Have We Reached Success in Training?

How can one measure whether the training done with a dog has been successful? What parameters should one have in order to evaluate if success has been reached?

While the questions above are valid, we also must keep in mind that behavior and changes in behavior are quite complex which requires that we take a different approach to measuring the end results.

Behavior and measure of success are complex for many reasons; here are some of the main ones.

1. If the motivation for performing a given behavior, which was present when the dog was learning it, fails to be present when the dog must perform the behavior, it will go into extinction. What I mean is that it will not happen. There must ALWAYS be some reinforcer for the animal (and this includes all animals - not just dogs BTW). Good news is that the reinforcer should not appear on a one-to-one ratio (also known as continuous schedule of reinforcement) but it should be on a thin (or at least a variable schedule) of reinforcement. In other words, the dog will get reinforced on occasion, which in essence makes the behavior stronger, i.e.: less likely to go into extinction which is, of course, what we want. Now, in the real world there might be more than one motivator for behavior, so again, we are moving away from the notion of … dog sits, dog gets a click and a treat… as we might do when teaching dogs new stuff.

2. The environment is now substantially different to where the (initial) behavior was taught. Part of reaching success in training is to have the dog perform the behavior where we want the behavior to take place. This is called proofing. If I only work with the dog in the garage, I cannot expect the dog to perform at all or at least with consistency when I am out and about with my dog. This is a big problem when working with clients. They fail to understand that in order for their dog to perform, they must continue to practice where it matters. As I am here thinking of this example (funny because I don’t play tennis), but think about a tennis player that always trains on a rubberized, artificial court. Now when she goes to a tournament and the court happens to be clay she will quickly realize that the ball has a different bounce or that she is always late to reach the ball because the ball falls short due to the different terrain. In order for this tennis player to have a chance at success, she must practice under those conditions and adapt to the new environment. Same for our dogs :) with the difference that even a slight variation between the learned environment and the new environment can throw our dog for a loop. These are just but a few examples that might derail good learning.

 

Now, back to my client’s and their expectations for success. It is also part of my job to exactly learn where the dog is engaging in behaviors that are problematic for the owner – dogs never have a problem when they act as dogs - it is always us! Or in the case of aggression, etc. behaviors that can prove dangerous so that I can resolve them in that context. When it comes to teaching new stuff, I also need to know where would the owner want the dog to exercise the behavior at hand… In their backyard? The kitchen? Out and about? Or all three?

I think it is fair for clients to want to know if their dog’s behavior will improve since they are spending considerable amount of resources in the form of payment and time invested in the training of their pooch. So for this reason, the conversation of… is their dog not engaging in “x” or “y” behavior as a result of training or performing with more consistency the target behavior as a result of training, needs to take place.

As such, another conversation that I also have with my clients is to make them absolutely aware that behavior is not a “straight” line but more of a fluctuating experience. Once again, the same can be said about us: consistency eludes us and we have the bigger brain. :)

By doing my job correctly, their dog performs. But only as long as they continue to practice where it matters. They also must keep in mind that their dog is not a robot but also subjected to influences in its environment and as such, not always consistent in responding. And this my friends, is part of adaptation, which we all need in order to thrive.

Does Punishment Work? Part 2

As promised, the conversation as to if punishment works continues. Before I dive into the explanation and exploration of this topic, I want to make my position clear: I DO NOT ENCOURAGE OWNERS TO PUNISH THEIR DOGS. INSTEAD I ENCOURAGE PEOPLE TO LEARN HOW TO USE EFFECTIVE AND HUMANE METHODS TO GET TO THE SAME RESULTS.

I am writing about this important topic because I realize that people are punishing their dogs. What’s even more heartbreaking is that most people, and a large number of professionals, do not understand how to effectively apply punishment which also makes it even more cruel for the animal *(&^^)%$#.
I strongly suggest that if you are going to use punishment with your dog, you invest the time and energy to learn how to do so! And as you will see it is not that simple… The rendition below is based on the Science of Animal Learning. It is not speculation, wishful thinking, and bleeding-heart thinking. Instead, it is based on many, many studies and experimentations on animals and their learning. Some of these experiments were so unethical that they are not allowed to take place in the name of science anymore!

Conditions of Effective & more humane (?) - You decide Punishment:

Based on Dr. Pamela Read book: Excel-erated Learning

• Punishment should be a learning tool and not a means of revenge. If your dog is not learning as a result of using fear, pain or intimidation then it is not really punishment it is callous revenge.
• Punishment MUST be initially SEVERE. Let me say that again, PUNISHMENT MUST BE INITIALLY SEVERE. Yes, I know, all those shock related products advocate for starting “small” and increasing the shock or whatever mechanism is being used to inflict pain, discomfort etc. as needed. But what these manufactures do not understand (or fail to report - as the case might be) is that this is not how punishment works.

The rationale behind this is that punishment should be strong enough to create an impression on the dog. It hurts so much, scared so much that the dog chooses not to engage in “x” behavior for fear of being hurt again, etc.

In other words, the behavior we choose to modify has been suppressed.

Now, here is the rub: Very few individuals (yes, unfortunately they are quite a few cruel people out there :() will not have the stomach to really use severe punishment. And, even if the person really is trying to help the dog by applying this rule of learning, we are left with another practical and ethical conundrum: How strong is strong enough? How can we measure this?

If less severe tactics are employed, in the name of being humane, the problem is that the behavior will not be suppressed long-term.

Another interesting twist to this dilemma is the fact that some animals will adapt to pain… so here we have to crank the pain outlet even more. Ouch! Real ouch!

Thirdly, when an owner produces a mild punisher, the response as discussed above will be lukewarm but it might be present, which in turns reinforces the owner to continue to punish the dog. What a sad life for both of them…

The presentation of punishment should be timely. Any delay in this presentation will be ineffective.
Consequences must be presented immediately and as a consequence for behavior in order for a dog to learn to link these two events together. Again, failing on this will not only constitute lack of learning, (suppression of behavior) but frankly how unfair can this be! The dog has no idea why he is being mistreated. So in choosing a punisher, we must consider the immediacy of the delivery. This is at the crux of fair/correct use of punishment.

The punishment must happen EVERY-SINGLE-TIME-THE-DOG-ENGAGES-IN-THE-BEHAVIOR… good luck with that! Last time I checked people in general have such a difficult time being consistent with anything. To put this one differently: The more times the behavior is followed by a punisher, the less likely the behavior will occur - the dog has learned something.

The punisher is always part of the punishing process. It is very difficult to separate the pain, etc. from who is delivering the blow. So bad news for the owner because most likely the dog will become to associate him or her with pain, fear, etc. (This is, of course, thanks to classical conditioning or associative learning - one of the most powerful ways of learning for creatures).

The punishment rarely teaches the dog an alternative. I HATE this! So the animal is getting hurt but does not know what to do instead. Oh, how very wrong!! The moral of this one point is to make sure we give the dog an out. We teach her what she can do instead to save her skin from pain.

Last, a very real and cruel result of using aversives to teach animals is that it is quite possible to end up with an animal that does not “behave” because it has learned - if punishment is not elegantly applied, that behavior produces pain and there is no sure way of escaping it. What is even more troubling is that many owners want a dog that does not behave. The animal’s joie de vivre gone!
I sure hope that writing about this topic makes people think many, many times prior to taking the “easy” (easy for whom?) way out.

Remember: Before you expect your dog to learn something ASAP, be honest with yourself about how painstaking it is to learn new stuff or go about making meaningful long lasting changes to your behavior. As such, let’s hope that we practice empathy for our dog's learning curves -especially when it comes to emotional issues (anxiety, fear) that are truly out of our dog’s control.

 

Do you really love me? Does it matter?

Laika our eight-year-old Kelpie-mix was born, I am guessing because we adopted her at the age of one, with weak hind legs. Her front legs and chest area are super strong. Because of the weakness in her hind area and because I am a trainer that loves a good challenge, I along with the help of a couple of great physical therapists designed a series of exercises to strengthen her rear legs.

We work on these exercises three times a week – similarly to a person going to a gym religiously. Her sessions of course are much shorter and she does not have to wear the ultimate in sports attire.

For the most part we work in the garage where I have a series of props and we move from one exercise to the next swiftly: me paying attention to the placement of her paws, her gait and balance, the wagging of her tail with open grin and making sure she gets treated for almost every well performed rep. A few weeks back, I started working with her on the same exercises outside. We make use of the natural terrain that is soft on the joints yet it represents a challenge because it is on an incline.

The sessions go like this: I work with Deuce - our young and energetic border collie on impulse control in a high-intensity game of whippet. If you have no idea what a whippet game is… stay tuned…. as I will write on this fantastic training tool/game in a future post. While Deuce relaxes and catches his breath, Laika and I work on her routine. A few days ago, as I launched her treat for her to “find” she remained with nose glued to the ground in a hopeful search for another treat. I called her after a few seconds to no avail. As I stood on our exercise “hill” the question: “do you love me”? popped into my head.

Well, then if you love me you would mind me and come rushing to our exciting game/practice…. If you love me more than a piece of kibble you would…Really? Does it matter if my dog loves me? Yes! It matters to my ego and to whatever motherly instincts I have (all of course directed towards four legged species- that is just my speed).

From the perspective of behavior and behavior modification however, my question is a moot point! It really matters little if my dogs love me because what really governs behavior (any behavior: our dogs, yours or mine) is MOTIVATION.

I have had more than one client feel somewhat embarrassed when after expressing (just like me) their desire to be loved by their pets followed by me telling them (as gently as I can) that it really does not matter if their dog loves them. In other words: to give up on the need to be loved by their dog!
Motivation should not be an undesirable “dirty” concept in our minds. It is actually natural – that is it is evolutionarily advantageous for any animal (yes, us too) to act out of sheer self-interest.

What is even more fascinating is that the motivation my dog can have at a given moment; including her lack of interest in our joint practice has so little to do with the wonderful, warm and fulfilling relationship we have.

Now take note: while I have let go of wanting my dogs to love me with adoration, I do want my dogs to trust me! I want them to feel safe and “know” that I will do anything in my power to protect them, meet their true needs and give them a break when their motivation is not 100 % in line with mine.
Does it matter that you love me? No, but it is sure nice to think of the possibility…