed, when my clients apply punishment, I have also noticed that what may be in their own eyes punishment can be in the dog's eyes a reward. But how can this be possible? Simply, because animals are unique. This is what makes training animals and studying their behavior so interesting and fascinating.
What is Punishment and What is Reward in Dog Behavior and Training? A scientific Review
In order to further experience and understand this uniqueness in animals, I want to clarify the terms of reinforcement and punishment in dog behavior and dog training. This topic often causes people to navigate through murky waters. It is astonishing, the amount of websites out there, some even written by professionals, that confuse the terms. Steer clear of websites or trainers who claim negative reinforcement is synonym with punishment! This is far from true and the website owner or trainer needs to do some homework!
What Exactly is Reinforcement?
In behavioral science, reinforcement simply means increasing, thus repeating. A behavior is, therefore, reinforced when it increases in frequency. There are two different types of reinforcement; positive reinforcement and negative reinforcement. Behaviors resulting from using one or the other, have to result in an increase in frequency. According to B.F. Skinner, "behaviors that are reinforced, tend to be repeated and strengthen".
*Note: in behavior terms, the words positive and negative are not used to denote good or bad. Positive simply means addition and negative simply means subtraction! Some unethical people are calling themselves positive trainers. If you find a positive trainer who uses a choke, prong or electric collar, he is using the term positive to depict "positive punishment" not positive reinforcement!.
- Positive Reinforcement
In positive reinforcement, the term "positive" means adding something for the purpose of making a behavior increase (reinforcement). If you give your dog a cookie every time he sits, you are adding something (addition, positive) for the purpose of making the sitting behavior increase and become more reliable over time.
- Negative Reinforcement
In negative reinforcement, the term "negative" means removing something for the purpose of making a behavior increase (reinforcement). If you raise puppies, you may decide to pick up a puppy one day and hold him in your arms. If the puppy struggles, you may stop holding him (subtraction, negative) and put him on the floor once he relaxes . With time, the behavior of relaxing will increase and become more reliable over time.
What Exactly is Punishment?
In behavioral science, the term punishment simply means reducing, thus stopping. A behavior, is therefore, punished when it decreases in frequency. There are two different types of punishment: positive punishment and negative punishment. Behaviors resulting from one or the other have to result in a decrease in frequency. According to Thorndike's Law of Effect: "responses that produce a disadvantageous effect become less likely to occur again in that situation".
*Note: in behavior terms the word punishment, is not meant to denote anything hostile, just as the word positive is not meant to denote anything pleasant, both terms simply depict behaviors that increase or decreases in frequency.
- Positive Punishment
In positive punishment, the term positive means you are adding something to make a behavior decrease. For instance, if you say the word 'NO" (addition, positive) when your dog is scratching the door, and your dog stops the scratching behavior (punishment), you used positive punishment.
- Negative Punishment
In negative punishment, the term negative means you are removing something to make a behavior decrease. For instance, if your dog likes to nudge at you to be pet, if you turn away and leave (subtraction, negative) your dog will decrease the nudging behavior (punishment).
While trainers may elect to use mainly a couple of quadrants or use a combination of these quadrants, it is also important to recognize that these quadrants are not only strictly taking place when you are training your dog. The outcomes of these four quadrants, indeed, tend to occur on a frequent basis in your dog's interactions with his environment. Understanding these quadrants may explain and unlock the reasons why some behaviors may be continuing despite your intervention. This is why it is important to recognize that punishment and rewards are ultimately subjective. The next paragraph will outline some of my personal observations.
Reward and Punishment From an Animal's Perspective
What inspired me to write this hub are my experiences and observations derived from my studies to become a dog trainer, experiences training different species of animals and eagerness to watch animal behavior unfold beneath my eyes. Behavior modification happens on a daily basis and explains why your dog may keep on raiding the trash, your cat escaping the yard and your puppy jumping on you. Reinforcement and punishment is highly subjective, and varies from animal to animal and species to species. Following are some examples and interesting dynamics from my personal observations.
Case# 1: "Why is my dog still jumping on me no matter how much I punish him saying no, pushing him away, yelling at him and even squeezing his paws?" A human's perspective versus, a dog's perspective. Rewards and punishment are subjective and vary from one species and another.
First of all, let me point out I do not encourage such forms of correction, but many people contact me with this type of question, so the above is just for the sake of an example; again, I do not advocate such strategies! Second, according to the four quadrants above, the owner thinks she is "punishing her dog" when in reality she is not. Let's remember that punishment applies when there is a decrease in behavior, in this case, the dog is still engaging in the unwanted behavior. Here is where it gets tricky: what looks like punishment from the owner's perspective is more likely reinforcing from the dogs' perspective! This is where, the notion of rewards being subjective and varying from species and species comes into play.
Let's see it from a dog's perspective. Kept alone for most of the day, it must feel like a long time to be socially isolated. The dog may hear noises near the property and he may repeatedly get up in hopes it is his owner. Yet, it is not. Exposed to these false alarms and bored, his eagerness to see his owner again grows by the hour. For a dog, 8 hours alone can be a very long time. I don't think there are any studies on this but I think dogs must perceive time differently than us. So when he finally hears foot steps and the key turning into the door, excitement builds up. Obviously, by the time he sees his owner he no longer can contain himself!
The owner looks at the dog after a long, stressful day at work, says no, (looking and talking to the dog are forms of attention), but since the dog does not stop jumping, she then starts pushing him away. Since this does not work either to stop the jumping jacks, the owner may like increasing her coercion-based methods. Problem is, when dogs are so excited, they may not feel pain as they normally do or perhaps they feel it, but the eagerness to be near the owner overshadows it. Ever seen a dog in pain at an animal hospital wag his tail upon seeing the owner? This lack of pain, or better to give the benefit of doubt, lack of pain manifestation, may therefore, lead to the belief that more and more coercion is needed which can get close to abuse! What is happening is that the dog is so eager of attention that he will get ANY form of attention (looking at the him, talking to him or touching him to push him away will do) even if it is negative!
So in order to stop the behavior, some form of punishment is required to decrease behavior. Since, excitement may blur any discomfort or pain and interferes with learning, positive punishment is the wrong way to go. Indeed, if an electronic collar trainer would apply shock to correct this behavior, he would likely find himself upping the level more and more. With time, the dog would require higher levels of shock. Not only, through associative learning there would be risks for the dog to start associating the presence of the owner with pain! Positive punishment, while effective, has the potential for deleterious side effects. Something serious to keep in mind before applying it or hiring a coercion-based trainer!
Training to Stop Jumping Using Negative Punishment and Positive Reinforcement
The best approach in this case, is to use negative punishment. This method decreases unwanted behaviors without having to resort to coercion-based methods. With the removal of positive attention (looking at the dog, petting the dog, talking to the dog) or negative attention (staring badly at the dog, scolding him, or pushing him away) the jumping behavior will gradually decrease with time. Avoid looking at your dog, talking to your dog, touching your dog when you come home from work. Act neutral. If your dog jumps, turn your back away and if he continues jumping, leave the room or sit down and completely ignore him. The jumping behavior should dissipate as the excitement fades away.
When your dog is calm, then you may use positive reinforcement. Call him to you, ask for a sit and pet him. The addition of petting will cause an increase in sitting behaviors. Also, since the excitement is away, he is better able to learn. If he starts getting excited, again, use negative punishment again. Then revert to positive reinforcement once he is calmer and displays wanted behaviors. As an opportunist, your dog will soon learn which behaviors are the most advantageous!
Case #2: Why is my dog still barking even if he is getting squirted with a water gun? Rewards and punishment are subjective and vary from one animal and another.
A while back, I had a client with a Lab and a barking problem. She took very good care of this dog. Tired of hearing him barking and willing to implement some rules, she took her dog to a pet store for training. The trainer's advice was to squirt him in the face every time he barked by using a squirt gun. The owner graduated from class, his dog learned how to sit, stay, lay down etc, but he STILL had a barking problem. She then contacted me.
I went to her home and watched the behavior. She got her squirt gun from a drawer, sat on a chair and here came the Lab which started barking in a high-pitched tone of voice. She squirt him in the face, he opened his mouth and appeared unfazed by the squirt. He then barked again and the cycle continued. She said ''See? the behavior does not stop, yet, I am doing exactly what the trainer said and the trainer even said my timing is good".
While she was confused, I had a darn good idea of what may have been going on. I asked her to put the dog in another room and place the water gun in the drawer. The lab in the meanwhile started barking at something out of the window. So I went to get the squirt gun from the drawer and surely enough here came Rover. He immediately barked and I squirt the gun and stopped and he barked. Squirted and then stopped and he barked. A pattern disclosed what was obvious to me but puzzling to the owner: this Lab loved water! Indeed, his barking was high-pitched and upon being squirted, he tried to catch the water with his mouth and then barked when I stopped simply asking for more squirts so he can play more!
When I told the owner she was amazed, but honestly not too much. After all, she recalled he was doing the same behavior in the summer when the sprinkler was on. Barking and playing with the water. I told her to retire the old squirt gun and increase his level of exercise and mental stimulation. His barking decreased except the occasional extinction burst (read more about extinction bursts in my hub: dog behavior modification terms.
This is another perfect example of how rewards and punishment are subjective from animal to animal. In this case, the owner thought she was punishing her dog, when instead, her dog found the water amusing, indeed, the barking behavior increased rather than decreasing. The trainer also thought it was punishment because this method had worked with other dogs in the past. Rewards are subjective. If your dog is not decreasing a behavior, consider if there is a rewarding element somewhere before rushing to aversion-based training. If you remove that rewarding element you may get close to the resolution of the problem in a humane, force-free way.
Case # 2 Punishment and Rewards are Subjective and vary from animal to animal. How temperament and past life experiences affect operant conditioning outcomes.
As I mentioned earlier, animals are unique and their perception of the world is unique too. In the following examples, I will be talking about my cats and dogs. Upon moving to Arizona, we noticed our 3 acres were covered in something I never was aware of their existence before; burrs! These nasty elements were everywhere! I could not find a way to get rid of them!
My female Rottweiler Petra, could not tolerate such burrs. When sent out to potty, they would get stuck in her paw pads and she would stop walking, lift her leg and look at me imploring for help. I therefore, had to remove the thorny mess one by one. From her perspective, the yard became a place of positive punishment, the addition of burrs, decreased her will to be out of the yard. She started looking at me sadly when she had to go potty. I had to take her on walks four times a day to prevent this issue and I had 4 acres of land!
On the other hand, my male Rottweiler Kaiser, cared less about the thorns. When he was sent out to potty, he did his business and then went searching for wild rabbit poop, which he perceived as a delicacy. From his perspective, the yard became a place of positive reinforcement; the addition of the rabbit poop, increased his will to go out. Indeed, when I said let's go potty, he was happy and eager to look for his dessert. The rabbit poop was also much more salient than the thorns, meaning they had more value than the burrs in his opinion! Of course, I walked him too four times a day before clearing an area from burrs and fencing up to stop all the rabbits from coming in!
Case Number 3: Rewards and Punishments are subjective from one animal to another. Observations of my cats.
A similar experience occurred with my cats. In the first weeks of owning two rescues cats, at dinner time, we got into the habit of feeding them food and closing them in a separated room so we could eat in peace. From my male cat's perspective, the addition of food made his desire of getting inside the room worth it. Indeed, he frequently often walked us up to the room to be fed when he was hungry. The room was a place where positive things happened. Because he was rewarded by walking in the room with food, his behavior of walking to the room increased, thus, positive reinforcement took place.
My female cat, was a whole different story. She really dislikes being closed in any room. I have no idea why she has this issue as she was a rescue. Anyhow, upon closing her, she cared less about the food. She just started crying and crying. From her perspective the action of adding the food to the room, had assumed a bad stigma, suggesting she would be closed. Through associative learning, the food was a predictor of bad things. Upon seeing the food, she backed away. While before she used to sleep in that room, now her behavior of going inside the room reduced and even extinguished, thus positive punishment took place.
Case Number 4: Rewards are subjective too! Petting a frightened, stressed dog. Another case of a pet store trainer's bad advice!
This story comes from when I was contacted by a lady with a newly rescued dog that started biting. The owner took her dog to a class in a pet store. Because this dog was very frightened by being around other dogs and people, she would not take treats. She told me during class her dog was shaking for most of the time. This dog was not ready for classes! She was just adopted, given little time to get acquainted with the owner and was being flooded with stimuli which were too much for her at once!
The pet trainer told the dog owner that since she did not eat treats, she was to pat the dog on the head instead. So when the dog sat, she pat the dog on the head. So this is what she did over and over. The training was not working though. Instead of sitting more, she was sitting less and less (let's remember that reinforcement increases behavior, not decreases it) so the owner started pushing her rump down and then pat her on the head to reward, until one day she bit the owner on the hand . She had to stop attending classes due to the new aggressive display, so she came to me.
Moral of the story: the pet trainer assumed that just because many dogs like to be pet, petting was the best reward for this fearful dog. I noticed right away this dog was head shy and the trainer failed to see the dog's fearful body language. Being pat on the head frightened her and was perceived as positive punishment. The addition of petting caused the sitting behavior to not increase; actually, it was never allowed to establish at all!
Another case showing how rewards and punishment are subjective and why no cookie-cutter approaches should be applied in dog training. If you need to change behavior, put some effort in recognizing what your dog finds most rewarding and identify reinforcers that may be keeping unwanted behaviors alive.
Victoria Stillwell Explains why Positive Reinforcement is the Best
How do I train? My Choice of Quadrants
Positive reinforcement, negative reinforcement, positive punishment and negative punishment are often referred to as the four quadrants of operant conditioning. I tend to train using mostly positive reinforcement and, at times, negative punishment. Some balanced or "old-school" trainers often tell me that if I fail to use all quadrants of behavior, I am not training to the full extent I should. They often tell me that a certain level of positive punishment and negative reinforcement is needed to stop unwanted behaviors. I refuse to apply those quadrants. But nature will at times; -just think a dog being sprayed by a skunk, after some time, he may decide to avoid all skunks in fear of being sprayed. I don't want to become a skunk, ie a predictor of negative experiences. I want to create a bond and have fun in training because the dog wants to, not because he must out of fear and negative connotations. Studies, interestingly suggest, that dogs that are shocked, with time, learn to associate the presence of their owner and commands with receiving shocks, even out of the training context!
Electronic collar trainers and old-school trainers use a good chunk of positive punishment in their training. They add shock to stop unwanted behaviors (momentary correction, positive punishment) or remove shock to increase wanted behaviors (continuous correction, negative reinforcement). Balanced trainers may use a combo of all quadrants instead. So how can I train using mainly positive reinforcement and negative punishment? How can I train dogs and solve major behavior problems using gentle, force-free training? Easy as pie and scientifically proven.
When training, I rely on Skinner's revision of Thorndike's Law of Effect. While Thorndike's Law of Effect claims "responses that produce a satisfying effect in a particular situation become more likely to occur again in that situation, and responses that produce a discomforting effect become less likely to occur again in that situation.” Skinner's revision claims that in reinforcement, behavior which is reinforced tends to be repeated (i.e. strengthened as in positive reinforcement); whereas, behavior which is not reinforced tends to die out-or be extinguished (i.e. weakened as in negative punishment).
When doing behavior modification, I use humane dog behavior modification. I work on the dogs inward (emotional state) rather than suppressing the outward manifestation. To read more about these positive techniques read: A guide to behavior modification terms and techniques.
My training methods are, therefore, not only humane but also highly effective. They are based on science and are recommended by major organizations. The American Veterinary Society of Animal Behavior's position statement (AVSAB) recommends that QUOTED:
"Training should focus on reinforcing desired behaviors, removing the reinforcer for inappropriate behaviors, and addressing the emotional state and environmental conditions driving the undesirable behavior".
They also advocate against the use of choke chains, pinch collars, and electronic collars as a first-line or early-use treatment for behavior problems due to potential adverse effects such as the inhibition of learning and increase of fear and aggression.
The Association of Pet Dog Trainer's position statement claims QUOTED:
"Physical or psychological intimidation hinders effective training and damages the relationship between humans and dogs".
"The APDT advocates training dogs with an emphasis on rewarding desired behaviors and discouraging undesirable behaviors using clear and consistent instructions and avoiding psychological and physical intimidation. Techniques that create a confrontational relationship between dogs and humans are outdated. Modern scientifically-based dog training should emphasize teamwork and a harmonious relationship between dogs and humans that fulfills both species' needs. Most of all, it should be a fun and enjoyable experience for everyone involved."
Hope this guide on dog rewards and punishment helped, happy training!
Alexadry© All Rights Reserved, Do not Copy
For further reading
- Dog Training: How to Shift From Compulsion to Reward...
Looking for kinder and scientifically approved training methods? Reap the benefits of training reward-based training methods for a better bond with your dog!
- Evaluating Cesar Millan's Dog Rehabilitation Methods
Dog Whisperer dog rehabilitation methods evaluation. Pros and cons of Cesar Millan's training methods.
- Dog Behavior: Can you Reinforce Fear?
We are always told not to pet, cuddle or comfort a fearful dog because this may reinforce fear. But can fear really be reinforced? Learn what the experts say about this.
- Dog Behavior: A Guide to Behavior Modification Techn...
Say NO to coercion and say YES to rewards. Learn effective modern based dog behavior modification techniques. As the words imply, behavior modification entails modifying a dog's behavior for the purposes of increasing or decreasing wanted and...
- The Use of Aversives in Dog Training: Pros and Cons
Training with aversives Dogs are exposed to various stimuli on a daily basis which are for the most part perceived as positive, neutral or negative. Positive stimuli are obviously those that are associated with positive emotions, either through an...