Dogs do something because they like the result.  As Murray Sidman says in his book "Coercion and its fallout", "consequences control behaviour".  To apply a human analogy, we would soon stop putting money in a drinks dispenser if the drink never came out!  However, if you successfully received your drink on one occasion, you may use the machine again.

In doggy terms, if a dog sits by the dining table when the family are eating their meal and the result is that he is given food from the family's plate, he will almost certainly try this strategy again. In fact, next time, he will probably sit next to the dining table, drooling, in anticipation of the food that he believes is coming his way. If the dog is never successful (never receives any food/attention) he will give up - eventually! (dogs can be very hopeful!)

Training dogs is simply about managing the "consequences" of the dog's actions. To do this you need to understand the concept of the reward/success. There are rewards that we can provide for the dog, including, attention, fuss, titbits, toys etc). But there are also "self-rewarding" activities and you need to understand these in order to manage the consquences appropriately. A dog with a high chase instinct (such as a border collie) may chase cars, squirrels, birds etc.  The mere act of chasing is rewarding. It stimulates adrenalin and a rush of "well-being" that the dog then wants to repeat.  Understanding what your dog finds rewarding is vital for training success (think digging, chewing etc).  For example, a dog that jumps up at visitors is being rewarded.  He wants the attention and fuss.  Even if the visitor pushes the dog away, this can be perceived as attention (and thus reward). By jumping up, the dog has successfully gained the visitor's attention. He has been rewarded and he will do it again.  If the dog had not been successful, or had achieved success (got the visitor's attention) by employing a different and incompatible strategy (sitting and wagging his tail for example), he would be less likely to jump up in the future. This is why we reward behaviour which is incompatible with the offending behaviour. A dog cannot be jumping up at the same time that he is sitting and wagging his tail.

The consequence of an undesirable behaviour should be that the dog does not achieve success. For example, if the dog is barking to come inside [and you don't want him to bark], you must not let him in whilst he is barking. He must not be successful in gaining your attention (don't shout at him to be quiet) and must definitely not be successful in getting you to open the door.  By waiting until the dog is quiet and then opening the door, the dog is less likely to bark in the future --- because it didn't work last time.

So, where does the clicker come in?

The clicker is a way of "marking" the bit of the behaviour that is good. It forms a "bridge" between the actual behaviour and the delivery of the reward. The value of the reward is important. The reward needs to be something the dog wants.  If the dog doesn't really want what is on offer, he will not find it rewarding and the value of the "consequence" will be lost.

Dogs are often greedy creatures, and this is one of the reasons that food rewards can work well. However, sometimes other rewards work better. Play, including chasing, running, tugging can be more potent and you need to work out what your dog likes best. The better the reward, the more chance that any competing distractions will be ignored.  For example, would your dog prefer to eat a piece of biscuit or chase a squirrel?  I know which my dogs would prefer...

The clicker is not a reward. The sound means nothing to the dog initially.  It has to be conditioned. This is what we mean by a conditioned reinforcer.  The sound of the click (or any other sound) becomes associated with the reward and becomes a reinforcer. The reward must be delivered every time the dog hears the noise for an expectation to be created. Precise delivery of the click lets the dog know what you are rewarding.  He will start to wonder what he did to make the click happen (and the reward follow).  He will start to try out different behaviours to see what would make you click.  Everytime you click & reward a behaviour, he has been successful and you increase the likelihood that this behaviour will be repeated.

Why not use punishment?

You can - but I would not advocate harsh punishment for anything less than a potentially life threatening scenario. But what is punishment?  Punishment could be considered to be anything which makes it less likely that the dog will repeat the behaviour.  If a barking dog is sprayed with a carefully timed/aimed jet of water (and doesn't like it) he may be less likely to bark in that situation again. (As a strategy this is probably only likely to be successful in few circumstances as no consideration is given to the reason that the dog is barking in the first place...)

Punishment can be used in a much more subtle way.  For example, if you are playing with a puppy and he nips you, a good reaction can be to yelp and stop the game (put away the toy and walk off). The puppy will find this "punishing" and will eventually learn that his behaviour causes the game to stop. He will be less likely to do this again. Stopping/removing something the dog wants can be as punishing as delivering something the dog doesn't want.

Shaping

The main value of clicker training is the ease with which we can communicate what we want.  To create a behaviour, we shape it. This means that the behaviour is built up from small successes.  Shaping can be "free" (with no help) or "lured" (with a few hints). With free shaping the trainer needs to click and reward for a behaviour which is a "step in the right direction".  For example, to teach a dog to turn around in a circle, you would click for a step to the side or a turn of the head.  Once you had repeated this a few times and the dog was reliably offering a movement to the side you would withold the click and wait for the dog to do something else. With a bit of patience, the dog usually offers something else. It may be the wrong thing (just ignore this) but, if it is a further step in the right direction you click and reward again.  By continuing in this way it is possible to build up (shape) a whole turn/circle. 

The same can be achieved (a bit quicker) by encouraging the dog to take a movement in the desired direction. Offering food to the side of the dog usually does the trick. The exercise still needs to be built up but can get to the end point much quicker than without a lure.  Some people argue that the exercise is much more thoroughly taught using the "free shaping" method. However, I think it is quicker and less frustrating to lure. It is also essential to lure a behaviour that the dog is not likely to attempt naturally. Try it and see what you think!?

Proofing

This is about making sure that the exercise is thoroughly understood.  Sometimes we think the dog has understood but actually has some confusion.  For example, if I leave my dog on one side of a jump/hurdle and stand on the other side, I could teach my dog to come to me by jumping the hurdle.  Have I taught my dog to jump a hurdle on command? Not really, he will probably only do it if he is heading towards me. I have to teach it again with the dog going away from me. What if I set the dog up out of line with the jump? Would he still jump the hurdle or would he come to me around the hurdle? This shows how exercises need to be proofed.  We need to be sure that the dog understands the exercise completely. To proof an exercise you need to encourage the dog to "go wrong" so that he learns what is "not right".  For example, teaching a dog to stay involves clicking and rewarding when he stays put. Although this can be built up to longer stays, are we sure that the exercise is thoroughly understood?  To proof the exercise we encourage the dog to move (perhaps by dropping a ball or titbit on the floor, clapping our hands etc) then communicating that he has been unsuccessful.  Repeat the exercise until he stays put and ignores the deliberate distractions - click and [big] reward!

[Home] [About Us] [Clicker Training] [Ellie's Diary] [Canine Epilepsy] [Obedience] [And there's more] [Links] [Disclaimer] [Site Map]

© Copyright 2004 all rights reserved