<< Return to Index     Next: Rewards >

Positive Reinforcement-Based Dog Training


Why use positive reinforcement at all?

Traditional training methods, which use "jerk and praise" techniques, rely primarily on fear and pain and the avoidance of both to motivate the dog to comply with commands. These methods were used in the military-hence the tradition of dogs heeling on the left, so you can carry your gun in your right hand. Since then, many trainers have changed their methods to use positive reinforcement, the desire for good things rather than the avoidance of bad things, as their primary training tool. Many of the techniques we use came from knowledge gained from training marine mammals-after all, using punishment with a killer whale is simply not a safe option.

It's true that it is possible to train a dog using aversive techniques, but the affect that the fear and pain has on your relationship with your dog must be taken into consideration. Most of us have dogs because we want to connect with them. It is much easier to connect to your dog in an environment full of trust, generosity, and empathy than one clouded with confusion, frustration, and fear.

A concern that many people have with using positive reinforcement is the reliability of behaviors trained using food rewards. Instead, we would like to our dogs to comply with our requests because they love us. Think of the best boss you've ever had. You really liked them, and you may have even loved them, but how long would you have continued to work if they stopped paying you? No matter how much you love your job, you still look forward to days off, when you can do what you want, rather than what your job requires. When we give our dogs a command, we are asking them to drop whatever they're doing and comply with our wants. In order for most dogs to think that it is worthwhile to leave the smell they are sniffing, or their soft spot on their dog bed, we need something more to offer than "good dog, Fido." After you've proven yourself to be a generous, interesting trainer, your dog will be motivated to respond to your requests even where there is no obvious reward present.

How does positive reinforcement work?

Any behavior that is rewarded is repeated. To train, we deliver or withhold rewards to influence which behaviors are repeated. This method has two main parts: the behavior, and the reward. The behavior is anything your dog does: sitting, watching, laying down, running, walking, wagging, jumping, biting, digging, peeing, pooping, eating, drinking, barking, sleeping. The reward is anything your dog likes: cheese, hot dogs, cheerios, walking, barking, laying on the bed, watching squirrels, and chasing balls. The key to success is finding out what is rewarding to your dog at any given point in time, in any environment, because it changes. If you know what your dog wants, you can use that knowledge to deliver or withhold the reward based on your dog's behavior. If your dog jumps on you, chances are he wants attention. If you withhold your attention (any eye contact, touching, or speaking) until your dog has all four feet on the ground, you are training your dog to get what he wants (attention) by exhibiting a behavior (staying on the ground).

The Bridge (Click or "Yup!")

To help our dogs figure out which behaviors earn rewards and which behaviors prevent rewards, we use a bridge. A bridge can be the sound of a clicker (a small plastic box that makes a clicking noise when pressed) or a word such as "yup," "yes," or "right." The bridge tells the dog exactly when he has earned the reward. For example, to train your dog to sit, you may click or say "yup" at the very moment that his bottom touches the floor, and then deliver your reward, a piece of cheese. Your bridge is only powerful as a training tool if it always predicts delivery of a reward. It is a way to give your dog information.

Cues/Commands/Signals

Because dogs don't understand English (or French or German for that matter), we use our bridge and delivery of rewards to communicate with them. Cues are added after a behavior is learned.

Once a dog has learned that a behavior is rewarding, we can teach them what it is called. If your dog is sitting every time you hold a treat above his nose, you can say "Sit" or use a hand motion just before he sits, click/"yup" as he sits, and then deliver the reward. Eventually, your puppy will learn that hearing the word "sit" or seeing the hand signal predicts that sitting will be rewarded at that moment.

It is important to say your cue once only. We want your dog to learn to respond to "sit." Not, "sit, sit, sit." It is often easier for your dog to learn a hand signal than it is to learn a verbal signal. Dogs pay more attention to what we do than what we say. If you want to teach your dog a verbal signal after he has learned a hand signal, you can add the verbal signal before the hand signal. Your dog will learn that hearing "sit" predicts the hand motion that tells him that sitting will earn a reward, and will start to sit in response to the word "sit," without waiting for the hand signal.

Help your dog succeed:

1. Tell him what you want him to do (not just what you don't want him to do)
2. Prevent him from failing. Use a crate, baby gate, and leash so that he can't make poor choices.
3. Reward every small success.
4. Follow the 2 Failure Rule: if you ask for a behavior twice and he doesn't respond, make it slightly easier and take smaller steps toward the end result. For example, if your dog doesn't understand how to lie down, reward them for dropping their head, then for dropping their head and crouching, then for crouching lower and finally for laying down with their elbows touching the floor. Taking small steps that are successful makes training go faster and keeps it fun.