Dog paying attention
This dog is paying strict attention. Photo by Dr. Teri Ann Oursler/VIN
The article, Dog Training Using Rewards: Why, covered why positive reinforcement training is so beneficial and important for you and your dog. The second part covers some of the basics in how to get started in reward-based training and some of the concepts behind this training. The basics work with any animal, including people, but the details may have to change depending on who you’re training. For example, most dogs work well for liver in many different forms: liverwurst, freeze-dried liver, cooked liver. But for many people, liver is not at all rewarding and a better choice would be chocolate.
Rewards and Reinforcers
In theory you can use anything your dog likes as a reward. Food is the easiest reward to use in most training situations. Some dogs are motivated to work for access to a favorite toy, especially the chance to tug with someone. Other things like belly rubs or the chance to snuggle on the sofa may not be practical in all situations, such as when you’re out on a walk or riding in the car. Petting can be distracting, turning a training session into a petting session can interrupt what you are teaching your dog. Save rewards like petting or snuggling for behaviors your dog already knows, not when you’re teaching something new.
- Find a food that your dog loves, preferably something that the dog can eat quickly without a lot of chewing and that isn’t too high in calories. Freeze-dried liver; cooked skinless-boneless chicken; semi-moist dog treats; and cut up hot dogs or cheese are all good choices.
- Use tiny pieces so you can give lots of rewards without your dog getting full or distracted between rewards. For small dogs, treats should be broken into pieces as small as a lentil or split pea. For larger dogs, the size of a kidney bean or Cheerio works well.
- The more difficult the behavior, the higher value rewards may have to be. A dog who will sit perfectly in the house for a piece of his regular food may need something better, like hot dogs or cheese, to sit outside while walking when he sees a squirrel. Think of it like hazardous duty pay.
The best way to train a new behavior is through continuous reinforcement. This means you need to reward (reinforce) your dog each and every time he chooses to do the right thing. That’s why the small, low calorie treats are so important.
Once your dog is 80-90% reliably performing the new behavior, you can switch to intermittent reinforcement. This means you become like a slot machine. He never knows when he will hit the jackpot so he keeps working to get the “win.”
Rewards must be given immediately. Timing is crucial. Remember to only reinforce the behaviors you desire. For example, if you wait a little too long to reward your dog for staying, he may get frustrated and bark. If you give the reward after the bark, you just taught your dog to bark for the things he wants.
Because we can’t always have a food reward available the instant we need one, it helps to pair food with a sound. This sound can be praise or a simple word like “Yes” or “Good” or a clicker. The sound becomes a bridge (promise) letting the dog know that a treat is on the way.
Troubleshooting Reward Based Training
How do I get him to do what I want in the first place?
Rewarding your dog for doing the right thing means giving him the opportunity to offer the behaviors that you want to reward. There are three basic ways to do this.
Dog lying down
The dog is laying down on command. Photo by Dr. Teri Ann Oursler/VIN
- Capturing: Reward your dog when he naturally offers a behavior you like. This is where using a marker, like a clicker or a word, can help. For example, when you sit on the couch you know your dog will lie down at your feet. Reward him any time he offers this behavior (lying down). If you don’t have a treat in your hand when he lies down, mark the behavior with the click or word you taught him means “Food is on the way,” then fish the treat out of your pocket. Next, as he’s getting ready to lie down, you say, “Down,” then reward him for lying down. Pretty soon, you will notice that you can say, “Down” even when you’re not near the couch, and he will lie down waiting for the treat to appear.
- Luring: Luring involves using something that your dog follows in order to get him to perform a behavior. For example, teach a puppy to sit by holding a treat above his nose then slowly move it towards his tail. As he tries to keep his eyes on the treat, his head goes up and his back end goes down. Wait until he’s sitting to give the reward. As you pass the treat (lure) over his head say, “Sit.” Eventually you can phase out the lure and just say “Sit” to cue the desired behavior.
- Shaping: This method involves rewarding small steps towards a final behavior that you want. Shaping may be combined with either capturing or luring. In agility training, a dog can be taught to walk up and tip a teeter-totter by having her:
1. Step just her front feet on to the teeter for a reward.
2. Then introduce a lure to get her to follow it until she has all four feet on the teeter.
3. Gradually, giving rewards each step along the way, work to the point where she will walk all the way across the teeter without the lure.
I know he knows it but he won’t do it.
Dogs are smart but they don’t always think the way we humans do. There can be several reasons why a dog isn’t “obeying” a person. We have to look at things from the dog’s perspective to problem solve what may be going on.
- He doesn’t understand what you’re telling him to do. Dogs communicate through body language. When you’re training him, you may think he understands what you’re saying, but in reality, he may just be following your body language. A common example of this sort of misunderstanding occurs when people train behaviors like “Sit” while standing straight and tall in front of their dog. The dog might not understand “Sit” if the person is lying down or sitting in a chair. This is why it’s often easier for dogs to learn hand signals than verbal cues. Hand signals also can be easier to transfer from person to person since they don’t depend on pronunciation or a specific tone of voice.
- There’s something much more interesting to do. Dogs are often distracted when they see things that they want to play with or smell something to eat or roll in or hear interesting sounds around them. It’s hard work for my terrier to shift his attention away from a squirrel on a walk. I have to reward him with high-value treats and continuously reward him for focusing on me until the squirrel is gone.
- The dog anticipates something unpleasant. Even if we are training with rewards, we may accidentally do something the dog considers unpleasant. This is a common problem with many dogs’ recall (“Come”). People often call off-leash dogs to “come” to them when it is time to leave the park or put the dog back in the house before leaving for work. The dog quickly learns that if he comes when called, the fun is over. In this case, it’s important to practice the recall in a variety of different situations with really high-level rewards. Also sometimes call the dog to come to you, reward him, then let him go back to playing.
Anxious English setter
This dog is anxious. Photo by VIN.
- The dog is too anxious or fearful to follow your cues. Anxiety turns off the rational, thinking parts of the brain and turns on the survival parts (fight or flight). At lower levels of anxiety, working for rewards can help refocus the dog’s mind and reduce his anxiety. But if anxiety increases, the dog finds it harder and harder to pay attention to what is being asked of him. A dog who is afraid of thunderstorms may be able to follow cues when the sky starts to get dark, but he’s slower than during calm weather. As the rain falls and the thunder starts, he becomes too anxious to do what’s asked of him.What do I do if my dog isn’t “getting it,” doesn’t “obey,” or is “being stubborn?”
People get frustrated if training isn’t going perfectly and start to look for quick fixes. This frustration often leads them to try punishment-based options or so-called “balanced trainers” who use rewards to teach a behavior followed by punishment later in the training process. There’s a growing body of research showing us that dogs trained via positive reinforcement versus punishment-based training methods (especially using shock collars) are happier and less anxious during training. This is true whether they are learning something new or practicing cues they already know. Learning occurs more efficiently with reward-based training than other methods. Physically punishing an anxious or unsure animal can lead to aggression.
Simple Keys to Success
Keep it fun: If you and your dog aren’t enjoying the training session, go back to simple things that he already knows so you can have success or just take a break from training altogether. Always train with a happy, upbeat voice and relaxed body language to signal to your dog that this is something fun.
Watch what your dog is telling you: If you see signs that your dog is getting anxious or frustrated, change the situation. Do more simple training or move away from what’s making him anxious. Early signs of anxiety include ears held back, tail down or tucked, eyes darting around or seeing the whites of his eyes, hunched posture, panting when not hot, taking treats more slowly, as well as responding slowly to cues he knows well.
Set realistic goals: Don’t try to do too much at once. Keep training sessions short so they stay fun. Break down what you want to teach into smaller pieces to make it more manageable. For example, to teach a dog to lie down on a dog bed while you eat dinner you need to teach him to go to the bed, to get onto the bed, to lie down on the bed and to stay on the bed. Teaching him to stay on the bed involves starting with just a couple of seconds and gradually building up the duration. It also involves teaching him to stay there while you move further away from him and take your focus off of him.
Be consistent: It’s hard to learn something new if the rules are constantly changing or you keep getting different instructions. Make sure everyone who is training your dog uses the same cues (words, hand signals) and knows how to give rewards, especially the importance of giving rewards immediately when teaching something new.
Make yourself look great: If you want to look like the best trainer in the world, take something cute your dog naturally does and put it on cue. For example, if your dog rolls on his back when you take his harness off after a walk, mark that behavior and give him a reward. When you take off his harness say, “Roll” then reward the roll. Pretty soon you can show off this trick to your friends. One of my small dogs would stand on her back legs to check things out. I gave it a name and rewarded it. She really impressed people when I would ask her “What does a Meerkat do?” and she’s stand right up.
Reward-based training isn’t just good for your dog. Once you and your dog get the hang of it, it’s also fun and quite easy. It really is a gift that keeps on giving.
VIN News Service commentaries are opinion pieces presenting insights, personal experiences and/or perspectives on topical issues by members of the veterinary community. To submit a commentary for consideration, email firstname.lastname@example.org.