Excerpts from the Maxwell Award winning "Best Magazine Series" of 1995
By Gary Wilkes
NOTE: If you are serious about trying clicker training, the Click & Treat Training Kit is the best place to start. For more information click here!
Getting Started: Now that you have a clicker, you are probably wondering how it works, and what you can do with it. It seems like such a simple gadget, it is hard to believe that it can control animals as big as a whale or as small as a tropical fish, and almost everything in between. At zoos, marine parks, and now at dog training schools, elephants, whales, baboons and our faithful dogs are learning new behaviors pleasantly, thanks to this little tool.
How it Works : In essence, a clicker is an abbreviated way of saying "good boy." It identifies for the animal exactly which behavior "caused" a reinforcement. Behavioral psychologists, who first used this tool for shaping behavior, call it a secondary reinforcer. Primary reinforcers are the actual things that animals work for -- food, water, physical affection or fetching a ball. A secondary reinforcer is a signal that is associated with actual reinforcements until it takes on some of the qualities of those reinforcers. To understand this, merely pick up your dog's leash and ask if anyone would like to go for a walk. The leash is not the actual "walk", the primary reinforcer, but can act as a secondary reinforcer by triggering an almost identical reaction. The clicker can be associated with many actual reinforcers, such as food, affection and play. Once the clicker takes on these properties it can be used in a number of ways.
- · The clicker accurately identifies correct behavior. Because the clicker is faster than verbal praise, it is more precise. In the time it takes to say "good boy" an animal may perform the desired behavior and then move to an unwanted response, before the praise has time to register. In this scenario, the animal can't tell if the trainers liked the "sit" or the "jumping up on the trainer" that occurred a split second later.
- · The clicker can also work well from a distance. It is impractical to try and toss a treat at an animal's mouth at the exact moment that a desirable behavior occurs. The clicker bridges the gap from the instant the animal performs the correct response and the time it takes to actually deliver a treat. Marine mammal trainers actually call this a "bridging stimulus".
- · The clicker can take the place of the actual treats. Just as verbal praise has the ability to satisfy and animal in the absence of treats, the clicker can motivate and animal to work through "dry spells."
- · The clicker can take your dog's mind off the actual reinforcement. Some dogs are so food crazy that they cannot learn new behaviors in the presence of food reinforcers until the secondary reinforcer is established.
- · The clicker helps to define the end of the behavior. When teaching a dog to stay, for instance, the click indicates how long the animal must remain in one spot before a reinforcement is possible.
Charging up your clicker:
The first step in "powering up" your clicker is to associate it with positive reinforcers. If your dog already knows some obedience behaviors, merely replace your use of verbal praise with the clicker.
- Say "Sit"
- Fido sits
- Click and treat. (The sequence of "click then Treat" is important.)
* If your dog does not yet know any formal behaviors, simply click the clicker and give the dog a treat. Do this about 20-30 times until the dog visibly startles at the sound of the click.
Shaping your first behavior:
When offered a favorite snack, most dogs will sit expectantly and wait for the treat. After a few seconds of waiting, Fido is likely to get impatient and fidget in some way. He may turn his head, backup, speak or lift a paw. Wait for the first thing he offers, click and treat. (For this first session, the behavior you pick is not important.) If Fido turned his head a little bit for the first click, wait a few seconds, he'll do it again. Click and treat. (C&T)
Continue this process and watch how his behavior changes. If you continue to click and treat each time he moves his head, the behavior will become stronger. Now try waiting a second before you click. Try to get two "head turns" for the price of one treat.
Once you have a clearly definable behavior going (head turning), start saying "Turn your head", just before you think Fido is going to do it. If the behavior you shaped was lifting a paw, say "High five" just before you think he is about to perform the behavior.
Learning to use positive reinforcement to shape behaviors is a fun process. For your first project, learn to relax and see what behaviors you dog wants to offer. Your initial goal should be to simply watch how your dog's behavior changes, and see how the clicker helps you to identify correct responses.
NOTE: A proper treat is anything your dog will actively work to get. Treats should be small, bite sized and easy for the dog to swallow, whole. Soft treats can be swallowed quickly and are preferable to hard and crunchy treats which require considerable time to chew and swallow.
Commands and Signals:
A common training practice is to chant commands while attempting to push, pull or tug the dog into the desired position. For instance, the trainer says SITSITSIT while pushing the dog's rear end to the ground. With this method, the animal is forced to chose between paying attention to the trainer's words and learning the behavior. This practice often leads to one of two problems. Either the animal understands the behavior but must be told several times to do it, or the command is obeyed instantly, but in a sloppy fashion.
If teaching the command and the behavior simultaneously causes poor performance, the obvious solution is to teach them separately. Since a command without a behavior is pretty useless, a reasonable sequence would be to teach the behavior first, and then add the command. While this may sound logical, the implications of this concept may seem alien. If one does not talk to the animal first, how can the behavior be taught? If one trains silently, how does the command become connected to the behavior?
Before the strangeness of this concept causes you to reject it automatically, consider how you taught your first Click and Treat behavior...
Your first goal was to select a behavior such as "shake" or "head turning" and reinforce each occurrence. Within a few repetitions, your dog was offering the behavior consistently. It was after the animal started offering the behavior that I suggested you add the cue -- just before you thought the behavior was about to happen.
As you tried this method, you probably found that teaching the behavior without chanting the command was incredibly easy -- yet thinking about the concept somehow felt incredibly "wrong". The secret to overcoming this problem is simple -- stop thinking about it and just do it. OK? Here goes.
Down/stay: Start with the dog in a seated position.
1) Touch a treat to the dog's nose and slowly move it straight to the ground, to a point between her paws, under her chest. As her nose catches up to the treat, she should be seated, and slightly humped over. (click and treat.) Repeat this routine several times until she can easily follow your hand to the ground. If she attempts to stand, say "uh-uh", or "wrong" , in a neutral tone of voice, and return her to a seated position. Try it again.
2) After she is consistently following your hand to the ground, add a second stage to the behavior. Move your hand straight to the ground and wait for her to touch your hand with her nose. Once her nose touches your hand, begin to move the treat along the ground, toward you. The dog's nose will follow the treat. If she moves either of her front legs forward as she follows your hand, click and treat. Try to get her to stretch a little farther each time. It is unnecessary to get the whole behavior at one time. Be satisfied with steady, small progress. If she raises her rear end, say "wrong" and try it again.
3) As your dog stretches farther and farther, she will eventually walk herself into a "down" position. (click and treat.) As she begins to gain some confidence and speed, wait a few seconds, before you click and treat, so that she must hold the "down" position for several seconds. If she jumps up, say "wrong" and repeat the behavior. Try to get 10-20 repetitions of this simple behavior.NOTE: If the dog jumps up AFTER the click, give the treat anyway. Remember, the clicker also acts to end the behavior. After the click, the repetition is over and the dog does not have to "hold" the "Down".
4)Now that your dog can easily follow the treat to the ground, make things a little tougher. Touch the treat to her nose and then quickly hide it behind your back. If she moves to get the treat, say "wrong" and try it again. Tease her in this way several times. If you see her make any attempt to lie down, click and treat. If she continues to try to circle behind you for the treat, lead her to the ground a few more times. Once the behavior is again happening consistently, go back to several teases. Give her about 20 seconds on each "tease" to think about the target behavior. If you have sufficiently strengthened the behavior, you will be pleasantly surprised. After one of the teases, she is going to lie down on her own! If it takes more than three teases in a row to get her to lie down, go back to leading her to the ground a few times to refresh her memory. Reinforce any voluntary attempts at lying down, such as front paw movement, or head dips. Don't be surprised if this process takes a whole training session.
5)Once she is lying down each time you tease her, start saying the word "down", JUST BEFORE she does it. Click and treat.
6)Now stand up and try the sequence again. Touch a treat to her nose and say "down". If she lies down, click and treat. If she seems to forget the behavior, lead her to the ground a few times and try your "teasing" routine again. Your goal at this stage is to get her to lie down on command, even if you are standing.
7)Start varying the reinforcement. A simple way to vary the reinforcement is to ask the dog for longer or shorter "downs". Another approach is to give additional reinforcements for better performance. Since the clicker acts as an "end of behavior" signal, by withholding the click you can easily turn "down" into a "stay."
8)Start introducing the new command/behavior in a training session that includes other behaviors. If your dog also knows how to "speak", or "high five", or "turn your head", try to get the dog to "down" after a "high-five, and then ask for sit, come or any other behavior the dog knows.
9)Phase out the clicker once the behavior is correctly integrated into the animal's repertoire. The clicker is primarily a construction tool that should be used as needed to build behaviors. It is often impractical to click during a competition, or while actually working or performing with an animal. Switch to verbal praise for those types of situations.
These exercises can be the foundation for a great deal of fun for you and your dog. While pushing and shoving can develop control over behaviors, hands-on training is very limited. Using the click to signify "Yes, do that again" and reinforcing the behavior, you are limited only by your imagination and your dog's physical and mental abilities.
To expand your dog's repertoire, try this new behavior:
Identify an object , by name.
A) Hold an object (Car keys, wallet, newspaper, etc.) in front of Fluffy's nose
B) Click and treat if he investigates it.
C) repeat the process several times.
D) When Fluffy is repeatedly touching the object for treats, start naming it , just before he touches it.
E) Teach at least two more object names
F) Put all three objects on the ground and ask for them by name. Only reinforce correct identification.
- · Make sure that you only give a command once. Saying SITSITSIT will eventually necessitate always repeating yourself.
- · There is no need to bark a command or yell it. You dog is perfectly capable of hearing a potato chip hit soft carpet -- or a word, spoken in a normal tone. For sensitive dogs, a yelled command may actually be interpreted as scolding.
- · A dog can perceive a spoken command or hand signal with equal ease. Hand signals are best given so that the movement of your hand or arm is in silhouette. If you are backlighted or wearing clothes that match your skin shade, your dog will have trouble discerning the movement. To teach a hand signal, merely present it before you say the spoken command.
- · Once you start adding the cue, stop reinforcing unrequested behaviors. Your goal is to tell the dog that he must pay attention to what you say.
- · You can have many signals for a behavior, such as a hand signal, whistle or spoken word. Each command must be recognizable from every other command. Sit, vs. Down works fine. "SIT-DOWN" is ambiguous. GO may be confused with BEAU and NO.
Some basic rules for Click and Treat training:
- · If a behavior fails, drop your standards and review the behavior from the beginning. In the early stages of shaping a behavior the dog can become easily frustrated by repeated failure. Dropping your standards allow the animal to get back on the right path. Once the behavior is re-established, start gradually raising your standards again.
- · It is important to get "two-fers" (two behaviors in exchange for one reinforcement) early in the shaping process. You must teach the animal "if at first you don't succeed, try, try again. Animals that are experienced with this method of training are often willing to perform long routines in the absence of actual reinforcements.
- · Don't be afraid to shape behaviors just for the fun of it. Your skill and timing can only be improved though practice. Learning tricks can be a rewarding and stimulating experience for you and your dog.