You Can Cross Over,
But You Can't Cross Back,
by Donna Duford
(Reprinted from the Latham Letter, Fall 2000)
The term crossover has been used to describe a person or dog who has "crossed over" from traditional, command-based training to positive reinforcement-based clicker training. It is often challenging to cross over. Dogs who are used to command-based training will have trouble understanding that they can offer behaviors freely. They are used to waiting to be told what to do before being allowed to do anything, and, in fact, have often learned that offering behavior results in some sort of aversive consequence like a leash correction or verbal reprimand. People who cross over face the challenges of breaking old patterns, learning a new training system and facing peer pressure. They often also look at training and indeed their dog's behavior in a completely different way. But once you have crossed over, you are there for good.
When I crossed over, I did it gradually. I started with my own dogs, experimenting along the way. When it was going well for me and I realized that this stuff could actually work, I introduced clicker training to some of my individual clients. Not all, mind you, as I didn't think it was the thing for everyone. I loved what I was doing with my own dogs and was committed to making my crossover complete - no corrections, yet I was not convinced that these same techniques would work with inexperienced people. I thought that people had to be somewhat dog or training savvy to understand clicker training and do it properly.
At first I found it simple to use one method with my own dogs and teach a different one to my clients. After all, they were inexperienced; it was reasonable to modify techniques to accommodate their lack of skill. Success is highly reinforcing, and the great results I got with my experienced clients motivated me to try clicker training with my novice clients. I started using it in private lessons only, as the thought of using it in group classes was daunting. I just couldn't fathom inexperienced people learning this technique without the benefit of individual instruction.
What I discovered when teaching "non-dog" people clicker training was that it was no harder than teaching any other technique. In fact, inexperienced people often learned it more easily than their more experienced counterparts, most likely because they were not laden with the confusion and frustration of relearning. When I realized this, I had reached The Point of No Return. There was no turning back. I gave all of my students clickers and taught everyone - experienced and inexperienced, individuals and groups - to train their dogs with a marker signal and positive reinforcement. I did some experimenting along the way, but I never looked back.
I have known for years that I have no desire to go back, but I did not realize until recently that I can't. This issue became clear to me when two of my colleagues and I recently started a pilot study with the Pryor Foundation on the difference in the speed ofacquisition of learning a new behavior between traditional training and clicker training.
Each of us was to train a dog, one with command-based training, and the other with clicker training. We decided on a 10-minute session for each dog. To make sure we ruled out as many variables as possible, we decided to use shaping exclusively with half the dogs and physical manipulation with the other half. The clicker-trained dogs were rewarded with food, while the other dogs were given verbal praise and petting. The manipulation we used was mild. We either slid the front legs out to guide the dog down, applied pressure on the shoulders with one hand, or used downward pressure in the collar. All three of us had used physical force of varying degrees in the past and felt comfortable using gentle pressure for the purposes of this study.
The clicker training sessions were fine. Some of the dogs caught on quickly and started to offer "down" on cue. Others didn't get quite as far in the process. Regardless, they were relaxed and engaged with the trainer. The first thing we observed with the force-trained dogs was their stress. We noticed it early and watched it escalate quickly. We started with mostly exuberant, friendly dogs who within minutes were subdued, upset and socially disengaged. And none of them figured out what we wanted. In fact, a couple of the dogs who had followed a hand motion to the ground at the start of the session would no longer respond to it after a few trials with force. It did not surprise us that the clicker-trained dogs learned the behavior faster. We were overwhelmed, though, by the amount of stress we saw in the force-trained dogs.
The techniques we used were gentle compared to some of the ones we had used in the past, yet once we got started we found ourselves agonizing over them. Though our agreed training period was 10 minutes, we each wanted to stop training by five minutes into our session. We couldn't wait for the timer to go off. We each tried to avoid using physical manipulation in a different way. One trainer contorted her body, without realizing it, in an effort to prompt the dog. Another endlessly petted and caressed her dog, prolonging the time between trials. I tried to convince the other two to shorten the time criteria. We became unsure of ourselves and stressed. Though we were once confident with and proficient at traditional techniques, we found it very uncomfortable to go back to them.
Why can't we go back, or become what many in the dog training field call "balanced" or "hybrid" trainers? Largely, it is a matter of conscience. If you use aversive techniques because you believe they are in the best interest of the dog, you can justify their use. If, however, you know you can get the same results without them, it is unethical to do otherwise. Once you learn a more effective, more humane method, you can't train any other way. Scientifically speaking, we don't cross back because clicker training is reinforcing. It works. Not only does it work, but it is fun and the dogs love it. This is all very reinforcing to someone who loves dogs.
One of the products of clicker training is better observational skills. To execute a well-timed click we must pay close attention. When we do, we are attuned not only to our dogs' behaviors, but to their emotional states as well. We see the signs of stress earlier and more clearly. It is more apparent to us when a dog is confused or frustrated. Because we can relate to these feelings we become more empathic as trainers. Our increased empathy prevents us from crossing back.
So what happens to our study? To continue the research we will have to find non-clicker trainers to do the force training. Our anxiety and discomfort disqualify us from that part of the study. We know now that we can't cross back. We can't even fake it.