Monday, February 23, 2015

Reinforcement, rewards, treats and toys

This post is for anyone who is concerned with using food in training, who is afraid their dog will "only do it if they have treats" and who think that their dog should "know it" by now and be willing to "do it" just because they said so.

I continue to have these questions asked even by people who I have been working with for a while and I feel I need to address it.  Perhaps I am not clear in my way of explaining things and have failed in some way to help people understand how reinforcement really works.

Reinforcement by definition in terms of training and behavior refers to something that makes a behavior happen again and more often.  We use things like food, attention, access to things, treats, toys, play and more for reinforcement.  Here is the kicker, the LEARNER, meaning the dog is the one who gets to determine the value of the reinforcement, not the trainer.  If what you are using as "reinforcement" is not serving to increase the frequency of the behavior then it is not reinforcement! So, you can pat your dog on the head all you want or say "good girl" all you want, if the behavior is not getting stronger or even maintaining, you aren't reinforcing and in fact, you aren't training your dog to be more reliable.  The first thing we need to determine is whether or not something is actually reinforcing to the animal. If you are insisting that you will use only praise to reinforce your dog but the dog does not find your praise valuable enough your dog's training will not be reliable, plain and simple.  Dogs and other animals don't come programmed to just do whatever we say because we said it.  Neither do people, actually.  We all behave and do things because of the outcomes it brings.  Yesterday, I went to the vet and picked up medication for my close friend's dog.  She was very busy, was going away this weekend and needed the meds and I had time so I ran and got them and brought them to her.  I did not do this because I would be reinforced in some way or because I would get something.  I did it because my relationship with this friend is based on years and years of a trusted friendship that frequently involves us helping each other when needed. I do things for her and she does things for me because we CHOOSE to based on our history which includes a lot of positive reinforcement.  This is not unlike the relationship with my dogs.  In the beginning, when I was building a relationship with them I used food to reinforce everything they did and learned.  I do my best to set them up to get it right, so that I can reinforce their correct behavior and continue to add "insurance" and more reinforcement history so the behaviors are strong and remain strong. Over and over and over until they pretty much always gave me the response I was looking for.  Then, and only then, did I begin to add in other reinforcers to the mix.  Only after I was truly convinced that they had a solid understanding did I begin to ask them to do things and use different things to reinforce them. Things like attention, scritches, access to things, play, toys or sometimes just a wink and telling them how clever they are, but I am always continuing to reinforce them. 

Take a look at this video.  It shows a variety of different reinforcers used for different animals.  It is critical to keep in mind that just because a reinforcer is effective in one instance or environment does not mean it will it be in another environment.  For instance, Sonic is shown enjoying playing tug and fetch in the video, however, I know from experience that this would not be the case outside or in a group class.  In a class setting I would need to use food with him.  Reinforcers may need to change depending on the level of distractions, the environment or just the level of difficulty of the behavior.  https://www.youtube.com/watch?v=CE_CYoNMvAM.

When training a new behavior we may use food as a lure.  In other words, use food to guide the dog into position to accelerate the dog being able to do the behavior.  Frankly, when training my own dogs I use a lot more shaping.  I don't do as much shaping in classes and with clients because it is more challenging to learn how to do this, but honestly people tend to get frustrated having to reward such small steps and having to wait for things to happen.  So, I use a food lure which of course, has it's own challenges.  When using food as a lure, you must stop luring as quickly as you can.  I will sometimes lure a dog only two or three times before waiting for the behavior to be offered or prompting without food, however I am STILL reinforcing with food after they do it. It is critical to stop luring with food as soon as you can or you may get into a situation where the dog "only does it when you have food".  This does not mean using food is flawed or that the dog is being manipulative, it means that the execution of training was not done correctly.  Food was used as a lure too long.

This next part is probably the most important thing.  It is critical that if behaviors are to be strong, the animal continues to be reinforced for them.  After the behaviors are learned and reliable we can begin to play with different reinforcers and change things up, but don't fool yourself into thinking that you can simply stop reinforcing the animal.  People are quick to admit that they would not go to work without some form of pay or positive reinforcement for their work, but want their dogs to behave like little robots.  We have to remember that while we consider them "ours", dogs and our other companion animals have their own lives, feelings, strengths and weaknesses.  We cannot expect them to just do what we say because we said it.  We also cannot expect behaviors to remain strong and reliable if we drop off the reinforcement for the animal.  It simply doesn't work that way.  If it did, it would be easy, simple and every dog walking around would have impeccable manners and be from from behavioral issues.  It takes time to build reliability.  It takes working them around different distractions and in different environments always reinforcing them heavily when they make correct choices.  

It is time that we stop viewing dogs are little automatons and seeing them for what they are, intelligent, thinking, feeling individuals.  

I always try to look at my training and my dog's reliability honestly.  When something isn't going how I would like, it is ALWAYS my issue, not the dogs.  I haven't practiced enough, I have not proofed enough, I have not worked on something enough.  If I had done my part better, my dog would be more successful.  

So please, use food with training.  Reward the dog often