In this post, you will learn how to use science to make dog training easy. There are lots of theories about dog training, but only one method has been scientifically proven to work the best- positive reinforcement. I use dogs as examples in this blog but the truth is that positive reinforcement works for any species- even humans. You’ll learn the basic principles of dog training and how to get your dog, cat, or even child to do what you want. You can even improve yourself with positive reinforcement.
If you REALLY want a deep dive into operant conditioning, the absolute best book on the subject is Don’t Shoot the Dog by Karen Pryor. While the book is based on science, it is very practical and applicable to real life, and even a very fun read!
The Science of Operant Conditioning
Reinforcement theory, or operant conditioning, is the science behind dog training. There are 4 basic ways to train a dog:
- Positive reinforcement
- Negative reinforcement
- Positive punishment
- Negative punishment
Positive and negative here do not mean good or bad. They are referring to either adding or subtracting something to the training system. If you add something to the training system, such as treats or shocking the dog, it’s positive. Negative means you are taking something away, such as your attention or forceful restraint. Reinforcement increases the frequency a behavior is offered by the dog, and punishment decreases it. Sounds simple, right? It can get a little confusing because of our alternative definitions of negative and positive but it’s really pretty easy to understand if you stick to the definitions I just provided.
So if you use positive reinforcement, you are adding something to the situation to increase the likelihood that a behavior will occur again. Positive reinforcement is easy to understand- if your dog does what you ask, they get a treat, and they do more of what you want. There are really only 2 tricks to successful positive reinforcement: your dog must understand what behavior is being rewarded, and they must want to work for the reward. That sounds easy, and in theory it is. In practice, it’s a different story.
Timing is crucial in training, especially if you are trying to reinforce a behavior that doesn’t last very long. This is why trainers introduced a marker, a way to communicate to the dog that the thing they are doing right now is the behavior you want. Some people use a clicker, some use a voice cue (Good or Yes). There are pros and cons to each- I don’t always have a hand available to click when my dog is doing something I want. I almost always have my voice- maybe not so much the day after a hockey game. The clicker is much more consistent, it always sounds the same. Plus a clicker doesn’t sound like anything else that happens in real life, it’s specific to training. My voice is not as consistent, because I don’t always use the exact tone or pitch when I am training. And the words “good” and “yes” aren’t exactly specific to training, we use those words all the time in every day conversation. So the dog may get confused to hear you say “good” or “yes” and think they are doing something right and then expect the rewards. That said, I use the word “good” and rarely use a clicker to train. It’s a personal preference.
Like I said, the marker is used to tell the dog that the thing they are doing right now is good. The reason we use a marker is that it is usually impossible and not always desired to give the reinforcer at the exact moment the dog is doing the desired behavior. If you try to give a puppy a treat while they are pooping outside, they’re going to get distracted and not empty their bowels and then come inside and poop in the house. If you wait until after the puppy poops, then call your dog to come inside and give then them a treat, you’ve trained your dog to come when called, but not to poop outside. So you need a way to mark the desired behavior as it is occurring to let the dog know that is the thing you want them to do, and that they will get reinforced for that behavior.
A reinforcer can be a treat, petting, playing, a favorite toy, anything that the dog is willing to work for. Some reinforcers have a high value- the dog would do ANYTHING to get it. Other reinforcers have lower value, the dog likes the thing but isn’t going to work that hard for it. Here’s the kicker- the dog decides what is high value and what is low value. I hear all the time that the dog won’t work for treats. Maybe not the treats you’re offering now, but there is something that the dog wants and is willing to work for in most situations.
It’s true that some dogs are much more food-motivated than others. But all healthy dogs are motivated to eat- they have to in order to survive. It’s also true that some people will feed a dog, or worse, food is always available, and they still expect them to work for treats. Nobody is motivated to work for food when their belly is already full. You have to set up a training session correctly for it to be successful. Train before feeding if you’re using a food reward as the reinforcer.
Pairing the Marker with the Reinforcer
Before you can start training your dog, you have to make sure the dog understands that the marker means a reinforcer is coming. This is especially important for dogs with no training experience. This is always the most fun part of training for the dog. All they have to do is look cute, hear the marker, then get a treat. At first, you mark and treat at the same time. As your dog is getting the treat, you mark that behavior with a click or “good.” I fed my dog several meals kibble by kibble to make sure she understood that “good” means food is coming. You don’t need a high-value treat for this exercise because you’re not asking the dog to do anything. Start by hand feeding, then start tossing the treat and have your dog get it. As soon as they get nose to treat, mark it. Again, timing is crucial! And it will take practice for you to get it right.
Next, you’re going to start varying the time between the marker and the reward. At first, just by a second. Click or say good, then wait a second, then give the reinforcer. Alternate between giving your dog the reinforcer at the same time as the marker, and few seconds later. You want your dog to pay attention to you the entire time they are waiting for their reinforcer. Eyes locked on your eyes, total focus. You can gradually increase the time between the marker and the reinforcer up to a minute even. You are successful if the dog is completely focused on you between the time they got the marker and the reward. You may need higher value reinforcements at this stage. Continue this exercise, varying the time between marker and reinforcer, until you have a dog that is completely focused on you.
Dogs are always learning even if we are not in a training session. Give your dog a marker for any behavior you’d like to reinforce, such as laying calmly at your feet when you are working or watching TV. If you’ve trained them correctly to understand the marker means a reward is coming, they will focus on you, so immediately toss them a treat.
Once a behavior is learned, we don’t always mark and reinforce it. I don’t tell my dog she is good for laying down at my feet when I am working, or give her a treat for doing so. But sometimes I do. Rewarding a learned behavior intermittently has been proven to increase the likelihood that the behavior will recur in the future. It’s like gambling, you have to play to win. If you always win, that’s not a particularly fun game. If you never win, why bother playing. But if you win sometimes, and you don’t know when, you’re going to keep playing over and over in hopes that you hit the jackpot.
True fact: intermittent positive reinforcement is the strongest most reliable way to train a dog to perform the wanted behavior when you ask.
Negative reinforcement means to take something away from the situation to increase a behavior. Negative reinforcement isn’t often used in training because it means you are applying an aversive first, then removing it once the dog performs the wanted behavior. For people, having your car buzz at you until you put on your seatbelt is a negative reinforcement. You are more likely to put on your seatbelt before your car makes that annoying noise in the future. Some people train their dog to sit by applying pressure to the dog’s rear end until their butt hits the floor. That can be confusing for the dog. If you tell your dog to sit, then force them into the position, the dog may think that “sit” means you’re going to force them into position. They won’t know they have the opportunity to avoid the force by sitting first because you never gave them that choice.
What's Wrong with Negative Reinforcement
Negative reinforcement can be taken way too far. Forcing a dog into something often increases fear. Holding a dog down to trim their nails in hopes that you are training the dog to accept nail trimming is absolutely backasswards. If you need 3 people to trim your pet’s nails, I promise you’re doing it wrong. The pet is only fighting because you made the nail trim scary from the beginning. It’s a vicious cycle. The more you try to force a pet to do what you want, the harder they fight. If the pet actually does submit, it is only because they gave up, just like an abuse victim who stops fighting. What does that make us?
We see this a lot in our practice because some veterinarians are just too busy to care. A lot of our patients are scared half to death of any sort of vet procedures and act aggressively if anything remotely resembles vet care- even if they are really sweet pets under all other conditions. I’m so glad that pet owners call us, even if they are only hoping things will be different because the pet is at home. Unfortunately, often, it’s not the building that scares the pet, it’s the treatment they received by people who are sworn to protect animals’ well-being. It takes a lot of very good experiences to overcome even a few really bad ones. How long does a bully need to be nice to you to prove that they are no longer a bully?
This is why we joined Fear Free and take the principles so seriously. It’s why we simply won’t make your pet do something that makes them uncomfortable. We love animals, and we will never risk causing irrevocable harm by trying to force them into submission.
If you are using positive punishment, you are adding something to the system that makes the behavior less likely to recur. Timing and consistency are crucial for positive punishment to work. And the punishment, or aversive, must be strong enough to overcome the dog’s motivation for the unwanted behavior.
Let’s use the puppy potty training example again. Your puppy poops in the house when you are away at work. You come home, and find your dog has pooped in the house. You yell at your puppy, maybe rub their nose in it. I’m sure your dog looks sad when you yell at her, but she didn’t learn not to poop in the house. She just doesn’t like being yelled at. What has your puppy learned? That when you come home from work, she will be punished. Why? Whatever the dog is doing at the moment of punishment is what the dog thinks you are punishing for. Your timing is WAY off because dogs learn in the moment. Just like with positive reinforcement, you have to catch them in the act of the unwanted behavior, and mark or punish them while they are doing it.
OK, got it. When you are home, you watch your puppy very closely. Whenever you catch her pooping in the house, you smack her with a newspaper. That ought to teach her, right? Problem is that when you are at work, nobody is there to take her outside, and she still poops in the house. Now you’re not punishing her for pooping in the house when you are away, but you are when you are home. What did your puppy learn? To poop in the house when you are not home. Why? Because you aren’t punishing the behavior consistently enough for her to learn that it is never OK to poop in the house.
Fine. You’ve had it up to here with poop in the house. You tell your boss “I can’t take it any more. I need a week off to house train my puppy.” Your boss, being an animal lover says of course. You’re determined to get this puppy house trained, so you watch her like a hawk. Every single time she poops in the house, you catch her in the act and smack her with a newspaper. You got your timing and consistency 100% perfect. This puppy is never going to poop in the house again. Or so you think. But she does! Now what is wrong?! The punishment isn’t aversive enough to overcome the relief of emptying her bowels.
More Problems with Positive Punishment
These are the inherent problems with positive punishment. It’s almost impossible to be 100% accurate with timing and 100% consistent with the application of the punishment. The single exception to that is an invisible fence. It is always going to deliver a shock to your dog whenever the dog gets within the fence boundaries. The problem is that the shock may not be enough of an aversive to prevent your dog from gritting their teeth and going through the fence anyway. So even if you are fine with the idea of shocking your dog, which DOES HURT, it probably won’t even work in all situations. The aversive, whether it’s a shock, choke, pinch, smack, squirt, loud noise, or anything else, isn’t always enough to overcome the internal motivation for the unwanted behavior.
IF PUNISHMENT WORKED, YOUR DOG WOULD STOP DOING THE UNWANTED BEHAVIOR.
If you’re using positive punishment and the unwanted behavior isn’t actually decreasing in frequency, you’re not training your dog. You’re just being a jerk by repeatedly applying the aversive. Doing the same thing over and over again and expecting a different result a definition of insanity. Let’s stop the insanity, PLEASE!
Positive Punishment Increases Fear
On top of that, for fear-related behavior, you’re likely to make the problem worse. Let’s say you’re afraid of snakes and scream every time you see one. I punch you for screaming. Are you any less afraid of snakes? I doubt it. You’re probably even more afraid to see a snake when I’m around, because now there’s a snake AND you’re going to get punched. If I’m not there to punch you, are you going to scream? Yes, because you’re still afraid of snakes. What do you think of me for punching you? That I’m helping you overcome your fear of snakes, or that I’m a jerk? I’m sure the latter.
It’s the exact same way for a dog who is dog-aggressive on a leash, for example. A dog on a leash knows their ability to defend themself is limited. If they are afraid of other dogs while on a leash, they will bark and lunge and growl and act a fool to make the scary dog go away. Their behavior is reinforced by the other dog owner, because they’re going to walk away from your aggressive dog. Your dog got what they wanted- the scary other dog went away. That is a very strong reinforcing outcome for the aggressive behavior.
Now, if you decide to use a shock, pinch, or choke collar to “correct” the barking/lunging/growling behavior, that is the exact same thing as me punching you for screaming when you see a snake. This makes the underlying problem even worse. Now every time your dog sees another dog, he IS hurt, confirming his worst fears. You made the underlying fear even worse. IF you are successful in reducing the unwanted behavior- and that’s a big IF- you haven’t changed the fact that your dog is more anxious when approached by another dog on the leash. Your dog doesn’t bark, growl, or lunge any more, so you think you’ve fixed the problem. You allow the other dog to approach, and your dog bites the crap out of him without warning. Why? All you’ve done is repress your dog’s expression of their anxiety. You punished your dog for communicating with you. And you’ve made your dog more dangerous because he won’t give any warning before biting a person, either.
Let’s recap the reasons we don’t use positive punishment in dog training:
- You’re never going have perfect timing.
- You’re never going to be perfectly consistent.
- The aversive has to be bad enough to overcome the dog’s motivation for the unwanted behavior (ie, you have to hurt or scare your dog).
- It will worsen fear-based behaviors.
Negative punishment isn’t as bad as it sounds; it simply means to take something away in order to decrease a behavior. If you turn away from your dog when they are jumping on you, you are taking away your attention to decrease the jumping.
The only real problem with negative punishment is not following up by teaching your dog an alternative behavior. A dog can’t not do a thing. They do something, or do something else. They can jump on you, or they can go to a mat and sit calmly, but they can’t not jump on you.
Therefore, we use negative punishment to decrease an unwanted behavior along with positive reinforcement to increase an alternative incompatible behavior. A dog can’t jump on you AND sit calmly on a mat. Only one of those behaviors is going to be rewarded, so guess which one your dog chooses? Yes, of course, to sit on a mat. If you don’t teach your dog an alternative behavior, they’re going to offer a bunch of random behaviors, including jumping on you, and you’ll never actually solve your problem.
Unfortunately, dogs are accidentally intermittently positively reinforced when your plan is to negatively punish them. Remember how I said that intermittent positive reinforcement is the strongest way to increase a behavior? Giving your dog one nanosecond of attention while they are jumping on you strongly increases the likelihood that jumping is going to be the behavior they offer again and again when you come home. This goes for jumping, begging, pawing, getting on furniture, and any other unwanted behavior you’d like your dog to stop doing. Caving is the absolute worst thing. Allowing them to do the unwanted thing sometimes intermittently positively reinforces the behavior, making the unwanted behavior even more likely to occur in the future. You HAVE to be more stubborn than your dog.
So don’t just do negative punishment for a problem and expect the it to solve itself. Positively reinforce an acceptable incompatible behavior so your dog knows what you want them to do.
Teaching Your Dog New Tricks
Dogs are always learning throughout their entire lives, whether you are training them or not. You can always teach a dog new tricks, no matter what their age or previous life experiences. There is no situation in which positive reinforcement training does not improve the life of your dog. EVERY dog can learn, if the conditions are right. For dogs with severe psychological damage due to previous experiences (abuse by an owner, veterinarian, trainer), they can still learn, but they may need some medication to put them in a mental state conducive to learning. Check out our post on Behavior Modification for more information. The point is, every dog can learn, and deserves the chance for a happier, more confident life.
What a Fabulous Family Dog Really Needs to Know
This may come as a complete surprise, but there are only 5 essential behaviors that a fabulous family dog needs to know:
- Calm Down
- Come Here
- Walk Nicely on a Leash
- Leave Stuff Alone
Come to our Solutions Shop to learn how to apply the training techniques to teach your dog these 5 essential behaviors.
Training doesn’t have to occur in set sessions. You can do random acts of training in the course of your normal day and it really reinforces your dog’s formal training. I recommend carrying a clicker if you use one, as well as a treat pouch filled with at least half of your dog’s daily ration. These silicon treat pouches are perfect. Your dog should “earn” breakfast throughout the day, then fed the rest of their daily ration at dinner- as long as you’re not training after dinner. If you train after dinner, then give your dog breakfast, and make them work for the other half of their food in the evening.
There are 4 techniques that can be used in positive reinforcement training:
Capturing means that you mark a desired behavior as your dog randomly does it. This technique works extremely well for cats too! If your dog is quietly and calmly laying at your feet while you work from home, every once in a while, mark that behavior, and toss them a couple of kibbles of their food. If they go to their crate and rest, mark that behavior and toss in a few kibbles of food. Don’t get excited about the marking part, else the dog will also get excited, and you don’t want to train them to get excited, just to continue relaxing. If they’re laying nicely on the sofa while you’re watching TV with you, mark that behavior, and toss them a few kibbles. This technique is key to reinforcing calm independent behavior in the home.
You can also capture cute things your dog does. Annika lays down with her front legs crossed in the most prissy fashion as a natural thing. I can wait for her to do that, mark the behavior, and reinforce it. Once she understands what I am marking, I’ll put the behavior on cue. You will know your dog understands what you are reinforcing if they start offering that behavior multiple times in quick succession. Be careful not to mark or reward any other tricks your dog may offer because you don’t want to confuse them. Once Annika is offering the desired behavior in quick succession, I’ll give the trick a name, “prissy.” Every time she does the trick, I’ll mark it first, say “prissy” then give her the reinforcer. She’ll associate the cue “prissy” with laying down and crossing her front legs, and do it when I ask.
If you want to get really fancy, you can put an undesired behavior on cue. I do this for bark-begging. I know that huskies (my breed of choice) will offer barking to demand treats, so I wait for them to bark, and train them to do it on cue. Then I very rarely ever ask them to bark- NEVER when there is food around.
Capturing an undesired behavior is a great way to extinguish an unwanted behavior- put it on cue, then never ask your dog to do it. If your dog jumps on you whenever you come home in your nice work clothes, you can put jumping on cue. When you come home, give your dog a different cue, such as “place” or “sit” instead. You can still ask your dog to jump on you as a cute party trick if you want, or never. It’s up to you.
Here’s a video on capturing the behavior of staying with the trainer instead of ambushing the gentleman walking by.
Luring is exactly what it sounds like- tempting your dog to do a specific thing by offering a reward. This technique works great for a lot of the basic cues: sit, down, heel, come, and mat training. A video is worth a thousand words, so I’ll just show you what it looks like for each of those behaviors.
Sit and Stay
Shaping is the most difficult conceptually and in practice. When you are shaping, you get an approximate version of the behavior you want, mark and reinforce it. Then you slowly change the rules to be closer to the behavior you really want. People who train dogs for very specific behaviors like doggie dancing and acting will often use shaping techniques to get the dog to perform just right. You can train a dolphin to jump higher, a fish to shoot basketball, and a dog to do a backflip with shaping.
The dog in this video is already familiar with the concept of shaping. Teach a much easier trick, like getting all 4 paws on a mat or bed, before moving to something like this.
Chaining means to link specific behaviors in order. Dogs learn routines very quickly, which is a skill that can be harnessed to complete all sorts of seemingly complicated tricks. You break down a complicated task into smaller bits, and teach each of them separately.
Chaining can be done forwards or backwards. Sometimes it’s easier to start with the end because the terminal task is self rewarding. Otherwise, starting from the beginning may be easier for the dog to learn, or the task can’t be performed backwards.
Forward Chaining. One of my life goals is to teach Annika how to go to the fridge and get me a beer. This only really works in the forward direction, as shown in this video. This is an awesome video of a RESCUE dog who had been returned to the shelter multiple times learning how to get the owner a beer! Like I said, ANY dog can learn new tricks.
Back-chaining means to teach the dog the LAST part of a sequence first. It’s awesome for teaching a dog recall- to come when called and sit at your feet.
To train your dog, set you and your dog up for success. There are very specific things you can do to increase the likelihood that your dog will learn and perform the desired behaviors on cue.
- Use the proper equipment. Check out our Solutions Shop for our recommendations.
- If you are using a food reward, train your dog on an empty stomach so they are motivated to work.
- Remember, the dog chooses the value of the reward, not you. Just because you are offering your dog a hot dog doesn’t mean that your dog will think that it is worth the effort to earn it. Your dog may want peanut butter or chicken or petting instead.
- Carefully watch your dog’s behavior for signs of fear, anxiety, or stress. Stop whatever you are doing, and do something that you know your dog likes. For example, some dogs do not like having treats tossed in their direction- it scares them. Training is supposed to be fun and enjoyable. Don’t toss your dog treats if they don’t like it.
- Training should occur in an area without distractions until your dog is reliably responding to the cue. Distractions can be other dogs, people, a leaf blowing across the yard. Again, your dog decides what is distracting, not you.
- While training may occur in set sessions, your dog is always learning. Use that to your advantage.
- Training sessions should be short. Yes, most training classes last an hour, but we are typically only working the dog in 3-5 minute bursts at a time. Younger puppies have very short attention spans.
- Always end on a good note. If your dog’s attention is starting to wane, give them a well-known cue and a treat bomb to end the session.
If you want help training your dog, and you are in the Sanford, NC area, our Fear Free Certified trainer Jael offers group and private training courses to help you improve your relationship with your dog. You do not have to be a Family Veterinary Mobile Clinic customer for you to participate in our training courses.