BEHAVIOUR and TRAINING
[Behaviour] [Conditioned Reinforcer] [Shaping] [Sit] [Attention] [Down] [Come] [Biting][Jumping up] [Pulling on lead] [Begging] [Rewards] [Variable reinforcement] [Jackpot] [Problems] [Fearful] [Quiet] [Punishing] [Collars] [Positive]
To learn to train your dog you need to first understand how your dog learns to exhibit a behaviour, the chances are that you are already a good dog trainer without even realising it. For example, have you taught your dog to answer to the 'name' you call your dog? To 'pull' on the lead when going for a walk? and to jump-up on everybody? and to come when called for dinner? If so, then the chances are you are already using positive reinforcement to train your dog.
Every time your dog does one of these it is followed by something your dog finds rewarding! You use your dogs name to call him for a reward. You give your dog a jackpot reward for pulling you to the park. And 'muddy' paws ALWAYS get attention!.What you see as a "problem" with your dog’s behaviour, like the pulling on the lead, jumping up on you and your friends (even worse, with muddy paws), biting your hands, never coming when you call in the park, are not problems at all to your dog, only to you. The only reason that our dog behaves in this way is that these behaviours have been rewarded in the past. Yes, you have already trained your dog to behave this way. To change your dog's behaviour you have to stop rewarding your dog for these behaviours and teach your dog what behaviours you 'DO' want instead, sound simple? it is.
"Behaviour that is not rewarded will decrease in intensity, duration and/or frequency and in time disappear".
"Conversely, behaviour that is rewarded will increase in intensity, duration and/or frequency".
By selectively rewarding (reinforcing) or not rewarding (removing the reinforcer) your dog’s behaviour, you can change almost any behaviour good or bad.
"provided it is not serious or dangerous behaviour, if so, seek the professional help of a qualified canine behaviourist".
No need for shouting, slapping or punishing, no rolled up newspapers, sticks, or 'choke' chains, all these can destroy the relationship you have with your dog. Your dog will learn faster and better if you only use positive reinforcement teaching methods, these same methods are used to teach marine mammals, dolphins, sea-lions and killer whales, birds, cats, rats, and can in fact be used to teach any animal.
The methods used are based on the sound scientific principles of how all animals learn known as the laws of learning. In dog training we mainly use operant and classical conditioning. As you want to have a positive relationship with your dog, you need only focus upon the positive aspects of these laws, namely Positive Reinforcement.
To start you must find out what your dog finds inherently rewarding or motivating, this can be anything that your dog wants, food is an obvious one but others are toys, attention, petting, going for walks, playing, meeting other dog’s, to name a few. Anything that your dog finds rewarding may be used to reinforce wanted behaviour, but for a start, the easiest to use is food, not just ordinary dry kibble but something really special and tasty like chopped up (cooked) gravy beef, chicken or cooked meats, what ever your dog really likes.
The Conditioned Reinforcer (CR)
The first step in training is to prepare your dog for this teaching process, the tool you are going to use is called a "conditioned reinforcer" CR, you can have many different conditioned reinforcers to use with your dog. Some may by very precise like the "clicker" others not quite as good, like a short verbal "Yes" or "Bing" but all are useable and conditioned the same way. They must 'consistent' and delivered in a neutral manner.
The timing of the conditioned reinforcer is very important as it gives your dog information :
"Whatever behaviour you are doing at this moment (instant) in time is rewarding"
It bridges that behaviour with the reward, (so it is also known as a "bridge"), the CR tells your dog.. "You have just 'earned' a reward!" and lastly, it marks the end of a behaviour that is rewarding.
When you use your CR your dog may stop and look for the reward, that's OK, likewise your dog may carry on with the behaviour (self-rewarding) and put the reward in the bank, but you should always expect to pay your dog the reward for good behaviour. When you have the behaviour changed (shaped) as you want it, you can then attach a "cue" or command to that final behaviour. Then you can give the "cue" and get the behaviour you want from your dog and reward your dog.
We only use our CR for shaping behaviour.
Now, armed with some tasty treats (in very small pieces) you call your dog and give a "click" [ or say the word "Bing" or better still the word "Yes" if you don't have a clicker] follow this by giving your dog a treat. Repeat this twenty or thirty times, don’t forget you must allow for this extra food when feeding your dog or you will soon have an overweight dog. After a time you will see your dog 'alert' upon hearing the CR, and look for the reward, you now know you have "conditioned" the "click" or word to be a CR.
The "click" [ or word ] you use now becomes a "conditioned" reinforcer for behaviour, and can be used to communicate to your dog 'which' behaviour is good behaviour, as it is happening. One "click" and then reward your dog.
We must always follow the "click" [ or word] with a reward.
The CR is used to give information to you dog of "which behaviour" is rewarding, very quickly your dog should figure which behaviour it is, and start to 'offer' this behaviour to make you "click" (CR) and so earn rewards, as you have set the criteria for this behaviour, you only "click" when your dog's behaviour meets this criteria. Soon all (100%) of the 'offered' behaviour will meet this criteria. Your dog changes behaviour to meet your criteria and in so doing earn rewards. Now if you raise your criteria a very small amount so your dog is still earning rewards most of the time (at least 80%), your dog will again change the behaviour so that all behaviour again earns rewards (100%). By changing your criteria in these very small amounts you do not 'loose' the behaviour, or cause frustration in your dog, but cause the behaviour to change to meet your new criteria, this is called "shaping a behaviour".
You can change behaviour in several ways:
-
Change the duration, or the length of the behaviour.
-
Change the intensity or amplitude of a behaviour.
-
Change the frequency or speed of a behaviour.
-
Change the effect of external stimuli on the behaviour.
The most important aspect of shaping behaviour is changing the criteria by very small amounts. Large changes in criteria can produce spurious behavioural responses or frustration, a smooth change, a small step at a time can avoid this. Once the behaviour is shaped you can attach the 'cue' to the behaviour.
The first and easiest behaviour to teach is the "sit", however there is not much sense in saying "sit" to your dog until you have the sitting behaviour to attach the word "sit" to. By getting the behaviour first you can then attach the name to a known behaviour.
There are several ways of getting the sitting behaviour from your dog, wait until it happens, your dog decides to sit (operant behaviour) and "click" [ or "yes" ] and reward that behaviour.
Another way to get started is to "lure" the behaviour with a reward held high over the dog’s head until the sitting behaviour happens "click" as the dog is sitting, followed quickly by the reward. If you are using the second method, you should go to the first method as soon as you can, usually after three or four "lures" stop luring and allow operant behaviour, this will fade the need to lure.
Attaching the name to the behaviour is easy once you know the behaviour IS going to happen, you give the name immediately before or as the behaviour happens, then CR "click" and reward (reinforce) it.
Soon, when your dog is happily offering sits, you can use the name "sit" for this behaviour, and follow the behaviour with the reward or treat. This fades the "clicker" as you only use it to shape and mark the behaviour. Remember too that "sit" means to sit anywhere, "wherever you are put your bum on the ground".
The name "sit" is not a command but rather a request (or cue) for the behaviour (of putting your bum on the ground), if it has been rewarded in the past then there is a good chance it will be repeated and your dog will sit. Repeat this teaching your dog sitting in lots of different places to generalise the behaviour. Train with short training sessions of a few minutes each and always reward good behaviour with something your dog finds rewarding.
One of the most useful behaviours to teach your dog is to focus attention on you using the cue "watch me" to get your dog’s attention. This can be useful in many situations where you need to distract your dog from a potential reward, and focus on you instead. If your dog looks at you reward this behaviour.
With your dog sitting in front of you show your dog a "tasty" treat held high up at your eye level or out of reach to your dog, now, move the treat with our dog watching it out to your side with your arm extended. Your dog should continue to watch the reward, soon frustration will build up in your dog from not getting the reward and your dog will look at you in this unsure situation, as soon as your dog looks at you "click". You have now caught your dog’s own operant behaviour of deciding to look at you, and to look away from a reward to do so. Looking away from a reward has earned a reward "click", this is attention!
Repeat till your dog continuously looks at you when you extend your arm. Use different arms, different arm positions to generalise the behaviour. Attach the name "watch me" to this behaviour [ the same way we attached "sit" to the behaviour of sitting, once you know the behaviour is going to happen]. Start fading the use of the arm, use the cue "watch me" to get your dog's attention when distracted by any rewarding stimulus.
Having your dog look at you is a very important behaviour so start rewarding any eye contact that your dog may offer, a smile and a pat can work wonders, try not to ignore your dog , especially when your dog is not demanding attention from you. You should always acknowledge (reward) good behaviour from your dog.
Like all operant teaching you can wait till your dog goes into a down position, and then reward your dog’s own behaviour, or you can "lure" the behaviour a number of times first to get it started. Using a chair (for a large dog) lure the dog under the chair following a treat, when the dog goes into a down "click" followed by a reward. For a small dog you can sit on the floor and use your leg 'arched' to lure your dog under it and into a down position. The name for this behaviour is "down" and is attached to this behaviour in the same way as you attached "sit" to the sitting behaviour, in fact this method is used for attaching names to all behaviours you teach your dog. Repeat teaching the "down" in lots of different places. Always keep the 'tone' of your cues soft and friendly to your dog, never shout or get angry with your dog if the behaviour does not happen, Changing the stimulus (cue) may change the behaviour. If the behaviour doesn't happen it is a training issue. (usually the behaviour is not generalised) It is not your dog being 'disobedient', dogs do not think this way!
When you taught your dog to sit, where your dog sat was not important, but with the "come when called" behaviour you want your dog to sit in front of you. "Come" is the behaviour of sitting in this position. As before, you can wait till your dog does this behaviour "click" and reward, or you can "lure" your dog to sit in front of you (which you have probably taught anyway when teaching the "sit"). You attach the name "come" to this behaviour of sitting in front, as you reinforce this behaviour lean over and touch your dog’s collar so that later on you can attach your lead to it without startling your dog with strange body language. Only call your dog when you have your dog's attention and always reward your dog for coming to you.
Many dogs never learn how sensitive human skin is to their sharp teeth, so you have to teach your dog what is called "bite inhibition". This is telling your dog that any hard contact with your skin is painful (even if it is not). You start by letting your dog bite your hand and give a very high pitched "yelp" or "ouch" [do not frighten your dog] just loud enough to let the dog know it hurts, (or pretend it hurts, but don't jerk your hand away), this should cause your dog to back off and show some calming behaviours like head turning, shaking, licking, scratching or sitting. If your dog comes back and continues to bite hard, you give another high pitched "yelp" and walk away from your dog, you should go somewhere your dog cannot reach you. After five minutes or so you return and repeat as often as is necessary. Do not shout at or try to punish any "bad behaviour" do not do anything to your dog as it may only serve to reinforce that bad behaviour by giving attention, or worse still frighten your dog. Soon your dog will learn that the behaviour of biting is no longer rewarding as you refuse to play and give attention, this non-reward should stop the unwanted biting behaviour.
This is the most common unwanted behaviour you can teach your dog, when did you teach it? Every time you lifted our puppy towards our face, every time you gave attention when your dog jumped up to greet you, is how you reinforced this behaviour. To stop your dog jumping up all you have to do is consistently remove the reinforcement your dog gets for this behaviour, if your dog jumps up, just look up and away (no eye contact, attention is rewarding) totally ignore the behaviour, wait till your dog stops jumping up and either sits, or stands quietly in which case request a "sit" and instantly "click" and reward your dog for sitting. Repeat this till your arrival becomes a stimulus for sitting. Next repeat with friends approaching, if your dog jumps up, your friend turns and walks away for a few minutes or so, and then repeat till your dog sits as the friend approaches "click" and reward. Always reward good behaviour. Because you had 'trained' your dog to jump up to earn rewards in the past, your dog now has to be allowed to jump up to learn that jumping up is no longer rewarding.
With all your interactions you have with your dog nothing is so rewarding as taking you for your daily walk, pulling like a husky on a snow sleigh, or a bull on a plough, (some owners even fit their dogs into harnesses so that they can pull even harder), the reward for this behaviour must be powerful, it is, you let your dog pull you.
This is no different from any other behaviour, first remove the reward and the behaviour will after a while, start to diminish and eventually disappear. You must always be consistent with your response to a behaviour especially an unwanted one.
You must stop every time the lead goes tight and your dog starts to pull, wait till the lead goes slack and "click" and reward a loose lead. Yes, it is that simple, if you do not let your dog pull you, and reward your dog walking on a loose lead instead, soon you will have a well behaved dog walking on a loose lead, in the same way you attached names to all other wanted behaviours you can attach the name "Lets go" to the behaviour of walking on a loose lead, your dog can be anywhere provided the lead is kept loose. (MORE ON PULLING)
If you reward your dog with food scraps from the table when you are eating, you will very quickly teach this rewarding behaviour of begging at the table. To stop this behaviour you must first stop feeding your dog from the table. Then you ignore your dog 'begging' during dinner, and after dinner put whatever scraps you want to feed your dog into your dog’s dish in the kitchen, soon you will have your dog waiting by the dish in the kitchen, "click" and reward this wanted alternate behaviour.
Rewards can be anything your dog finds rewarding (remember it is the dog that decides what is a reward). Most dogs have favourite likes and dislikes, this is where you can experiment with different "safe" rewards. Chocolate and Onions are TOXIC to dogs, never allow your dog to eat them. For training rewards should be small and moist, dry foods can be used but can be distracting if your dog has to stop and chew them. Make a list of what your dog finds rewarding and then put them in order of preference, this can be used to vary the type and value of the reinforcement your dog earns.
some food examples: Cooked meats, Dried meats, Cornflakes, Kibble, Cheese, Sausage, Bread, Apple, Carrots, and of course your dog's dinner, etc.,
Other rewards can be: Attention, Praise, Toys, Playing, Going out for walks, Run in the park, Meeting other dogs and people, Sniffing, Petting, Scratching, In fact anything your dog likes.
All behaviour is in a state of constant change so after you have taught your dog a wanted behaviour, you need to reinforce that behaviour once in a while to keep it as you want it, otherwise it may start to change. It is a good idea to revise your training on a regular basis and re-shape behaviour that may have varied due to non-reinforcement since you trained it. As with most pet dog training it is better to always reward wanted behaviour, it does not have to be always food treats, a "good boy" scratch and tickle may very well do, but it always has to be a reward for your dog.
The use of different rewards to vary the reinforcement your dog earns is a powerful tool in increasing wanted behaviour.
Another very useful tool you can use when teaching your dog wanted behaviour is the random use of large jackpots, these are large rewards for a good example of a behaviour you want from your dog. Jackpots should be used sparingly to be most effective. When shaping behaviour a good response may still earn only one "click" but a larger/better reward. Jackpots can increase the rate of learning in your dog.
Many, if not most, of the behaviour problems your dog may have developed are all learned behaviour, this in effect means that if you are to apply positive teaching methods to these problems, you have to remove the reward your dog is receiving for unwanted behaviour, and more importantly teach your dog alternate wanted behaviour to replace the problem one.
If there are any serious problems with your dog’s behaviour it may require expert help and advice to overcome, talk to your Veterinarian, some problems may have other causes, either physical, medical or in rare cases a mental component that needs to be addressed first.
In some very stressful situations your dog may exhibit fearful or frightened behaviour, just like all other behaviour if this is rewarded with attention, petting, or your soft talking and soothing voice, this behaviour may increase. If you react with anger, loud voice, or in an aggressive manner the behaviour
will get worse, it may even develop into fearful aggression. You should always be aware of how you interact with your dog, pulling the lead tight, changing how you interact with our dog may also reinforce your dog’s concern in that stressful situation. While aware of any safety aspects in a situation, a happy attitude or "jolly" routine may well save the day. Feeding can reduce stress and fear if they are at a level your dog can cope with. You may need to seek expert help and advice if the fearful behaviour continues.
This may seem unusual but if you reward your dog for doing "nothing", lying quietly at your feet, in fact any calm quiet relaxing behaviour, these will increase and over time you will have calmed your dogs general behaviour which can make teaching your dog so much easier. You can also benefit from practicing calming behaviour around your dog. Hyperactive behaviour may be learned through 'hyperactive' playing with your puppy, if your puppy gets hyperactive stop playing. Hyperactivity may also be caused by reactions to some foods or food additives, seek advice from your veterinarian.
When teaching your dog you are also developing a relationship with your dog based on mutual respect, mutual trust, and understanding. If you apply any negative or punishing corrections you start to destroy that relationship and trust (the bond between you and you dog). You also inhibit your dog from trying to figure out what you want, from offering any new behaviour for fear of being wrong and earning a punishment instead, your dog never knows which will follow the behaviour, a reward or a punishment, this inhibits learning. The use of punishment to stop unwanted behaviour does nothing to address the cause of the behaviour, it does however suppress the behaviour, But ever increasing levels of punishment may be needed to maintain that suppression. Remove the cause of the unwanted behaviour not the symptoms. Suppressed behaviour can return at any time, this is known as spontaneous recovery of suppressed behaviour. Another risk when a behaviour is punished is that of 'displacement' behaviour, usually it is far worse than the behaviour being punished, and is unpredictable. Punishment can increase fear and cause aggression.
Punishment ---------- suppresses behaviour.
Non-reinforcement ---- extinguishes behaviour.
If you never use corrections or punishments when teaching your dog you get what is known as accelerated learning, you can then make rapid progress in training your dog and enhance your relationship with your dog at the same time.
Some unwanted behaviour may be self-reinforcing, to change such behaviour you have to look at what precedes the behaviour, change it, by first reducing the effect of this stimulus, and then teach an alternate wanted (rewarding) behaviour to follow it.
There are many different types of collars available for dogs, some however are unsuitable because they abuse the dog, they automatically correct or punish a dog’s normal behaviour, both good and bad alike, without any regard to what causes the behaviour. They can increase the danger of aggressive behaviour from a dog caused by the frustration and fear of always getting corrected, every time the collar gets used.
Another aspect to consider with some collars, in particular the " Electronic" and "Choke-Chain" is the increased danger of both physical and mental damage to your dog, not to mention the damage it can do to your relationship. There is no 'correct' way to ABUSE a dog.
The safest and ONLY collar to use for general training is the standard "flat" collar made from leather or webbing. The collar and lead should only be used as a SAFETY device,
COLLARS should not be used to train the dog.
There are some specialized collars that can be used in conjunction with the flat collar to help manage a particular problem (like pulling) while working on modifying that unwanted behaviour, some examples are the "Halti"™, or "Gentle leader"™ type head collars, just like a horses halter, they gently turn the head and should be used with a gentle smooth flowing movement. They allow you to 'manage' the problem, as you train your dog.
There are a number of lead types available, avoid the "chain" leads as these can damage your hands. Leather or cotton webbing leads are the best to use.
Positive reinforcement dog training
So, there it is, training your dog using positive reinforcement training methods that work. They are better for you and your dog and are more efficient. Rather than trying to make your dog do some behaviour, it allows you to teach your dog "to want to" do the behaviour you want from your dog. It’s simple, easy to use, and it’s fun, why would you want to train your dog any other way?
Happy training.
Robert Loftus
İRLoftus1997