The text below is the synopsis of David’s presentation on the negative impacts of training dogs using an electric shock collar originally given to the Scottish Parliamentary Cross Party Group for Animal Welfare at the Scottish Parliament on 13th September 2006.
My background is in police dog training – I am employed by Cumbria Constabulary as an instructor, although the views I express here represent only my own.
I have been a police dog handler for over 25 years and an instructor for thirteen. I also have a postgraduate diploma in Companion Animal Behaviour Counselling from Southampton University, which, along with Edinburgh University, is acknowledged as a world leader in that field. As a member of the APBC I take behaviour cases on veterinary referral only and run veterinary referred puppy and dog training classes. I currently work with a four-year-old police German Shepherd bitch.
I have never felt the need to use an electric shock collar, in fact, as I am sure you know, electric collars are not allowed to be used in the training of police dogs.
This is an emotive issue and there are strong opinions on both sides, so I’m going to dodge it. I’m going to present the facts, and let you decide.
The principle of an electric shock collar is to cause the dog pain when it performs a behaviour the trainer designates as undesirable. In pure learning theory terms this is one of the four ways in which trainers can affect operant behaviour: – positive punishment, and, technically, it will work.
Positive punishment involves applying an aversive stimulus contingent upon the undesired behaviour, which the dog then strives to avoid.
This is the first consideration of which the trainer must be aware: when the dog is trying to avoid any behaviour, she is in a negative state of mind, rather than being motivated to achieve a particular result in a positive state of mind. So, if I’m training a recall using punishment, the dog is trying to avoid not coming back, as opposed to actively wanting to come back.
For positive punishment to be effective, it must fulfil certain criteria, namely:
- It must be contiguous. It must happen at exactly the same time as the behaviour for maximum effect. The less in time with the behaviour, the less likely the dog is to associate the stimulus with it.
- It must be consistent. It must happen every time the behaviour happens, or the association between the stimulus and the behaviour is less likely.
- It must be intense enough for the dog to consider it sufficiently aversive to prefer not to perform the behaviour
As a trainer, I need to consider how these conditions apply to electric shock collars, because if I don’t get it right, the behaviour won’t stop and all I am doing is applying pain to the dog unjustifiably.
It would seem to be easy to be able to apply the remote controlled shock at the time of the behaviour, but is it?
Although we have a particular behaviour in mind when we press the button, we have to be sure that the dog has the same behaviour in mind. Our definition of the behaviour and the dog’s must be the same.
For example, if we wish to punish ‘bicycle chasing’, we should press the shock button when the dog looks at the cyclist and is thinking ‘I’m going to chase that’. Split second timing is necessary, or we’ll have punished another behaviour, for example, ‘Looking back at the owner to see if I can get away with it’.
It can be extremely difficult, even for professional dog trainers, to observe a dog’s body language and thereby assess her state of mind, to come to a decision and then to act on that decision, by applying the shock. Reactions have to be extremely quick. My experience is that very few dog owners are capable of it.
Contiguity is not as easy as it appears.
We have to be able to observe and punish the behaviour every time the dog performs it and, with good organisation and preparation, this may be achievable. Dog owners rarely train their dogs using good organisation and preparation, but often in response to evolving events.
As with contiguity, making sure our definitions match is essential. We have to punish the same behaviour consistently. If our definition is ‘failing to come back when called’, the dog’s definition may be, ‘searching for rabbits’ the first time, ‘chasing cats’ the second time and ‘peeing on lampposts’ the third time.
Which brings us onto Stimulus Generalisation. When a dog receives a punishment, she scans the environment to assess what caused it. All the stimuli the dog thinks are relevant are associated with the punishment. As well as the behaviour the dog is performing, these may include the location and the trainer, or rabbits, cats and lampposts. If the dog attaches most significance to the wrong stimulus (lampposts?) not only will the avoidance to the desired stimulus is reduced, but stress may be felt every time the ‘wrong’ stimulus is present.
How is the dog to know that the trainer wants her to come back? She has to dismiss everything else in the environment, other than ‘not coming back’. I’ll explain why they often do come back, shortly.
Consistency is not as easy as it appears.
The correct degree of intensity is the most difficult to achieve. The first point is that the type of punishment must be viewed from the dog’s perspective. Not only is punishment breed specific (some breeds are more touch sensitive than others), it is also individual specific (not all individual’s sensitivities are similar). My explosives search dog Sandy was extremely pain tolerant, but social exclusion was devastating for her.
The second point is that the level of punishment is also breed, individual and mood specific. The dog’s emotional state will dictate the level at which she finds a stimulus aversive.
A dog that is sniffing a lamppost will require less aversion to make her stop than a dog in full flight chasing a rabbit. When dogs are in a highly charged emotional state, particularly during the predatory hunting sequence, they close down extraneous senses, for example the ability to hear. Endorphins released as a consequence of predatory behaviour also increase pain thresholds.
Incremental aversive stimuli don’t work. (Azrin, Holz and Hake, 1963)
So we have to come in at the highest level possible in order to be successful. But that level is dictated by the breed, the individual dog and the emotional state that the individual dog is in when we press the button.
Intensity is guesswork – and every time might be different. Once the dog learns to ‘work through’ the pain, it becomes ineffective and has to be increased – but she’s already learned that she can work through it once.
Pain as Punishment
I next need to look at the consequences of the type of punishment an electric shock collar dispenses – pain. Sudden pain causes a fear response in dogs (and other mammals). Dogs have three genetically prepared strategies for dealing with fear. They are Hiding, Running Away and Fighting.
When dogs return to their owners on being shocked, they are trying to hide from the pain, looking for comfort. When shocked, dogs come back because they are frightened of the pain.
Usually they are running back to the owner, but not necessarily.
Pain often stimulates an aggressive response, particularly if the dog thinks the source of the pain is a possible combatant.
That is why it is not a good idea to shock dogs that are actually chasing things like cyclists. If they perceive that the pain is from the cyclist, they may well respond with aggression towards them – which is a worse scenario than the behaviour we are trying to change! There is the additional problem that if there is no obvious target, the trainer may be viewed as the combatant. ANY aggression will only escalate on the application of pain. If you are in ‘fight’ mode and you feel pain, you respond by inflicting more pain.
The worst effect of using electric shock collars is that they take no account of the state of mind of the dog that is being trained.
They are marketed ‘to stop dogs barking’. When I checked my books, I found eighteen different reasons for dogs to bark.
There will be some overlap, some examples are: alarm, hunting, play, fear, frustration, attention seeking, territorial defence, anxiety. Can you tell the difference between territorial defence and anxiety barking?
If anyone can buy a shock collar, there is no diagnosis, by a competent person, of the reason the dog is barking in the first place. If the dog barks in excitement, it might be worth considering punishment as a behavioural modulator, although I still wouldn’t use pain. But many dogs bark because they are lonely, anxious or frightened. Imagine the mental state of a dog that is already sufficiently in emotional deficit to cause barking, which is then electrically shocked.
Finally, there is no assessment of the overall behaviour of the dog. Many ‘training’ problems are complex and can be seen as almost a cry for help from the dog.
Chasing other animals is often a sign that the dog is failing to fulfil its inherited drives. Dogs inherit different levels of instinctive behaviour they need to exhibit. Simply punishing the behaviour does not provide fulfilment and the problem manifests itself in a different way, for example, self-mutilation, tail chasing. These may be less troublesome to the trainer, but are indicative of a reduction in the welfare of the dog.
The Forces in an Electric Collar
I do not pretend to be an expert in electricity, so I asked someone who is, because I wanted to know what affected the force applied to the dog’s neck.
Bruce Englefield is also a member of the APBC with an MSc in animal behaviour. He is an international sheepdog trialist and also an ex Sound Engineer and Technical Director for the BBC. He knows about electricity.
This gets a bit technical, but stay with me and it will become clear.
Bruce told me that that Ohms Law states that Voltage = Current X Resistance.
In properly manufactured Electric Collars, the Voltage always stays the same. That means that if the Resistance goes down, the Current goes up.
The Power – the force applied – (measured in Watts) is the Current times the Voltage.
The Resistance is in the dog – the bit of dog between the two electrodes. If that Resistance to the Current goes down then the Power goes up.
The Resistance will vary depending upon the electrodes’ degree of contact with the dog’s skin, how tight the collar is, how hairy she is, the amount of body fat she has, and, crucially, how wet she is, because water is a good conductor. If that was difficult to understand, the bottom line is that not only will the Power of the shock vary with each dog, it is also going to be extremely difficult to judge whether it will be the same as the last time. Is the contact as good? Is there any hair in the way? Have the electrodes moved? Is she wetter, or drier?
We already know that we should apply the highest degree of aversion possible, to be effective. So I set the collar to what I think is the maximum pain this dog can stand and let her go. What if she runs through a puddle before I press the button? How much more pain would it administer than I intended? I certainly don’t know.
I don’t want to use an electric shock collar for changing behaviour that involves emotional deficit, fear or aggression, because pain will exacerbate the problems, causing deeper welfare issues for the dog.
I don’t want to use an electric shock collar for problems involving instinctive behaviour (chasing, searching) unless the owner provides for the dog’s needs by allowing it to express that behaviour in an alternative way, because that too will cause a decrease in welfare.
I don’t want to use an electric shock collar for any behaviour the dog might associate with people, dogs, or other animals, in case the pain prompts a fearful or aggressive response.
I don’t want to use an electric shock collar for any behaviour I want to train the dog in a positive frame of mind, like coming back to the owner.
I don’t want to use an electric shock collar for any situation in which I can’t determine the sensitivity of the individual dog at the time I press the button.
I don’t want the general public to use an electric shock collar because I don’t believe they have the skills necessary to assess the behaviour of the dog, the intensity of pain required, or the ability to time the pain effectively, which will result in repeated applications of electric shock with no discernable improvement in behaviour.
Since I started to train dogs methods have become increasingly less punitive, as our society becomes more civilised. What is certain is that, eventually, these products will be banned. The only question is: will that be sooner or later? Perhaps Scotland could take the lead.
David appears as an expert legal witness in canine behaviour and advised the Scottish Parliamentary Cross Party Group for Animal Welfare on the negative implications of training dogs with the use of electric shock collars, on behalf of the Scottish Society for the Prevention of Cruelty to Animals. He presented the APBC’s response to the Scottish Government and Welsh Assembly Government consultations into the use of electronic training devices.
David’s particular interest is in inherited predatory motor patterns and the lengths to which pets will go to find a way to express them, usually despite their owners’ best efforts.