Introduction:
There are four quadrants of operant conditioning, which is when you provide consequences in order to increase or decrease behaviours. They are: positive and negative reinforcement, and positive and negative punishment.
Reinforcement is a consequence applied to encourage a behaviour; punishment is a consequence applied to discourage a behaviour. Positive refers to adding a stimulus, while negative refers to removing a stimulus:
Positive reinforcement:
Positive reinforcement is when you add a reward (usually a treat, but sometimes toys or interaction with an owner) to encourage a behaviour. You are teaching your dog ‘when I do this, I am rewarded’ so they will try doing it again. Clicker training, which I use, is an easy way to positively reinforce behaviours you want to see from your dog.
Examples: Your dog comes when you call his name, so you give him a treat. He may come more frequently, because he associates coming when called with being rewarded.
Your dog jumps up at you when you get home and you pet and speak to her soothingly (‘It’s okay, Daddy’s home now, calm down’). Now the dog has learned there is a positive reward for her behaviour (jumping up) and will do it again.
Negative reinforcement:
Negative reinforcement is removing unwanted stimulus (pain, loud noise, bad smells, etc.) in order to reward for performing a desirable action. It operates by teaching the dog ‘when I do this, what is upsetting me goes away’, leading to the repetition of the behaviour.
Examples: Your dog wears a collar which causes discomfort when she pulls on the leash. When the leash is slack, she is rewarded by the removal of discomfort, and learns it is nicer for her to not use up all the slack when walking.
Your dog is approached by an annoying child who insists on putting their hands in his face. The dog bites the child, making them retreat, crying. The dog has learned that biting gets rid of the unwanted attention, and will likely do it again.
Positive punishment:
Positive punishment is adding a negative stimulus (pain, loud noises, etc.) to discourage a dog from performing an undesirable action. It teaches the dog ‘doing this causes an undesirable result, so I should not do it again’.
Examples: Your dog starts chewing on your carpet. When you see, you make a jarring noise, and the dog stops. You make the noise whenever you see the dog chewing something he shouldn’t be. The dog learns that chewing the carpet when you are around results in an unpleasant noise, and avoids it.
Your dog escapes your yard and runs down the street. You finally manage to coax her back with some tasty food, but when she comes, you smack her across the nose to ‘punish’ her for running away in the first place. Unintentionally you have taught her that it is best not to return to you even when food is offered, as you may hurt her.
Negative punishment:
Negative punishment is removing a positive stimulus in order to discourage an undesirable action. It teaches the dog ‘when I do this, the fun stops’.
Examples: Your dog plays with you excitedly. As soon as you feel skin hit your teeth, you get up and walk away, ending the game. The dog learns that playtime ends if she bites or mouths you.
Your dog is gnawing a bone when your mother comes to visit. Since you want the dog to pay attention to your mother and not his treat, you take the bone away. Your dog has learned that guests arriving = the removal of treats.
Which is best?
Each kind of operant conditioning can yield positive results or unintentional negative ones, as shown above. The truth is, to have a well-trained dog you are likely to need to use a combination of quadrants. Each quadrant has drawbacks:
- Positive punishment – Rated as the most intrusive on the Humane Hierarchy of dog training. Positive punishment can be relatively mild (making a displeased noise), but it’s also where abusive and extreme methods lie (slapping your dog, ‘alpha rolls’, etc.) It can also be difficult to apply correctly/make your dog understand why it is receiving the punishment. It must be applied IMMEDIATELY.
- Negative reinforcement – Can often be considered cruel and increase stress in a dog, which in turn is not beneficial for training, often making this method unreliable. An example would be fitting a dog with an uncomfortable prong collar. The hope is they will learn not pulling stops the discomfort. However the dog may become too stressed to learn, it may take a long time to understand, the dog may associate pain/discomfort with the walker/owner instead of pulling, etc.
- Positive reinforcement – Rated the least intrusive method of all the quadrants. Generally not considered cruel at all, but it can be difficult to reward the correct behaviours at the right time, and teach the dog which behaviours you actually want.
- Negative punishment – The main issue with negative punishment is that it must be applied consistently, which can be tricky. An animal who receives positive reinforcement for a behaviour 1/10 times and negative punishment 9/10 times will keep ‘gambling’ on the chance that this time their behaviour will be rewarded. For instance if you try to teach your dog to be calm when you get home by ignoring when he jumps up at you when you get home, but every so often you forget to be consistent and give him affection when he does it, he may repeat the behaviour.
My experience and combining the quadrants:
In my experience it is generally more effective to rely mostly upon positive reinforcement, with other methods of training supporting it. Not only is it the least intrusive method, and made easy through clicker training, it also prompts the dog to try more things in order to seek a reward.
A dog who is trained only through punishment or negative reinforcement may stop exhibiting novel behaviours and avoid doing anything that could possibly lead to a punishment. That’s why positive reinforcement is the method I use most.
Here, I’ll give some examples of how I use combinations of the quadrants to train our dogs:
- Buddy begins chewing the carpet. I tell him ‘no!’ sharply (positive punishment), making him stop suddenly. I offer a more appropriate chew toy (redirection – another element of good training!), and when he happily chews that I praise him heavily for it (positive reinforcement).
- Lacy sees another dog she wants to play with, and begins barking excitedly. I turn and lead her away from the dog, removing the stimulus/opportunity to play (negative punishment). When Lacy encounters another dog and wags politely, I praise her and allow her to go play with them (positive reinforcement).
- Both older dogs like to pull on the leash. When they get to the end, I stand still so they experience the uncomfortable pressure of the leash becoming taut (negative reinforcement) and cease interacting with them (negative punishment). When they move backwards I praise and reward them (positive reinforcement) when the leash is slacker.
As you can see, even when I use other sections of the quadrant to train our dogs, I always ultimately fall back on positive reinforcement. Everybody has their own method of training, but it’s my belief that the best methods rely heavily on positive reinforcement and not punishment.
Hopefully this helps you to understand the quadrants of operant conditioning and how to use them! Next week I’ll post about classical conditioning and how it can help with other issues, and on Thursday I’ll share an introduction to clicker training.
That Dog Kid 