In my mind, the definition of a reward would be a pleasant consequence for a behavior that one would like to teach or reinforce. Ok, so rewarded behaviors increase in frequency, I get it, I get it, there is no need to think any further. I hold a cookie for the dog and tell the dog to sit. The dog sits, I feed the cookie…end of story.
Well, not so fast! There are quite a few things that deserve consideration here.
The most common misunderstanding is the lure vs bribe scenario. Let me give you an example: When first teaching a puppy a basic behavior such as “sit” you might hold a cookie or a toy over his nose. As the puppy looks up to see what’s there he will most likely end up with his butt on the ground which is the behavior you wanted to elicit. You then feed the cookie or play with the toy. Perfect.
However, if in the process you don’t fade the cookie out once the dog begins to understand your hand motion and ping pong your rate of feeding, then the dog will ONLY sit if you have a cookie in your hand. You and the dog become dependent on the cookie and the cookie is a bribe. Your dog will expect the cookie or mistake the cookie as a signal to sit and will only sit when the cookie is present.
The second scenario and the more subtle but worse scenario is: You wave your cookie and call the dog. The dog runs to you and you lock the dog into the kitchen or end some other fun game he was just engaged in. Or you wave the cookie, the dog comes and you start clipping his nails. All of a sudden the cookie turns into a predictor of “something bad is about to happen” and devalues the “wow” effect the cookie had before.
It is therefore important to consider the consequences as well as the value of the reward. If you want to habituate a puppy to something the puppy is scared of or feels uncomfortable about, consider very carefully the order in which the reward appears. Be sure to present the “scary thing” (the harness, collar, the brush, the stairs, the injection needle) first and then give the reward for tolerating the object of discomfort for a fleeting moment or at a closer distance without recoiling. Only then can the dog be classically conditioned to understand injection needle = YAY roast beef!!!
[…] post Rethinking Reward Based Training appeared first on NYC Dog […]Pingback by Rethinking Reward Based Training – DogRelations NYC on June 9, 2017 at 6:19 pm
[…] When I walk around the city I see all these dogs trained with different approaches. Putting aside the sad fact that there are way too many pinch collars and choke chains around there is an encouraging number of dogs who are taught with reward based training… or let’s say: a version of reward based training. Here is where I still see that there is a missing link of understanding at least in the way I see reward based training really working. Read More > […]Pingback by Rethinking Reward Based Training | Lifestyle Okanagan Blog on June 12, 2017 at 4:25 pm