Clicker Training for Dogs

Basic Clicker Lessons

Variable Reinforcement Scheduling

Withholding reinforcement for a "known" behavior, a behavior that has been previously reinforced, will cause the dog to intensify that behavior. Think for a moment about teaching your dog the recall and using a treat as reinforcement for him coming to you. If he comes in at a brisk trot every time he is called and every time he comes he receives the same treat in the same fashion, will he ever learn to come at a faster pace? No.

Fixed reinforcement schedules, reinforcement for every occurrence of a given behavior, do not give enough incentive to work harder. In fact, he could actually start to come more slowly. For the same reason that winning a nickel every time he plays a slot machine will cause a gambler to lose enthusiasm for the game, the dog will lose his drive to work for a fixed reinforcement. A good training saying is, "Don't expect different results unless you change your training technique." Why would the dog ever come any faster when the results are always the same?

Never mind that he is always earning a small amount. The knowledge that his reinforcement will always be there, that it is always the same, that it is a "static" reinforcement, will eventually cause his enthusiasm for playing the recall game to decline. The secret behind getting dogs and gamblers to continue playing their games with enthusiasm lies in them never knowing for sure when the "big one" is coming. The occasional loss will cause him to try again, often with a variation. Periodically paying the dog off with a jackpot for extra effort and not paying at all for poor performances, will cause him to put forth that extra effort that earns jackpots. Withholding reinforcement will increase the likelihood of extra effort being offered.

For over 10 years, it has been my observation, that clicker trained dogs vary their behaviors in hopes of performing one that will cause their trainer to mark it with a clicker. This is the way I've always done it and it's worked very well for competition training from herding to obedience. The dog isn't supposed to figure out how to get the treat after the behavior is marked!

I admit to no longer being a pet trainer, that I have always trained for competition work and that titles weren't as important as high scores and national standings. These reasons are why I disagree so strongly with the content of the treatless click idea. To my knowledge, the trainers promoting treatless clicks, while well-known and respected, have never put working titles on dogs.

There are significant differences between shaping a dog to find a cookie in a box of toys (something that doesn't require shaping at all), and shaping a dog to consistently perform a scenting exercise in Utility classes. On one hand we have a behavior that requires no shaping, and on the other, a dog that must be able to repeat the behavior of finding a scented object, while also performing waits, stays, fronts, sits, carries, finishes and attention; all with strange judges standing nearby, in different surroundings, and with stressed-out handlers. Shaping to this level definately sets a boundary around pet trainers who are experimenting with dogs at seminars and in classes.

This is NO "put down" of pet class instructors, who have a difficult job and serve a wonderful purpose.

Causing a dog to offer new behaviors or variations of known behaviors isn't done through rewriting the clicker training rules, but by using the clicker training rules properly and to your advantage. Withholding the clicker (bridge) rather than using the clicker and then withholding the treat (primary reinforcer), is an excellent way to get new behaviors. A dog that believes he should have earned a click when he didn't, will offer the same variations as the dog who was clicked and not treated, so why lie to the dog with an "empty" clicker.

I have to believe these trainers have forgotten that the sound of the clicker isn't just a marker, but is a powerful emotional trigger for the dog. The behavior marker, according to Pavlov, not only triggered a physical response, but also emotional and physiological changes the dog was unconscience of.

Caution should be exercised in using the treatless click technique. A serious, high-drive dog, when he finds out his earned treat isn't coming, could bite his trainer or a child standing nearby! Call me an alarmist, but extingtion bursts don't just happen in the pool with whales, and they don't just happen after a behavior is well known...bites happen in less stressful situations where the dog hasn't been teased with a treatless click. Call them "frustration bursts."

Another concern is for anyone wanting to clicker train for competition. The behaviors recommended and used in the articles I read were all for behaviors that are not useful. I think it's a wonderful idea to experiment with useless behaviors (if there's such a thing) and even recommend it. When the behavior doesn't count - when nobody cares if the dog can perform them again, go ahead and experiment to build drive and to learn about your dog.

With all this said, the reality of it all is that withholding the treat and withholding the click are both forms of variable reinforcement, so go for it. I hope to get some feedback from anyone who does and either likes or doesn't like the result. I might rewrite the clicker training rules myself, if it works.

Clicking without treating has always been "around." Seminar presenters, including me, have played with it for years and, thankfully, competition trainers never took it seriously. Lately there are seminar presenters who have convinced some trainers that withholding the treat after clicking is an accepted form of variable reinforcement. After receiving feedback, wherein readers responded favorably to an article on the subject, I just HAD to address it on this web site. But you need to try it to see what it's all about. Here's one recommended experiment:

• Click your clicker and don't give the dog his primary reinforcement (treat or game).
• Watch the variations of the behavior offered.
• Watch for completely different behaviors offered.

The undisputed role of the clicker:

• The clicker is a secondary reinforcer (bridge) that connects the dog's behavior with what he really wants, the treat (primary reinforcer).
• Use a clicker to bridge a behavior with...nothing, and it's no longer a bridge. Intead, it becomes a cue to offer new behaviors.