Updated Version of Clicker Training 101
I wrote the original "Clicker Training 101" thread quite some time ago. Reviewing it now, a year or so later I feel as though there is more I'd like to add, so below you will find my updated version of clicker training 101. As with the original thread, please feel free to use it as a central place to discuss anything clicker related.
p.s i've really missed Chaz
Pre-Notes: Why I switched to clicker training from aversive based training...and why I'm not switching back
When I brought home my first dog, Duke, four years ago the first trainers I went to for advice and basic obedience all had my using aversives in some form or another. While Duke grew into an "obedient dog" his fear based aggression, which he had begun exhibiting at an early age, had made little improvement. I decided to try to find more information about force-free training after reading "Don't Shoot The Dog" by Karen Pryor. Three years later, after finding new trainers to work with, and having expanded my own knowledge base by attending the Karen Pryor Academy and Clicker Expo conferences I now view the behaviors he presents on a day to day basis and our relationship, much differently then the days I was trying to dominate him.
In my personal experience with traditional and balanced training, it felt as though I was constantly looking for a mistake to catch before it occurred to ensure my correction was properly timed. Over time stress avoidance behaviors began to pop up during training sessions such as sniffing, scratching, lip licking and looking away making pre-obedience trail training difficult. The bottom line is, when training with aversives, whether to diminish behavior, teach behavior or correct mistakes, you are in a constant battle with your dog, because your core training is based around the mentality of: Do it, or I'll make you. Don't do that, or I'll correct you.
In stark contrast, with clicker training your dog wants to learn, you are the most interesting thing at that moment so there is never a power struggle going on. Overtime with progressive training and layering in distractions your dog learns he or she will always win by offering behavior you want in a wide variety of contexts. In the actual realm of learning, force-free marker based training takes your dog from a passive observer to an active participant and it really shows. Dog's are willing participants and love to offer behavior. When it doesn't work your dog learns to slow down and try something else, and in essence is learning how to think through situations. Lastly it creates a dog who’s not afraid of trying things, and this is the biggest difference I saw in Duke's overall attitude years after giving up aversives in my training. The worst thing that happens when he is wrong is he doesn’t get a click- oh well try again, traditional trained dogs often go into a behavior void when first learning via shaping (discussed below) because until this point, trying things has never gotten them far.
On a last note Duke is now just shy of four years old. In the last three years he has come from reacting to almost any dog or male within a 30 foot radius to confidently offering a look at the dog behavior as we pass dogs on walks, even face to face. The other week a tethered dog charged us from it's lawn and came to a halt a foot away from Duke, hackles raised and barking at us. Duke looked at the dog and immediately looked at me while wagging his tail, and farfetched sounding or not, these are the real life difference I have seen with my dog.
Everything below is just my 0.02.
The Basics of Clicker Training
Marker based training works by first pairing the marker sound (be it a bell, whistle or click) with some type of reinforcement desired by the student. Once the marker has been established over several repetitions of this pairing the marker can the be used to mark behaviors the student performs that a trainer whishes to reinforce. Depending on a trainer's personal preferences markers can be used to mark behaviors that we wish to reinforce positively, or to mark behaviors we are about to apply aversives too, in hopes of diminishing it.
Ian Pavlov was the first to discover a novel sound or smell could be closely paired with food. His famous ringing a bell followed by delivering food was the corner stone of the field of science we know as classical conditioning. Ian discovered that after a short time of pairing the bell with food, the bell itself soon caused the dogs to drool- in essence the bell had become a predictor of food. The clicker works on the same principle. Alone it is a novel sound- BUT when paired closely with something your dog desires it becomes a powerful event marker, and in training can be used to communicate to your dog EXACATLY what he did correctly to earn a reinforcer. Furthermore it acts as a bridge between the time he performs the behavior and receives the reward.
The clicker is NOT magical; it works because of classical conditioning, the process of pairing something novel with something the dog wants. Overtime this makes the novel item/ sound/ smell/ sight a predictor of good things to come.
The clicker acts as an event marker, telling your dog exactly what he did right
The clicker acts as a bridge between the time the behavior is performed to the delivery of the treat.
Guidelines to using the clicker effectively
For new pet owners and dogs new to clicker training it is recommended to reinforce after every click. You certainly do not need to click every desirable response your dog offers but it is very easy to discourage dogs new to learning through experimentation ( in hopes of earning a click and reward) by not following through with a reward after your dog has earned a click.
During the early stages of training with a clicker it is recommended to keep food out of sight to prevent dogs from becoming overly focused on the presence of food.
Nuts and Bolts of the Clicker Training Process
Below is a simple three step process to using a marker effectively in your training.
1) Observe the behavior- Clicker training requires a keen eye, the two types of methods to build behavior in clicker training are capturing (waiting for the behavior to occur on it’s own and then clicking it) and shaping (building the behavior gradually step-by-step)
2) Mark the behavior- Think of clicking as taking a picture of the behavior. Your eye sight is the LCD display screen and you want to take a picture of the behavior in progress. Near the end is fine, but it is always better to click early then late.
3) Reward the behavior- The only guideline for this is make sure you keep the reward out of sight initially as the dog is still learning about the clicker and at this point will focus more on the reward. Second make sure you find something your dog enjoys- not something you think he enjoys. There is a big difference. Try out 6-7 different food items and observe if they are gobbled up instantly, sniffed and then eaten or passed by and rate them from 1-5 in terms of interest level. Do the same with 5 different toys (a tug toy, chase a Frisbee, chase a ball, etc)
An example : Teaching your dog to sit using capturing
Note: Tethering when using capturing is a good idea as it limits your dogs options as to what behaviors he may perform.
Step 1: Watch your dog for "clickable behavior"
The instant your dog begins to crouch and looks like he/she may sit get ready to click.
Step 2: Mark the behavior you wish to reinforce
As your dog begins to sit, click.
Step 3: Reinforce the behavior you wish to see repeated
If you are using a food reward, deliver it to your dog. If you are using play with a toy, unclip your dog from the tether and have a 30-45 second play session.
The Basics of Shaping Behavior
While there are many excellent books written on the in-depth process known as shaping, below is a simple outline of how the process works.
Shaping involves teaching a behavior without any force, food lures or cheerleading on your part. In essence the dog must learn to offer behavior and is similar to playing the hot/cold game for humans.
You start by defining what your end behavior will look like and then brainstorm how you can break it down into small easily accomplishable pieces.
Many dogs new to shaping may benefit from shaping on a novel prop item, such as a ball, a platform, a chair, etc.
In the beginning you will mark and reward your dog rapidly for offering any behavior towards the prop. This could be as simple as glancing at it.
As your dog begins to catch on your begin to raise your criteria, always keeping your rate of reinforcement (how often you mark and reward) high and keep sessions very short, this is especially important with brand new dogs to shaping. Over time as a dog becomes more savvy you can begin to introduce longer sessions, but a good guideline in the beginning is keep sessions to 30-40 seconds. You can have multiple sessions in one training session.
<3's March because it's Epilepsy Awareness Month
“Violence begins where knowledge ends.” –Abraham Lincoln