Reblogged from Mirkka Koivusalo, KPA CTP's blog: Reward-Based Dog Training as Science Explains it
Most dogs have some annoying habits. We humans want them to stop as quickly as possible.
Some humans have a little handheld tool they use to stop behaviours they don’t like: a squirt bottle. A sharp spray of water in the dog’s face should stop jumping/chewing/nipping/barking pretty efficiently, right? Plus the gadget is cheap, easy to get and shouldn’t really hurt the dog, right? Well, not in my opinion.
I am a professional dog trainer, and I have, in my time, used a squirt bottle as punishment. It worked. It made my then-puppy learn to leave the curtain, rug and a corner of the coffee table alone. Would I use the squirt bottle as a training technique with my clients or own dogs today, with the current knowledge that I have on dog behaviour? No, I would not.
Today, I see that the reason for my success with my puppy over a decade ago was simply that I got really lucky. But I believe more in science-based methods of behaviour modification than I do in luck. What I know now is that squirt bottle training isn’t as precise or useful as other methods of training. In fact, it can be harmful. But that’s true of most kinds of punishment.
When you’re talking behaviour technical definition of punishment is that it’s anything that makes a behaviour less likely to occur again. But in order to be effective, there are certain laws to be followed. One of my favourite trainers, Steve White, has beautifully outlined “eight rules of punishment”, which all have to come true for the punishment to be effective. Let’s examine the spray bottle technique from the point of view of these rules.
Rule #1. The punishment must be something the animal dislikes and something the animal does not expect.
Does your dog dislike it? I can think of several dogs, who would just LOVE a spray of water in their faces! Often they would be the Labrador types who just want to experience water in all of its different forms. So the squirt in the face wouldn’t be much of punishment for them.
Is it something the dog doesn’t expect? This is so important, and that’s why I got so lucky with the squirt bottle the first few times I used it. I was stalking my puppy and managed to surprise her just at the moment when she was grabbing our curtains with her teeth. But to obtain the element of surprise I had to stalk her behind a couch or a kitchen counter and have the squirt bottle somewhere handy so that I could grab it without her noticing it. Pretty exhausting and challenging work! Plus the prerequisite for this to work was that she had no previous associations with the bottle. The first few times it worked fine, but when she started to experiment with other items with her teeth, she quickly caught onto me and only backed off if she sensed the squirt bottle despite my detective-like manoeuvres.
Rule #2. The punishment must suppress behaviour. If something is being used for punishment, but it does not suppress behaviour, it’s ineffective and often just plain abuse.
Yes, even if it is “only” a spray of water, if you have to use it a more than twice to try and suppress the same behaviour, clearly something is off. And even a squirt bottle can easily become abusive. We cannot choose what a dog finds punishing or reinforcing; maybe the water isn’t what works for that particular dog, or something else in the environment is reinforcing the behaviour and negating the effect of the (possibly) unpleasant water spray. This will easily happen with some very self-reinforcing behaviours like barking or jumping on people. Fair enough, the water is a little annoying, but oh boy isn’t jumping on visitors fun! Definitely worth a few squirts in the face! The challenge with us humans is that we still keep trying to use the intended punishment because pressing that lever on the bottle and seeing ANY kind of a reaction in the dog is so rewarding for us.
Rule #3. The punishment must be of the perfect intensity. Too much and there will be negative fallout. You’ll end up hurting your relationship with the animal and losing more than just that behaviour. Too little and the punishment will only serve to desensitise the animal and build resistance.
What happens with too strong a punishment? Many dogs are not as hard-tempered as my puppy was. If you squirt a sensitive dog, there are many caveats to this technique. The dog may become fearful of water or any kind of bottles. Bath times and any kind of medical procedures involving a bottle will become battles in the future. Or if the punishment occurred when your dog was greeting a specific visitor, you may get a fearful response from your dog every time your friend comes over. Not much fun for any party involved.
How about if you play it “safe” and just squirt the dog a little, just to startle them a little bit? Well, it doesn’t solve the problem behaviour if it’s just a little annoying to the dog. And chewing/jumping/barking or whatever it is that the dog is doing is most likely worth it despite the annoyance. Also, we are not able to experience the spray of water the way the dog does, therefore gauging the intensity is very difficult for us. Otherwise said: You can try to test it by spraying yourself in the face, but you still don’t know what it’d feel like because you don’t have a face like a dog.
Rule #4. The punishment must happen immediately after the behaviour it is to be associated with. Otherwise, a clear enough association between the wrong behaviour and the punishment will not be made.
Super-fast reflexes are required for using punishments! If we’re just a second or two late the dog may associate the punishment with something completely different. Such as looking out of the window at the neighbour coming home. Now in the dog’s mind the neighbour produces the evil spray in the face and the next time he sees the neighbour outside, a frantic barking sequence may follow. Personally I have noticed that the older I get, the slower my reaction time is. At least I know if I deliver a reward too late I will not be scaring the living daylights out of my dog. Very few of us are fluent enough punishers to make them work well.
Rule #5. The punishment must be associated with the behaviour, but not with the trainer. Otherwise, the trainer becomes part of the punishment and the animal starts fearing and disliking the trainer.
This rule comes back to the element of surprise. If you didn’t do your pink panther-like detective stalking well enough and your dog sees you and your arm approaching with the horrid bottle that produces the horrid spray, what will happen? The punishment will become equal with you and the dog will become fearful of you. Fearfulness often manifests itself as aggressive behaviours, so this may end up being a vicious circle. Even if your dog doesn’t become fearful, what most likely will happen is that your dog will still indulge in the unwanted behaviour when you are not around.
Rule #6. The punishment must happen every time the behaviour occurs. If punishment does not happen every time the behaviour occurs, the behaviour gets put on a variable schedule of reinforcement. Depending on the behaviour and how often the punishment actually occurs, the animal could decide that performing the behaviour was worth the risk of getting punished.
Exactly. This is the exact reason why it is tough to extinguish unwanted behaviours. To not reinforce a behaviour we should ignore it every single time. Like demand barking. An annoying habit, no doubt about it! It is such a self-reinforcing behaviour that it seems to wind the dogs up, and if we acknowledge the dog one time out of ten, BINGO, the dog just got a huge jackpot. They just learned that the tenth time works in getting a reaction out of us, so they need to try extra hard ten times to get what they want.
Same principle is true for punishment that happens every now and again. If we are not delivering the consequence, the spray bottle, consistently every single time when the behaviour occurs, for example we are not at home or are in a different room, the dog learns that they still occasionally get to do the reinforcing behaviour. Since it is punished only randomly, it is always worth trying it.
Rule #7. There must be an alternative for the animal.
This is the rule in positive reinforcement training: train the dog what TO do instead. Here is an example from my own life. I am Finnish and when I moved to Canada, I had no clue about having to tip people for their services. I mean, I had already paid them right? This what I was used to. When I landed at Toronto International Airport, I took a cab to my lodgings. When we arrived, I duly paid the cab driver the exact amount in the meter. He stormed out of the cab and literally dumped my two suitcases in the dirty slush in the street and drove off. I was in tears, horrified and shocked by his rude behaviour and had no clue why I had been punished this way. Nobody had ever taught me that in Canada I have to tip the cab driver.
This is pretty much how our dogs feel, they live in a human society which has our rules, not theirs. Therefore, as fair parents, it is our task to teach them what to do. If we punish a dog for jumping without teaching them that we people would appreciate a nice sit for a greeting, we are not being fair to them at all.
Rule #8. Punishment must never be used to the extent that punishment outweighs positive reinforcement (from the animal’s perspective, not yours!).
When we grab the water bottle and decide to spray that annoying behaviour out of our dog we easily end up sliding down a slippery slope. To successfully teach a dog to do any behaviour, the amount of rewards should always exceed the amount of punishments. If the “good” thing to do, such as leaving the couch cushion alone, is not adequately reinforced, it is therefore less rewarding than the “bad” thing (the pleasurable, reinforcing activity of chewing the cushion into pieces). If that’s the case, we are not really training the dog to do what we want. We are just falling into the cracks of failing to comply with the rules already listed above.
As tempting as it may be to us humans, the use of punishment is very challenging in training. I figured out a long time ago that I’m not a good enough trainer to be able to use it, even if it’s something relatively mild, like squirt bottle. Until I run into a training dilemma that I cannot solve with the combination of management and reinforcement of acceptable behaviours, I’m not going to feel the need to think about spray bottles again. But I don’t think that day will ever come.