Let us first address the concept of science based training. Many force free training advocates will tell you that their approach is supported by science. That animals trained with positive reinforcement are proven to learn new behaviors more quickly then if they are punished. Guess what! This is absolutely true. Unfortunately this is not the complete truth.
BF Skinner is the noted behaviorist best known for his development of the learning theory, ‘Operant Conditioning’ which is recognized as one of the most comprehensive behavior theories in existence today.
Skinner noted that humans and animals learned new things more quickly when they were rewarded as opposed to punished. There has also been similar research to support this assertion with a variety of animals including marine mammals.
However, here is where they don’t tell you the entire truth. There are FOUR quadrants to Operant Conditioning. They are as follows:
1 . Positive Reinforcement
Strengthens a behavior by providing a consequence an individual finds rewarding. For example, your dog is hungry. He sits and you give him a piece of food. He is now more likely to sit.
2. Negative Reinforcement
The removal of an unpleasant reinforcer can also strengthen behavior. This is known as negative reinforcement because it is the removal of an adverse stimulus which is ‘rewarding’ to the animal or person. Negative reinforcement strengthens behavior because it stops or removes an unpleasant experience. Example, you pull up on your dog’s leash and tell him to sit. When he sits you relax the leash. He is now more likely to sit when he hears the command to avoid the leash pressure.
3. Positive Punishment
Sometimes referred to as punishment by application, involves the presentation of an unfavorable event or outcome in order to weaken the response it follows. Example, your dog chases the cat. When he chases the cat you smack him on the nose. He is now less likely to chase the cat because of this consequence.
4. Negative Punishment
Also known as punishment by removal, occurs when an favorable event or outcome is removed after a behavior occurs. Example, you are holding a toy your dog wants, you tell him to sit, instead he lies down. You don’t give him the toy. He is now less likely to lie down next time you tell him to sit because he did not get the toy for lying down.
In both cases of punishment, the behavior decreases.
As dog trainers we are in the business of creating, managing and reducing various behaviors. Skinner proved that both reinforcement and punishment both impacted behavior. Failing to recognize and use ALL four quadrants of Operant Conditioning is an incomplete approach to dog training.
If all dog training involved, was creating one new behavior after another then indeed the exclusively positive / force free approach would have more traction. However, dog training not only involves teaching new behaviors like down or heel but it also involves making those behaviors RELIABLE or in some cases completely removing unproductive behaviors such as aggression. Positive reinforcement is not always the best tool to address all of these functions.
Mike has a terrier named Bruno. Bruno barks in the house every time a car drives by. Mike is getting complaints from the neighbors about the noise. He seeks help from a local trainer.
She tells him to issue a voice interrupter to Bruno’s barking behavior (basically when Bruno barks, Mike interrupts him with a word like “hey” or “no”) and to praise and give him treats when he is quiet. Mike tries this for 2 weeks.
After a few days of minor improvement things go south and if anything Mike notes an increase in the barking. He finds it difficult to always be on top of Bruno as he is often busy around the house. Also, Bruno still barks when he is away from home.
Why did this not work? Several things are at work here:
a) Bruno like many dogs derives satisfaction out of barking, while he probably knows that Mike does not like this behavior, he also knows that there are no real consequences for this action. Thus like many dogs he chooses to satisfy his need to bark.
b) Bruno was rewarded with food and attention once he stopped barking. This should theoretically have made it more likely that he remained quiet. The problem here is that all positive reinforcement has a beginning and an end. Mike cannot stay with Bruno and ceaselessly stuff food into his mouth. Nor can he pet him 12 hours a day.
When the motivation ceases Bruno has no more reason to be quiet. If anything Bruno likely figured out that his barking initiates the chain of events that lead to him getting attention and treats. Thus the solution to the problem actually made it worse.
c) Mike, like most people, is busy and cannot always be nearby Bruno to follow the local trainer’s strategy. This means that Bruno is receiving inconsistent feedback for his actions. Not only that, science tells us that dogs learn best when the consequences to their actions both positive or negative, occur within 2-3 seconds of the action. Unless Mike is the Flash he won’t always be able to make it to Bruno in time.
Mike has a terrier named Bruno. Bruno barks in the house every time a car drives by. Mike is getting complaints from the neighbors about the noise. He seeks help from Shield K9. The trainer there recommends that Mike obtain a Bark Collar for Bruno. In addition the trainer also teaches Mike a protocol to respond to Bruno’s barking.
The second Bruno barks, Mike issues a verbal correction followed by a physical correction if the barking does not stop. Mike repeats these steps as often as required. Mike immediately notes that the frequency of Bruno’s barking decreases. Especially when he is in the vicinity. However, Bruno still barks when Mike is out of the house.
Mike also purchases the Bark Collar and follows the trainer’s collar conditioning protocol for a week before actually turning the collar on. Within the first hour the collar is turned on Bruno hears a car go by and reflexively barks. The collar delivers a medium electric stimulation. This startles Bruno. He Barks again and receives the same consequence. Two weeks into the training Bruno no longer barks in the house.
Why did this work? There was no motivational techniques used. Surely Bruno would learn more effectively to be quiet when taught motivationally.
a) Bruno learns that the consequence for barking is an electric stimulation he finds unpleasant. This means that he is receiving positive punishment for his barking behavior. Initially this causes him some stress, however over a week or so he comprehends that only barking causes the positive punishment and thus understands how to avoid it.
Once he knows how to make the unpleasant stimulation cease, he no longer has to worry about it.
b) Mike’s goal was to make the barking cease, in other words he wanted to terminate / reduce that behavior. He was not teaching a new behavior but instead removing one. Bruno already knows how to be quiet, teaching him to be quiet again is therefor a pointless exercise.
We know based on Skinners behavior theory of Operant Conditioning that punishment is the most effective way to reduce or remove behavior.
Science tells us that we can affect behavior both through reinforcement or punishment. Bruno, like many dogs is simply exercising choice. His choice was to bark which was the wrong one from Mike and the neighbor’s perspective. When that option was removed he reverted to the only other choice he had which was to be quiet.
I will never disagree that the scientific method is the best way to train dogs. I see it every day in my practice. However, I cannot choose to ignore a whole portion of proven behavior theory simply because it makes me uncomfortable or emotional. Nor should any trainer!
Shield K9 Head Trainer
Hey guys, I will be holding our first Power Obedience Seminar