Clicker Training "Do's"
For the uninitiated, I figured a few "Do's" and "Don'ts" of clicker training might be helpful. We'll cover a few of the "do's" today, stay tuned tomorrow for some clicker training "don'ts"!
DO a trial run for sound sensitive dogs. Before you begin using the clicker in your training, click your clicker when your dog is relaxing in the house. Make sure that you are not clicking near the dog's ears. Do the dog's ears perk up, or does she tilt her head as if to say, "What's that?" If so, you're likely ok to begin using the clicker. If she appears fearful, running away, shivering, barking, with her ear or tail tucked back, go onto the next "do":
DO consider the use of other markers. While research has proven the clicker increases the acquisition rate of new behaviors, a clicker is by no means the only marker. My dogs respond to a number of markers - a tongue "click," a visual marker (pointing gesture), a verbal marker, and I should probably get to work on a tactile marker as well. Deaf dogs may benefit from having both a tactile and visual marker. In my training, the clicker is ALWAYS followed by food. My other markers predict more reinforcement variety and are more frequently paired with life rewards.
DO remember that the clicker is a teaching tool. Once the behavior has met your criteria, it's time to transition onto a verbal marker and introduce reinforcement variety and variability.
DO bring the clicker back out to "clean up" old behaviors which have deteriorated in some aspect of fluency. While I certainly don't click Mokie for sitting anymore, if we need to tidy that behavior up because she has been displaying a low latency, I will pull the clicker out, clean the behavior up, and once again abandon the clicker in favor of other markers once the behavior meets my criteria regularly.
DO feed your dog even if you have a poorly timed click. We all have great timing days, mediocre timing days, and days when we should just put the clicker away. If you have a bad click, for instance you are clicking your dog for sitting but his feet lift off the ground when going to greet someone, feed him anyway. It's not his fault you made a mistake! It is more important to keep the ratio of click to treat static at 1:1 - an occasional bad timed click will not generally undo dozens or hundreds of well-timed clicks and reinforcements. If you find you are always clicking the wrong behavior, talk to your trainer about doing some practice timing exercises without your dog to improve your skill.
DO make sure that the click and treat are two separate events. The click should predict the treat and therefore, must come before the treat. The treat should not be moving, visible, or otherwise distracting the dog until after the click. If you are fumbling with, reaching toward, or feeding your dog food as you are clicking, you are "blocking" the click - the dog is much more likely focusing on the food than on what he is doing to make the click go off. It should be click THEN treat not treat while you click. My students are required to have a "home base" - where will your treat hand be while you are clicking? Inside of your treat bag, fumbling with treats, or reaching toward your dog to feed him are not appropriate answers. Your treat hand should remain totally still until after you click.
DO work in short sessions. I'd much rather see someone do 10 one minute to 90 second training sessions in a day than one four hour long training session per day. It's better to "leave the dog wanting more" than it is to work on an exercise until your pooch is bored to tears and disengaging.
DO set your dog up for success by raising your criteria carefully and appropriately. This may involve: controlling distractions in the environment, "going back to kindergarten" by reducing your criteria temporarily if your dog is frustrated or confused, and raising your criteria when your dog is successful 80% of the time at the current performance level.
DO consider the use of a training journal. This can be a place where you create training plans, track your rate of reinforcement, your dog's response to various reinforcers, etc.
DO teach a concept first, then label it. For instance, if I am capturing "sit," I will click and toss a treat each time the dog sits. I will not say a word until I have a nice rhythm of "behavior, click, treat," ie., the dog is sitting, getting clicked, retrieving the tossed reinforcement and immediately returning and offering the "sit" position. Only then will I began adding a cue, first as the dog is doing the behavior and then beginning to back the cue up earlier in the sequence.
DO try to deliver your reinforcement within 1 second of the treat, particularly in "green" or novice clicker dogs.
DO have a purpose in mind for each reinforcement. "What is the job of this treat?" If I am working on duration for a stationary behavior like "go settle on a mat," I may want to feed the dog in position - reinforcing him while he is lying down on the mat. If I am first shaping the settle behavior, I will be tossing my treat away from the dog to reset him for another repetition. In settling on a mat, you can use treat placement to elicit more interaction with the mat - if I toss my treat so that the mat is between myself and the dog after retrieving the treat, chances are he will reorient to me after eating, thus setting me up for another chance to capture "four paws on the mat."
DO use food delivery to create excitement. My Chow mix Mokie thinks that a treat which she can a) catch in her mouth or b) chase down/hunt for in grass, etc. is at least twice as interesting as the exact same treat fed from my hand directly into her mouth.
I could come up with a bunch of other "do's" but will save them for another day. What are your favorite "clicker do's" that I may not have mentioned? Please leave them in the comments!