Wednesday, January 5, 2011

Teaching a Behavior

Teaching any behavior starts with the same basic process.  If you understand and master this process, you can teach any behavior you like.  The basic steps are outlined below.

First, you need to be able to make the behavior happen.  Once you can make the behavior happen consistently, you can add verbal cues and hand signals before the behavior.  Then you need to generalize the behavior to new areas and proof it in the face of increasing distraction.

When teaching any behavior, a trainer has four basic options to make the behavior happen.  We can lure, shape, capture or mold a behavior. 
  • In the case of luring, we use a treat to guide the dog into the desired behavior so we can reward it. 
  • In shaping, we wait and reward steps towards the desired behavior, slowly raising our criterion along the way.  Think of playing the game “hot and cold” as a child, you aren’t guiding the dog into the desired behavior but rather waiting for offered moves in the desired direction.  Shaping is most easily accomplished with a clicker to mark the behavior.  (Some people combine aspects of shaping and luring to acquire behaviors more quickly.)
  • Capturing can be used to train behaviors that a dog naturally engages in.  The handler waits until the dog performs the behavior and then marks and rewards the behavior as it naturally happens. 
  • Finally, molding involves actually placing the dog in the desired position through physical manipulation (like pulling up on his collar and pushing down on his butt to make him sit). 

I much prefer to lure, shape or capture a behavior rather than mold it.  For one thing, in molding the dog isn’t actually performing the behavior himself.  Instead, you are forcing the behavior on him.  Molding can be a valuable tool when a trained dog blows off a known command but is not the best way to initially teach most behaviors.  So for teaching new behaviors, I strongly recommend luring, shaping or capturing.  I frequently use luring in my group classes because it is generally the easiest training technique for handlers to pick up and become proficient at quickly.  For handlers and dogs that are comfortable using a clicker, I highly recommend shaping and, when appropriate, capturing.  We’ll talk more about clicker training in a later note.

Whichever technique I am using, when teaching a new behavior I don’t start using my verbal cue until I can consistently predict that the behavior will happen.  If I repeat my cue (“sit” or “down”) without the behavior happening, the dog will simply learn that the cue word has no meaning.  Alternately, if I repeat the word multiple times before he performs the behavior, he may learn that “sit, sit, SIT!” is the cue rather than “sit”.  For these reasons, I like to make sure that I can predict the behavior will happen with 90% confidence (9 times out of 10) before I add the verbal cue for the behavior.  Once I am consistently predicting the desired behavior, I add my verbal cue about one second before the behavior is initiated. 

In the case of luring a sit, I initially place a treat in front of the puppy’s nose and push it in a line towards the back of his head.  This will cause his nose to rock up and back and his butt will sink towards the ground.  When his butt hits the ground, I mark (either by saying, “Yes!” or with a click) and then I reward him with a treat while he is sitting.  Timing is everything.  You want to reward your dog while he is performing the correct behavior – not after he has gotten up to do something else.  (Clicker trainers have a little more leeway since the click effectively marks the moment in time that the dog is being rewarded for.) 

Your dog may not perform the entire behavior all at once.  If that’s the case, you will need to break the behavior down into smaller steps that you can reward along the way.  In the case of sit, you may initially reward the head rocking back and the butt tucking down just an inch or two.  Once the pup is comfortable, reward the butt tucking three or four inches, adding distance an inch or two at a time until the dog’s butt finally hits the floor.

Once the puppy is consistently sitting when I put the treat in front of him and start to move my hand back, I will start to fade the treat out as a lure and simply make it a reward.  At this point, I’ll do several repetitions luring with the treat before switching and making the same luring movement with an empty hand.  When the dog performs the desired behavior, I will mark it and reward with a treat from my treat pouch.  During this process, I want the dog to learn to follow the empty hand in anticipation of an as-yet-unseen reward.  It is really important at this point to reward 100% of the time when the dog follows an empty hand.  You don’t want him to learn that he has to see a treat in order to get a treat.

Once the dog is following an empty hand consistently, I introduce the use of the verbal cue “sit” just before I present the hand and start to move it.  The verbal cue needs to come before any movement on my part so the pup learns that the word predicts that I will cue the behavior with my hand and body next.  If I add the verbal cue while moving my body, the dog will continue to rely on the physical cues rather than just the word.  The first time the puppy responds to the verbal cue, before I use the hand/body cues, I mark the behavior and give him a jackpot reward (4 or 5 small treats fed in a row).  The jackpot reward makes a powerful impression on the puppy and increases the likelihood that he will offer the behavior (in this case sitting when he hears the word “sit” rather than waiting for my hand to move) again. 

I like to wait until the pup is following an empty hand before adding the verbal cue, some people add the verbal cue before luring with an empty hand.  I like my technique because the verbal cue is not associated with following a food lure but you can choose whichever technique works best for you and your dog.

Once he consistently follows the verbal cue, you can start to train a formal hand signal.  Simply use your new formal hand signal about a second before the verbal cue.  Jackpot reward the first couple of times that the dog responds to the hand signal before you say the verbal cue.  You can decide whether you want to train the formal hand signal or verbal cue for the behavior first but most people do the verbal first. 

While I used sit – probably the easiest behavior to teach – in this example, the progression holds true for teaching all sorts of behaviors.  Figure out how to lure, capture or shape the behavior consistently before putting it on the cue of your choice.  Once it is on cue, you can start to gradually generalize it to new environments and proof it in the presence of distractions. 

Remember that this progression is not a straight climb up a ladder.  There will be days when your dog is less motivated or more easily distracted and you need to lower your expectations so he can succeed.  If you move to an area with higher distractions, make sure you start back at the beginning and practice your baby steps before moving up through the process.  Realize that you will need to retrain each behavior several times in different locations before your dog generalizes the behavior to new areas.  If your dog is consistently failing to perform, evaluate whether you are moving too fast or have increased the distraction level too quickly and adjust your approach so he can succeed.  He needs to practice success and desirable behaviors, not failure, even if that means moving away from other dogs in a group setting or going back to baby steps on a behavior that he “knows” at home but cannot perform consistently in a more distracting environment.

No comments:

Post a Comment