Operant Conditioning in Dog Training

Okay, everyone knows this one. Tell the dog to sit, dog sits, give the dog a goodie. Seems simple, huh? Well, the problem is the dog will often turn it around on you.

For example: You like to click away on your computer. Fido likes you to pay attention to him and function as a 24 hour doggie entertainment system. You click. Fido sighs, gustily, from his bed in the corner. You click. Fido gives you pitiful puppy dog eyes. You click. Fido drops his fuzzy monster toy on your foot. You click. Fido shoves the fuzzy monster on your key board. You laugh, toss the toy and try to click. Ooopsie, there's fuzzy monster toy on the keyboard again. Toss. Click - Oh, looky, there it is again, right in the way.

Another example: Spot barks. You say "oh, what's the matter?" Spot wags his tail and his eyes light up. You go back to what you are doing. "Woof." You ask "What's up buddy? You want a biscuit?" Spot happily crunches up the biscuit.
"Woof."
"Now, that's enough, buddy, you be quiet."
"Woof!"
"Here's your toy, now Shhhhhh"
"WOOF"
Silence
"WOOFWOOFWOOFWOOFWOOFWOOFWOOFWOOFWOOF"

Dogs are very good at training us. They've been at it for eons.

Negative reinforcement is a little different than positive reinforcement. The dog is still rewarded, but he is rewarded by avoiding an anticipated undersireable consequence, rather than getting something.

Example 1:

"Rover, do you need to go outside?"

It's rainy and cold out there. Rover looks at you, not moving from his nice cozy bed.

"Rover, come on. Outside"

If possible, Rover gets even flatter in his bed.

"Rover. Here. Outside. Potty. Now."

Rover is one with his bed. All you see are eyes.

Oh, well, you think, he must not need to go out.

Example 2.

Trash is scattered all over the kitchen floor. Guess who did it?

"That bad dog. Rover! Come here!"

Rover hears your tone and knows he's in for a scolding.

"Rover, Here!"

Rover peers around the corner, takes one look at you and runs off.

"That's right, you had better skaddaddle, you bad dog."

Guess what? You've just taught Rover not to come when you call.

Whether it's called a "correction" or "scolding" or whatever, we are all very familiar with this one. It is the one that is overused, most times. Punishments extinguish behaviour, which means they make the dog stop doing something to avoid the consequence.

Remember the example under Negative Reinforcement about the tipped over trash? If you call Rover to you and he comes, he gets scolded. So, he gets punished for coming when you call. He gets negatively reinforced for running away, because he gets to avoid the punishment. Never call your dog to you to scold it, or to do any unpleasant thing to the dog. Go get the dog. You need your dog to come when you call and a bad recall is one of the most common training problems.

Positive Punishment, or "corrections" is one of the most overused techniques in dog training. It is easy. It is usually quickly effective. It requires virtually no thought or effort. But it doesn't teach your dog to do anything, it just teaches your dog to avoid something. Overuse leads to depressed, lethargic, overweight, or neurotic dogs. This is not to say that you should never correct your dog, simply that you should use the entire range of motivators rather than simply relying on squelching misbehaviour.

You've made some microwave popcorn and are sitting in the living room, watching television. Spike is watching you intently.

"Lay down," you say, and Spike drops like he's been shot.

You toss him a piece of popcorn.

Time passes. You get involved in the television show. Spike watches you intently, glued to the floor. But, no more popcorn comes. Simply lying there, quietly, being a good dog, stops the flow of popcorn treats. Spike whimpers a little, or wriggles, or does something to get your attention, and you remember to toss him a popcorn bit. More time passes. Spike learns being quiet doesn't get you anything.

In another example, you are housebreaking Puddles. Everything seems to be going well. You take her out, tell her potty, and she potties and you bring her back in. But, then things start deteriorating. She starts holding out on you. You have to walk her and walk her to get her to pee. Puddles has learned that if she goes immediately, you just take her back inside and ignore her, but if she holds out, she gets attention and a long walk.

There is another factor that greatly influences the training of your dog. It has to do with how frequent and consistant the reward is. If you say "sit" and your dog sits and you treat him ten times out of ten, he is very likely to quickly learn to sit when you say sit. If you treward him one time out of ten, he is more likely to wander off to find something stinky to roll in. So, the more predictable the reward is, the quicker the dog learns to associate it with the behaviour. Similarly, if you reward him ten times in two minutes, he is more likely to associate the reward with the behaviour than ten times over two years. The same is true if you reward him one second after the behaviour compared to one week after. Timing, frequency, and repetition are all very important.

However, while being very prompt and consistant with the reward is important in establishing a behaviour, any interuption in that absolutely consistant reward can quickly extinguish the behaviour. In "fixing" it so that it endures, a more intermitant schedule of rewards works better. What this means is that as you are teaching the behaviour to the dog, you reward each occurance, immediately, every time. After the dog performs the behaviour consistantly, than gradually wean him back from being rewarded every time to being rewarded occassionally.

When training a dog to perform a swimmer's turn off the box in flyball, for example, you don't just wait for him to do it. Instead, you might start by get a device called a "clicker" and clicking and treating until the dog looks expectant everytime he hears the click (this is called "loading the clicker") Then you teach him to touch the end of a stick with his nose and click when he does it. Gradually, toouching the stick requires more and more acrobatics from the dog. Eventually, the stick is used to lure the dog into bouncing off the wall or some other verticle surface. That surface is transitioned into to angled surface of the flyball box and the dog has learned to rebound off the box in a swimmers turn.

This concept of teaching a complex behaviour in bits and pieces that gradually conform more and more to what you eventually want is "shaping." Shaping is very important in dog training, because so many of the behaviours we ask of the dogs are too complex, or too far removed from their instinctual tendencies, for them to immediately perform successfully.

What happens when a subject is cannot avoid punishment? It was a question that inevitably came up and was investigated. A rat was put in a cage. The bottom of the cage was a wire grid that could be electrified, giving the animal a mild electronic shock. The top of the cage was closed, so the rat couldn't jump out and there were no places for it to safely perch and avoid the shock. So, the scientists started shocking the rat. At first the rats would jump and scramble and attempt to escape. Eventually, though, they seemed to give up. They languished on the floors of their cages, calmly submitting to being zapped over and over, and didn't try to escape, even though the top was opened and way to avoid being shocked revealed.

The rats had developed "learned helplessness." It's a concept very familiar to anyone who has worked with or attempted to treat depression.

There is a point to this horrible little story. Just like the rats, dogs can develop learned helplessness. It is very calm, submissive state and the depressed dogs will be very easy to manage. Behaviours like aggression and over-excitability will be extinguished. However, it is not a desirable nor a healthy mental outlook. There is a hazzard in training dogs to rely almost exclusively upon punishments and corrections, rather than taking the time to look for what is motivating and reinforceing the undersireable behaviour and shaping and consistantly rewarding more desireable behaviours. Inducing depression makes a more manageable dog, but there are better techniques and trainers out there.