One question rewards-based dog trainers who use treats often get is, “When can I stop using the food?” Invariably, this question is putting the cart way, way, way before the horse. The dog hasn’t been taught behaviors yet, and their people want to know when they can discontinue reinforcing with food.
First, it should be understood that the best animal trainers use food to train. Karen Pryor, who popularized clicker training, uses food. Legendary animal trainer Bob Bailey, who quite possibly has trained more animals of more species than any other living trainer today, trains with food. I have dozens of dog training books by the best in the business on my shelves, books by trainers and behaviorists with advanced degrees in animal behavior and longevity in the training business, and they all train with food. Trainers who work with aggressive animals, including the president of the International Association of Animal Behavior Consultants whose practice is limited to training aggressive dogs, train with food. Trainers who are concerned with their animals’ welfare as well as quality training use food.
There is nothing wrong with training with food, and many, many things that are right in doing so. Dogs and all animals learn by making associations and experiencing consequences of their behavior. If the associations are happy ones, the dog feels good about what is happening. A frequent example is a dog’s leash predicting that a walk will happen. Dogs generally want to be walked, so the association between the leash and walking is usually a positive one and the dog jumps about in anticipation of the walk. But for some dogs who fear being outside, the leash might actually have negative associations and the dog may avoid being leashed. The leash predicts scary things, not happy experiences for these dogs. The consequence for the first dog being leashed is a fun walk outside. The consequence for the second dog is a scary walk outside. The first dog approaches to be leashed, maybe a little too boisterously, but we can teach that dog to sit to be leashed and go out for his fun walk. The second dog avoids the leash altogether, moving away from his person and possibly cowering. If you find yourself dragging your dog out the door for his walk because of fear, this can be improved. With what, you ask? Food!!!!! There are details involved in this work, but the gist is to build positive associations to going outside using awesome food.
Providing food as the consequence for behavior in your training can increase positive associations both to training itself as well as to people and things in the environment when training takes place. Using food as an essential part of a positive reinforcement training program is also highly effective and facilitates learning (1), (2). We teach the dog what we want him to do and we reinforce that behavior using food, repeatedly, so that we will get more of the behavior we seek. We may vary our reinforcers, using play or activities a dog enjoys in addition to food. But we want to reinforce behavior we like a lot to keep getting more of it. One law of behavior, the Matching Law, is that you get what you reinforce. So, if you’re reinforcing a recall (come when called) half of the time, you will get a recall half of the time. If you want your dog to come to you all of the time? You have to reinforce that behavior all of the time, and reinforce it well because the environment is full of competing reinforcers like squirrels to chase. You also have to proof the behavior so that you stand a chance of getting it in relation to distractions like squirrels or when you’re at a distance from the dog. This is a graduated and repetitive process.
It takes successful repetition of tasks in many different environments for a dog to generalize the behavior so that he performs the behaviors reliably in various situations. Walking nicely on leash on a quiet suburban street is not the same as walking nicely on leash in Manhattan. Holding a down-stay at home is not the same as holding a down-stay at a restaurant. You have to train for that, which means you have to reinforce for that. And it’s the dog who determines what is reinforcing. If the dog’s behavior isn’t maintaining or increasing with what you’re providing as “reinforcement,” it’s not reinforcing to your dog.
It is sometimes said that it takes 10,000 hours of practice for a person to become an expert at playing the violin. It takes an equivalent amount of practice for a dog to be able to do many of the tasks we want the dog to do in all of the environments we want him to do them. Pay for those behaviors with food and other things the dog likes A LOT. Don’t worry about when you can stop with the food. You may never stop with the food. There are good reasons not to stop, such as wanting to maintain high levels of reliable behavior. Behavior needs to be reinforced to continue. Would you keep going to work without being paid?
You may be able to get more behavior for the reinforcement you provide, especially if that reinforcement is generous. (Again, there are details involved in successfully reducing rates of reinforcement while maintaining behavior. They generally involve having really high rates of reinforcement to begin with and periodically returning to it.) People tend to be stingy or misguided with their reinforcement and then think positive reinforcement doesn’t work. But it does, by definition, and according to science. Behavior that is reinforced maintains or increases. You get what you reinforce. Be generous about it!
(1). Hiby, E. F., Rooney, N. J., & Bradshaw, J. W. S. (2004). Dog training methods: their use, effectiveness and interaction with behaviour and welfare. ANIMAL WELFARE-POTTERS BAR THEN WHEATHAMPSTEAD-, 13(1), 63-70.
(2). Rooney, N. J., & Cowan, S. (2011). Training methods and owner–dog interactions: Links with dog behaviour and learning ability. Applied Animal Behaviour Science, 132(3), 169-177.