The Revision Guide for Student Nurses (Part I)

Clicker Training


What is a clicker?
A clicker is a small plastic device which emits a loud clicking sound when the metal tongue inside it is pressed. It is used as part of the process of training dogs (or cats) and treating their behaviour problems.

How is it used?
The animal is trained to associate the sound of the click with the delivery of a reward (usually food). This uses the same principle of conditioning that Pavlov used to train dogs to salivate when they heard a bell. Once this association has been made the click can be used to indicate to the dog the exact bit of behaviour that the owner wants to reward. The click has become the reward.

How is this different from traditional methods?
Dogs and cats are best at understanding punishment or reward when it comes within just two seconds of them doing something. Beyond this time and learning is poor or non-existent. The clicker can indicate precisely what we want to reward, but after the click we don't have to give the actual reward within two seconds. This buys us some time. This enables us to train at a distance and to choose a very precise moment to reward. It does not depend upon the animal learning a command word before it learns the action. Even if the animal is not capable of learning the whole sequence that we want to teach, we can train it using the click to indicate each stage until we get what we want. This is called shaping.

How do we start?
First we need to create the association between the click and the reward. Choose a place to do the training where there are few distractions. Get a plastic pot and fill it with small bits of food that the dog (or cat) likes to eat. Muffle the click for the first few times by holding it tightly in your hand and behind your back. Click the clicker and then give a piece of food after between one and five seconds. Vary the amount of time between click and reward. Say and do nothing else (don't touch or talk to the animal). Repeat this 30 times at the first session and then later on the same day for a further 20 times. From now on the animal should only hear a click when you are about to give a reward. You are now ready to train the animal.

Use small bits of food (the size of your small fingernail), stay calm and quiet during training and choose a time when the dog is hungry. Never point the clicker at the dog; it isn't a remote control! Initially concentrate on one behaviour at each session of training. Stop at once if the clicker frightens your dog, and always stop the training before the dog gets bored.

First we need the dog (or cat) to do something that we can reward it for, and there are two ways of getting this.

  1. Wait. If you sit and just watch your dog for 2-3 minutes (with the food pot and clicker at your side) he or she will go through various sorts of behaviours such as sitting, lying down and lifting a paw. Choose one that you want to encourage the animal to do and click every time as it does that behaviour. Remember to give the reward too. Just throw this onto the floor near to the dog. Soon this behaviour will be repeated more and more quickly. If you have chosen sitting or lying down you will need to throw the reward away from the animal so it has to move to get it. Then it can come back and do it all over again.
  2. Lure. Example: If you get a piece of food in your hand you can lure the dog into a sit by moving the hand from just in front of the nose slowly backwards over the dogs head so it has to look upward and crane its neck backwards. This encourages a sit, and you can click just as the dog does this. Don't forget the reward.

    Either of these two methods is fine, it just depends upon how you want to start. However, the lure has to be stopped early on so that the dog can think for itself. Otherwise he or she will only ever respond to a lure.

Shaping the behaviour
This is the most important and powerful feature of clicker training. You don't have to wait until the animal performs a perfect sit before you click. You can start by clicking as the dog does anything that appears to be leading to a sit. Then you can start to be more specific. For example you can wait until his rump is closer to the floor before giving the click, or until he is sitting more quickly. You should use this process to perfect the behaviour by clicking and rewarding only the best 'sits' or 'paw shakes'. If the animal gives you the perfect 'sit' then you can give an extra good 'jackpot' reward.

Once the behaviour is perfect we need to make it more permanent. Every so often, wait for the dog to perform the behaviour twice before giving a click and reward. Then occasionally wait for three performances. Vary this gradually until even you are not sure whether you are going to give the reward or not! This will make the obedience more independent of whether a click and treat are given.

Now we need to give the behaviour a name. Say the word you want to use once, quietly, just before you know the animal is going to do the behaviour. As the behaviour is performed give the click and then the treat. Repeat this 60-80 times over a period of several days. You can add other types of command such as hand signals or other noises if you wish.

Once this response is perfect we need to sharpen the animal's understanding of when to perform the behaviour. We do this by only giving a click for behaviours that have been asked for. If the animal looks at you and then does the behaviour spontaneously, then there is no click or reward. If you have several behaviours that have been given names then you can ask for each in turn and only reward correct performances.

The sequence for training is therefore:

  1. Choose and establish the behaviour that you wantby using a lure or just waiting until the animal does the thing you what.
  2. Shape the behaviour by rewarding only the very best performances. Then vary the number of performances that are needed to get the reward.
  3. When you are happy that the behaviour is perfect you can add the command word. Give the behaviour a name. Only give a click and reward when the performance has been asked for.

What else to do?
All of the above has been done in a quiet, calm environment. Now you need to practise in different places where there are increasing distractions. The quality of the reward may need to be increased (nicer food or a larger amount). The more places you practise this, the better and more reliable the response will be.

Introduce other types of reward. A short game or a pat on the head can be just as successful a reward as a bit of food. Vary the type of reward and you will get better success. Some dogs do not respond well to a food reward. This may be because the food is not what they want or they are not hungry. A short game or a pat on the head could be substituted for food here as well.

The clicker is only used to train each behaviour. Once the training of the behaviour is complete, the use of the clicker should be faded out. Start by saying: 'Good' or 'Yes' just before the click, and then gradually phase out the click altogether.

More advanced uses
The clicker can be used to correct problems of training.
Once the association between the click and the reward has been created, you can use this as a reward for shaping behaviour that traditional training has not perfected. Examples include dogs that don't come all the way back to you or don't walk to heel off the lead. Just use shaping to get the response you want. Imagine the ideal response and then 'click and reward' the behaviour that gradually brings the dog closest to this. Soon you will have what you want, and you don't need to use a command word again until the dog has learned to perform perfectly.

The clicker can be used to help treat behaviour problems.
Overactive dogs: watch the dog and then 'click and reward' any behaviour that is leading to quiet calmness. Choose sitting, lying down or just stopping for a few moments. Gradually shape this by rewarding the dog as it is resting closer to a place where it normally rests. Then only click when it is resting there. Gradually shape the behaviour to produce the end result that you want. Try to picture in your mind a set of easy steps that you could reward, and that will be easy for the dog to understand. It is up to you whether you want to apply a command to this training.

Nervous dogs and dogs that won't play: sitting quietly, put a toy in the middle of the floor where the dog can see it. Click and throw a reward as the dog goes to sniff the toy. Do that a few times and then only click and reward when the dog licks or nudges the toy. Then click and reward for picking it up. Gradually shape a response where the dog will approach, pick up, carry and play with the toy.

Essentially you can use the clicker to train almost any behaviour that a dog can perform.

© Jon Bowen 2000