Wednesday, February 29, 2012

Using Operant Conditioning for Dog Training

There are numerous methods utilized today for dog training - but which one(s) should you use to train your own canine?

What is Operant Conditioning?
Operant conditioning is the modification of behavior through the use of consequences (reinforcers and punishers). Although there are arguments against this, operant conditioning differs from classical conditioning in that it deals with changing operant behavior (or 'voluntary' behavior) versus reflexive behavior ('involuntary' behavior). That being said, whenever you're dealing with changing behaviors, operant and classical conditioning can work hand-in-hand.

Operant conditioning has two main tools for modifying behavior - reinforcement and punishment. Reinforcers increase behavior, while punishers decrease a behavior. These operate in two contexts - positive and negative. In this case, positive refers to addition; negative refers to subtraction.

B.F. Skinner was the "Pavlov" of operant conditioning, and actually outlined a third tool - extinction - which is the lack of any consequence. Though it might seem like doing nothing couldn't be an effective training method, it can actually produce results when used correctly.

Positive Reinforcement
An important thing to keep in mind when considering which of these tools to utilize is the long-term effects each can have on your dog. Positive reinforcement is - in my opinion - the strongest teaching tool - not only does it focus on increasing positive behavior; it teaches the dog to want to work with his owner and continue learning and trying.

The basic premise of positive reinforcement is: dog performs behavior, dog gets rewarded. The dog learns - at first - "every time I do this, I get THIS." After a desired behavior is consistently occurring, then the trainer will begin to decrease payment of the behavior, only focusing on the dog's best performances. This encourages the dog to try his best each and every time, since he only gets paid for his best endeavors.

Once a behavior is learned, the dog is only paid intermittently. He's not quite sure when he'll get paid, so once again, he is going to offer his best effort all the time. This form of intermittent reinforcement is often where most trainers who teach clicker training fail to understand the implications of operant conditioning. They teach their trainers to use treats and pay their dog for every single occurrence of behavior. This actually creates a dog who doesn't perform very well; since he's never pushed to give his best, he gets treated regardless of performance and often these dogs are lazy, don't perform behaviors on cue and tend to get lackadaisical about working with their trainers.

Negative Punishment
Negative punishment is also a strong tool to decrease unwanted behavior, since this method focuses on the removal of a desired reward when an unwanted behavior occurs. Trained consistently, the dog will learn whenever she offers a certain behavior, she loses something she's very motivated to have - i.e., a treat, a toy, another behavior (such as going outside, going for a walk, etc.). Other examples would be a dog gets overly excited while being petted, so the person ignores the dog until she settles and then continues petting; or a very a high prey-drive dog not being allowed to play fetch unless she displays the correct behavior. Fetch in this case is not only a game; it can be a very strong teaching tool. Of course, all dogs are different and are motivated - at different times - by different things. You need to choose the best motivator for your own dog in order to truly see results.

Negative Reinforcement
While the concept of training through negative reinforcement may seem counter-intuitive to positive methods, it can actually be a strong tool. Keep in mind 'negative' in this sense doesn't carry an emotional connotation. Negative reinforcement is the increase in desired behavior caused by the removal of an unwanted stimulus.


An example of negative reinforcement would be teaching a dog to respond to touch. If you place your hand on your dog's side and push, eventually he'll move away from the pressure. You're not hurting the dog, but pressure can actually be an adverse stimulus. When the dog moves away, you remove the pressure. This is how horses are trained to move while being ridden - it's called 'leg aiding' and applies the very theory of negatively reinforcing the horse for moving away from a rider's leg when pressure is applied.

Positive Punishment
There are many reasons I don't utilize this type of training with dogs - mainly because with this method you're merely telling the dog "don't do that!" Instead of empowering her with what she should or could be doing, you're focusing on one thing she shouldn't be doing.

In other words, dogs are always doing something, and if you use this method of training, you won't get very far, since it's a major trial-and-error to teach them the things you don't want them to do. That is, the communication focuses on an undesired stimulus (to the dog) such as a shock collar, a yell, even hitting the dog when she offers an undesired behavior (to the human). Even if you use positive punishment in conjunction with the other methods, another consequence from using this type of training occurs: your dog learns to be afraid to try new behaviors. This is because (through intermittent reinforcement) she doesn't know when she'll receive a shock (or other adverse stimulus) for a new behavior.

What you have, in effect, is a dog who doesn't entirely trust you - or herself.

Extinction
Extinction is the lack of consequence after a behavior. This can be utilized by a trainer to decrease behaviors. When a behavior consistently produces no consequence, that behavior will occur with less and less frequency. Keep in mind the dog has to be accustomed to the 'rule': desired behavior results in reward.


However, you might have heard of an "extinction burst." This refers to the dog taking more action in order to gain the desired consequence. For example, when free shaping a dog, you'd start by paying the slightest increments, and once the dog is offering the first increment consistently, you don't pay him the next time. The dog knows this behavior has paid before, so what he'll do is try harder to gain the consequence.


Of course, as mentioned before, if the dog never gets paid for the behavior, he will slowly stop offering the behavior, since there is no gain in performing it.

Remember, whichever method you decide to use, don't forget to have fun!

Thanks to Teaching Dog Obedience for this well written article!

No comments:

Post a Comment