Building a balance of trust with your dog

Reinforcement can take many forms.
Reinforcement can take many forms.

 

I attended the WOOF Training and Behaviour Conference recently. It was organised by Chirag Patel of Domesticated Manners and it was a truly amazing conference. The highlight for me, and many, judging by their comments, was hearing Dr Susan Friedman speak. Dr Friedman is a professor at the University of Utah and is a pioneer in the field of Applied Behaviour Analysis in companion animals.

Dr Friedman summed up her philosophy in one simple sentence, which I’m paraphrasing – Control the environment, not the animal, so that the animal makes better decisions and then reinforce those choices to make them more likely in the future.

To a dog, there is no difference to it jumping up or sitting as long as the behaviour is reinforcing. The only judgement put on the behaviour is by ourselves. We think of jumping on people as a “bad” behaviour and sitting as a “good” behaviour. To the dog, however, if both or either of those behaviours are reinforced i.e. they work for the dog, the dog will continue to do them.

One of the points Dr Friedman made was that the ability to make our own choices is very reinforcing. Imagine you had the most amazing life you could think of. You had all the money you ever needed, met the most interesting people, did the most fun and exciting hobbies and activities, ate most delicious and healthy food, had the best friends and a fantastic house with every gadget you could think of. Now, imagine that same life, but someone told you when to eat, when to do your hobbies, when to watch a movie and which one, when to see your friends and for how long and when to go on holiday and to where. The same, great life, might somehow now seem less satisfying. This is because we wouldn’t be in charge of our own destiny or decisions. Dogs need to make their own choices, at least some of the time. It empowers them and makes them happy. It’s our job as their owners to make sure the choices they make are appropriate and safe and to then reinforce those choices.

We can do this by managing the environment. If your dog tends to run off when you let him off the leash or bolt out of the door when it’s open, it’s because there is something reinforcing about doing this, otherwise, he wouldn’t do it. Now, if we put a leash on the dog when we open the door, we’ve now manged the dog’s access to the environment and her ability to self reinforce. We can now wait until the dog does something more appropriate, such as stop pulling or sit, and then reinforce this e.g. with food or access to the outside. This way, she then learns that dashing out the door no longer works, but sitting does, so she’ll offer this instead. I’ll give some more specific examples of this in upcoming blogs.

When the dog does make the right choice, we need to reinforce this, heavily and frequently. Think more in terms of reinforcement than rewards. When we use the term “reward” I think there is a tendency to view it as payment or compensation for the behaviour which has just been performed. What we are actually doing, is using the food “reward” or game/ball etc, to make the behaviour more likely to happen in the next time. When the dog offers the same behaviour again, we then have a further opportunity to reinforce this. This makes behaviours very solid. What happens quite often, and I include myself in this until recently, is the tendency to try to reduce the amount of reinforcement too quickly. What we are trying to do is build up a balance in a behavioural bank account.

Every time we reinforce a behaviour, we are making a small deposit into the account. When our dog performs a behaviour we want, and we don’t reinforce it, we are making a withdrawal. Unfortunately, the withdrawals are bigger than the deposits and if we don’t have an adequate balance in the first place, we can very quickly become overdrawn. To further illustrate, imagine you win the Euromillions lottery of £140m ( about Euro 150m or $200m). With that money, you make a £2m investment and you lose your money. Although it’s £2m, you still have £138m to fall back on so you probably won’t be too upset (which is pretty ridiculous given the amount of money we are talking about). Now, say it’s two days before payday and you only have £30 left to last you the two days and you need to pay a bill and buy food. If you lose £20 of that, the impact is so much bigger than losing £2m from the previous example, because the £20 has so much more relative value.

To apply this analogy to dog training, every time we reinforce a behaviour we are making a small deposit into the  “trust” account and make that behaviour more likely to reoccur. If we have a huge reinforcement history, built over months and thousands of reinforcements, we will build a huge balance. Our dog trusts us as there is classical conditioning occurring at the same time. What then happens is there comes a time where, for whatever reason, we are unable to reinforce the dog. We are cashing in some of that balance, but because it is so big, the trust between us and out dog isn’t affected (the lottery example). If we don’t have that same reinforcement history, we try to make a much larger withdrawal relative to the balance (the £20 loss before payday example) and the dog’s trust in us is affected.

Make loads of deposits into that account with your dog. He will do more of the behaviours you want him to do, less of the behaviours you don’t want him to do and your relationship will be stronger because he will trust you more. Until next time.

John McGuigan, Glasgow Dog Trainer and Behaviour Consultant

Advertisements

8 thoughts on “Building a balance of trust with your dog

  1. Nice post. Chirag is a very exciting trainer and some great stuff of his on YouTube. His classical conditioning video to teach ‘drop’ is inspired.

    How does Dr Friedman’s philosophy apply to reinforcement schedules when training? Any advice?

    I have trained my dog off cats using BAT. The idea was to allow her at a distance to safely make an alternative and more appropriate decision. i.e. turn away from the cat and look at me (reward is a ball toss and game of fetch). We have managed to work on this so we can now walk past them with little fuss. I used some ‘stooge’ cats who are stray but are cared for by people in the community. They live outside in their little cat houses on the edge of a car park, so I had the space and ability to work under threshold at a safe distance and gradually reduced the distance over time as we progressed.

    Environmental management and setting the dog up for success is key to building a history of safety so your dog will trust you. It requires making judgement calls about other people and dogs a lot of the time and also recognising that some environments may simply be too stimulating for your dog to handle. For my dog, a walk in the woods is anything but relaxing. Her high prey drive and subsequent arousal levels induced by squirrels, hot trails of deer, other loose and strange dogs equally aroused darting through trees makes her very difficult to control. No matter how hard one has worked on training, it can all go to pot if you have allowed your dog to become over stimulated. Therefore no walks in the woods for us. Know your dog and keep your expectations of her realistic.

  2. With regards to reinforcement schedules and information I’m learning from Ken Ramirez, we shouldn’t be overly concerned with quickly switching our dogs on to variable reinforcement rates, rather than varying the types of reinforcement. To build cast iron, reliable behaviours, a history or heavy and frequent reinforcement is required before we think about varying the rate.

  3. Thanks John – I’ve always felt that to be true in some regards but couldn’t really find much evidence out there to support my feeling on it. I do recall Jean Donaldson mentioning this in The Culture Clash though – there was strong evidence to support that continual reinforcement rates = higher rates of compliance though ( I think in marine mammals though) . Anyway, thanks!

  4. Lovely blog!

    Regarding variable reinforcement, I don’t really use it. I just continually crank criteria up (is any behavior ever perfect?) and reinforce, in some way, every response that is at or above criteria level. While I know we put behaviors on a VRS to avoid extinction, I’ve yet to have a behavior I want (except for the lower level responses which become sub-par as the dog’s skill increases) go extinct!

    In other words, while “sit” would technically be considered a behavior maintained on a VRS because a sloppy, lackluster performance no longer earns “pay,” gorgeous, fluent, fast, perfect sits at any distance in a distracting environment will be reinforced for a lifetime on a CRS.

    1. So if I understand this correctly, what you are saying is that raising the criteria for a behaviour creates the VRS in itself?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s