When Clicking Goes Wrong
To follow up on yesterday’s post today I will describe some lulu screw ups caused by my disrespect of the power of the clicker. To quote Bob Bailey “What you click is what you get.” Now this sounds pretty simple but you really need to think about every single click you make; what the dog was doing, how good was your timing and what would you be communicating if you continued to click that same response over and over.
I used a clicker to shape Buzzy’s weave poles, as I described in my book Shaping Success. Remember this was long before I came up with the 2×2 method to train weaves. Think about shaping 3 poles, when would you click? Of course when
the dog finishes what he is doing or in other words as he exits the last pole. Now think about what I said in that paragraph, what would happen if you were to continually click the same spot over and over. You create over-importance or too much value in the dog’s mind for just that one piece of the behaviour chain. By the time I had finished training Buzzy to weave 12 poles he had inadvertently learned to drive hard and fast through 10 poles and slow down so he could stop and target the side of his face and scream at the top of his lungs at the last pole. He actually would come to a complete stop and wait for his click before he moved out of the poles. It took some doing to un-train that c.r.a.p.
This wasn’t my first screw up with a clicker, nor my last. Many years before, in the early ’90’s when teaching my first Border Collie “Stoni” to do a go-out (or send away) for obedience my goal was to train her to drive across the ring and touch the wall with her front paws. Before I knew it I had created a dog that did a spread eagle on the wall and wouldn’t get off!
I have seen many students create bizarre behaviours while using a clicker when attempting to shape their dog to back up. The problem is you just don’t need a tool that is that precise to teach most things your dog needs to learn.
The trouble when training with a clicker is that you don’t really know what you are teaching unless you carefully analyze each of your dog’s responses and even then by the time you have created the C.R.A.P, you have reinforced that crap so much you then need to come up with a new training plan to un-do the crap! Been there done that let me tell you! Only the dog really knows what the click is isolating no matter how much you wish you were both on the same training page. I have decided I am not a good enough dog trainer to use a clicker for everything and I will respect its power and only take it out once and a while (I try to do it at least once a month to keep my mechanics sharp).
Using a word rather than a click to mark responses while shaping gives you some leeway. Your dog has a range of behaviour to zone in on rather than a ultra thin slice of the response that the click is isolating. Remember what you think you are training is not always what the dog knows he is learning!
Now what I am about to tell you is not science (am very specific to mention this as all of my training is science based). However this next bit of information has been observationally learned, therefore its accuracy has not been validated or disproved in a laboratory (that I know of).
One of the many reason I am very specific about “when” I will chose to use a clicker in training and this is that I have observed using a clicker while training actually increases some dog’s desire to vocalize. This next part may sound anthropomorphic so go ahead and jump all over me, but to me it is almost as if the clicker becomes an auditorial judgment to the dog. He gets more and more frustrated while to trying to get us to produce the click and this frustration often results in vocalization. Producing the sound of the clicker seems to actually become more valuable to the dog than the food you are presenting after the click. I have seen many dogs exhibit this behaviour (all of them where high drive dogs btw).
To me it all came to a climax when we introduced dogs here to the behaviour of nose targeting an electronic target about 7 years ago (yes long before you could buy one:)). Classical conditioning taught these dogs that when they produce the beeping sound (which is the same idea as the sound of the clicker) with their nose on the target, a food reward would be produced. It didn’t take long before some dog actually got so intent on creating the beep that they wouldn’t stop to take the reward for producing it. The reward became and inconvenience to these dogs as they worked fervidly to create the sound. Classical conditioning had turned around and now the value of the beep was greater than the value of the food that created the value for the beep in the first place!
This is a dangerous place to be, and perhaps I will address it more in the future, but you always want your toys, treats, and interaction from you to trump any value the dog can earn from work. I will share with your at a later date how to maintain that balance.
Today I am grateful for the smile I got by posting that picture of a much younger Buzz in the weave poles. Buzz turned 13 years young while I was away last week. I told him when I got home “all rules are gone now buddy!” however his hearing has gotten so bad that he looked at me as if to say “whose drool on John looks funny?” Dogs crack me up.