Archive for January, 2012

How do you know that positive reinforcement works?

Tuesday, January 24th, 2012

The proof is in the pudding!! Where pudding is the positive reinforcement, of course!

Many owners question me about how positive reinforcement works, and if it works at all. In trying to move away from the old style ways of forcing compliance in a dog, or in other words, being the alpha, implementing positive reinforcement techniques seems to contradict everything that dominance model training has taught us.

Positive Reinforcement is a technique founded in canine learning theory. This can be explained in a very detailed, very complex way, but I’ve found that if we just watch our own dogs for a little while, you can see it demonstrated right there in front of you. Your dog can learn without it’s human counterpart “training” him – those actions that seem to just come naturally are typically learned behaviours, we just didn’t realize that we were teaching them.

Consider the more common complaints of dog owners; my dog won’t come when called and when I try and catch him he runs away, or my dog is always jumping up on people, or the famous ‘she’ll do anything if there’s food around!’ (and I still don’t understand why this falls in the complain department!). There are also the less noticeable behaviors, such as sitting near the kitchen counter when it’s doggy dinner time, running over to the door when it’s time to go out, or dropping the ball in front of you during a game of fetch. If you really think about it, most owners didn’t teach their dogs these things on purpose – it just happens. But the dog has actually received positive reinforcement for each of these behaviours, which is why she does them reliably.

It is in these observations that I see the simplicity in learning theory – dogs will most always do what gets them what they want (sorry to burst your bubble – but they’re not doing it to please YOU!), and they will take the easiest route to get to what they want. Through trial and error, they determine what behaviours are most likely to grant them positive results. Simple as that! The key there is ‘most likely to’ – meaning that the desired result may not happen 100% of the time, but it does happen at least 51% of the time. Dogs are smart enough to do that math, and it applies to everything they learn in life – from humans, from other dogs, and from their general environment.

Here is the how and what of the above noted learning experiences:
• Poor recall – he doesn’t come when called and runs away when you try to catch him. Dogs love to play chase – he has learned that when you say Come, lean forward, then begin to approach him, you are initiating a very fun game of chase. It’s a favoured game, and he can rely on you to chase him after that sequence of events. Often, this happens at the park, and that is because he has also learned that if he doesn’t play chase with you and just comes over to you, his time at the park will be over, and he’s not ready for that yet.
• Jumping up on people – inevitably, if your dog jumps up, the person she jumps on, or you, will give her some form of attention (even if it’s ‘negative’ attention). She has learned that jumping up gets attention, and that staying quiet on the floor gets NO attention. Dogs are social creatures, and they want to interact (and smell what you’ve been eating!), so if sitting quietly gets them no interaction, and jumping always does get them interaction, they have learned that jumping works best.
• He’ll do anything when there’s food around – quite frankly, this is a positive thing! But, the aspect of learning is this – if owner has food (especially treats, which the dog knows by scent and history, are for him) it’s very likely that if he complies with the request and complies quickly, the treat will be given. This is the foundation of positive reinforcement (though we don’t always use food).
• Sitting near the counter at doggy dinner time – that one’s easy!! The dog has learned that certain cues (picking up the bowl, opening the food etc.) will result in being fed very soon. The closer she is to the food, the more quickly she’ll be fed. This one can be broken down even more, as it involves the skipping of cues. As a pup, before the cues became engrained, you would likely call your pup to the kitchen (or feeding area) once the food has been prepped. Now the adult dog has learned that they don’t have to wait to be called – once they hear that sound, some delicious food will be placed.
• Running to the door when it’s time to go out (or any consistent, natural behaviour your dog does to indicate that he needs to go outside to pee) – this often occurs after several behaviours have been tested out by the dog. Eventually, they find the one that gets your attention, and therefore they repeat this behaviour because when they do, it results in getting to go outside. Sometimes dogs aren’t so clear about their intentions, and the reward of going out happens by human coincidence. For example, my own dog has taught me that he needs to go out when he sits near me but with his back to me. It’s likely that this behaviour began to work for him because it’s an odd behaviour that caught my attention – though for a few weeks, it resulted in a list of questions starting with “do you want….??”. But now that he knows I’ve figured it out, he sits backwards reliably when it’s time to go.
• Dropping the ball in front of you during fetch – this typically comes quickly to the natural retriever, but the opposite (NOT dropping the ball) is also inadvertently taught by the owner. Dogs who regularly play fetch learn very quickly that the sooner they drop the ball, the sooner we will pick it up and throw it for her. In fact, I had one client who often played fetch while chatting about his day with his wife, and was slow in noticing the dropped ball. The dog learned not to just drop the ball, but to bounce off his owner’s chest with his two front feet, and then drop the ball. This was not a desirable behaviour for the owner, but for the dog, it was very effective in getting him what he wanted quickly – the ball was thrown right away. Dogs who refuse to drop balls often do so because it is at least 51% likely that the ball will be taken away from them – therefore, they have learned that it is in their better interest to keep the ball. These dogs are also often very good at initiating chase!! The dog who believes that it’s likely that his owner will pick up a dropped ball and immediately throw it is the dog who learns to drop the ball without being asked to.

So, when you wonder why some trainers are so emphatic about using reinforcement techniques when teaching your dog, it is because it works!!! The hard part can sometimes figuring out what that enforcement should be, and at what point in the behavior to use it. Ultimately, it all comes down to showing the dog that the behaviour we want will very likely be adequately reinforced the majority of the time. We teach the dog what works for them, and what works for us!


The Dog's Assistant - Canine Behaviour & Nutrition Consulting - © 2010.

Website Designed & Coded By - www.kylegallant.com

The Dog's Assistant is proudly powered by WordPress