Just a brief reminder about behavior…teaching using positive reinforcement is not about bribing, and most certainly is not about force. Scientifically speaking, positive reinforcement is a consequence of a behavior that either maintains or strengthens it. As trainers, we are using positive reinforcement to build value for a behavior by pairing it with something the learner values. To be used effectively in teaching new behaviors, that reinforcement should be delivered contingently (meaning ONLY if the behavior occurs) and contiguously (meaning very closely following the behavior.) In this way, you are teaching your student this: WHEN I do THIS, THEN THIS positive outcome will happen.
(Note that this was actually written several years ago for something. I just found a copy of it.)
I was thinking about this the other day when I was working on teaching our family dog, Sam, a new behavior. I was working on teaching him the leg weave (where he runs around a leg when you move it in a direction, then around the other leg), and he lost interest very quickly in the beginning.
I know that if an animal I am training is not getting the lesson plan, and would rather do other things at the time, that the animal is simply giving me feedback that I am not teaching in a way that motivates him to want to stay in the classroom.
So I took a step back and thought about what I was doing.
Was I setting him up for success or was I setting him up for failure?
These are some questions I needed to ask:
Was the environment too distracting? (Remember, it is important when teaching a new behavior that you begin in an environment with minimal or no other distractions so as to have your animal’s full attention. You only add distractions slowly as your student tells you by his ability to focus on doing the behavior that he is ready for it.)
Did I have high enough value reinforcers? (Please read my post on why knowing your pet’s List of Awesomeness is important)
How was my timing between when he did a behavior approximation and when I gave the reinforcement? (The shorter the time lapse between when the behavior occurs and the reinforcement is produced, the easier it is for the animal to learn that association.)
Was I going at the right pace for him? (In my shaping plan, were my behavior approximations at an interval that were enough to teach but not so much as to be too difficult? See my post on shaping.)
Was I completely focused on the training so as to catch his behaviors that I wanted to reinforce?
Was my training session short? (Training sessions should be short, 3 to 5 minutes, ending on a positive note.)
In this particular lesson, it was a very quiet living room with minimum distractions and so I was not using the highest value reinforcer. I was using Sam’s dog biscuits. Also, while I began teaching this by luring him around my leg (having him follow a treat), I did not give him a piece of the treat until he was half way around the leg.
That combination was not setting him up for success. Waiting until he moved half way around my leg was way too long of an approximation to keep his interest, especially when his reinforce for doing so was a piece of dog biscuit.
What did I do? I added some pieces of chicken to my reinforcer bag and, while I used luring at the beginning, I marked (and gave him a reinforcer) for his moving around my leg in very small increments the first few times. There were six locations around my leg that were my target points for marking and reinforcing.
That small change made a huge impact. In only a few trials, Sam was running around my leg and then around the other.
Next, I stopped at that success and practiced another behavior for a couple minutes, then another behavior for another couple minutes, and then went back to the leg weaves. Wow, was he charging around my leg by that point. I had made my classroom so much fun for him that he forgot he was there to learn…but learn he did! And with gusto!
The next step was to begin fading out the lure and using less frequent target points around my leg until I could get to the point where I would just give a hand cue and he’d begin to weave. Now we are working on shifting to just my leg movement as his cue to begin the weave.
My point with this post is that, whenever you are training an animal who just simply is not doing what you want him to do. Know that there is always a reason. And it is not that your animal is dumb or stubborn. As your pet’s teacher, step back and think through what you can do to make the lesson plan easier, more clear…and definitely more fun!
Yesterday, I watched as a client’s puppy jumped and wiggled when she tried to clip a leash onto his collar.
“Is that the behavior you would like for him to exhibit each time you present a leash?,” I asked. “No,” was her answer.
I explained how, from her puppy’s standpoint, he is learning that his jumping behavior must be the cause for her to put his leash on since she clips the leash on immediately after he begins jumping. In other words, since his behavior was continuing, and even strengthening, we then know jumping is being positively reinforced by clipping on his leash. I then showed her how to teach him based upon consequences what behavior it was that she wanted him to do instead, which was sitting with his head still while the leash was clipped.
Here is the thing that is so important to remember about behavior. It NEVER occurs in a vacuum. Behavior is simply a tool that animals use to get consequences. They either use behavior to move toward something positive or to move away from something aversive. It is whatever happens immediately AFTER a behavior that influences the future rate of that behavior.
I thought I’d write about those consequences to give you a little more insight. There are four quadrants of consequences. They are Positive Reinforcement, Negative Reinforcement, Positive Punishment and Negative Punishment.
A good way to think about them is in this graphic.
So, what exactly does all this mean? I’ll explain.
Scientifically speaking Reinforcement is any consequence of a specific behavior that INCREASES and/or STRENGTHENS the future rate of that behavior. Punishment is any consequence of a specific behavior that DECREASES the future rate of that behavior.
People often think that reinforcement only refers to things like food or a toy, but reinforcement by definition simply means ANY consequence that causes a behavior to increase. On the flip side, punishment does not only refer to scolding, spraying with water, or spanking. Punishment is ANY consequence that reduces the probability of a behavior. It is important to note here that ONLY the animal that is doing the behaving chooses whether a consequence is a punisher or a reinforcer. Always, that value is from the eyes of the learner.
As some examples, if, after your bird screams, you yell back at it; and the bird continues to scream and even scream louder, THEN we know that your bird is being reinforced for screaming by you yelling back at it. If you reach down and pet your dog after it jumps on you, then your petting it is reinforcing its jumping. On the other hand, if you walk out of the room when your bird screams and your bird screams less in the future, then, from a scientific standpoint, your leaving the room is punishing your bird’s screaming behavior. Similarly, if your dog begins barking and pawing in his crate when you walk away; after which, you turn around to tell him quiet AND his behavior continues when you begin to walk away; then you know your turning back around and talking to him (attention) is reinforcing his pawing/barking.
Let’s break this down even further.
Within Reinforcement, there is Positive (R+) and Negative (R-). R+ means something is added to the environment that is of value to that animal. To that animal being the key words. You may think you are giving your pet (or child or co-worker for that matter) positive reinforcement by giving it a treat or toy, but if the rate of that behavior is not increasing, then by definition it is not a reinforcer to that learner. (Please see my post on the difference between reward and reinforcement.) Let’s say for example, you are teaching your dog to come to you and when he gets there you give him a nice head rub while telling him he is a good boy. BUT he begins coming LESS to you when you call. Then, you are actually not reinforcing his recall behavior. In fact you are by definition punishing it. If, on the other hand, you give your dog a nice piece of meat or take out a favorite tug toy when your dog comes, and he comes MORE often and more quickly with a tail wagging, then you know you are giving an R+ consequence.
Looking at a scenario with a bird, if your bird begins to scream when you pick up the telephone; and the consequence is your putting the phone back down (which inevitably you will need to do eventually); and your bird screams MORE when you answer the phone, then we know that your bird is receiving R+ for screaming when you answer a call. (Note: Important to solving this behavior problem then is figuring out what that consequence is that is maintaining the behavior in that circumstance.)
R- on the other hand means something that the learner considers aversive is removed from the environment immediately following a behavior. If, after your dog growls at you, and you back away, then your dog is being reinforced negatively for growling and he will growl more when you approach him (probably at a specific time or circumstance). His growling worked to get something aversive, which was you, to back away. (Please read my post on what you should do if your dog growls.) If, an approaching stranger takes steps backward after a bird runs to the back of his cage; and that bird runs to the back of his cage more when someone it does not know comes close, then that bird’s behavior is being negatively reinforced.
Positive Punishment (P+) refers to a consequence that is adding something to the environment that is aversive to the learner immediately after a behavior. If you jerk your dog’s leash or give him a shock using an E-collar when he pulls on a leash; and he pulls on the leash less in the future, then you are using P+.
If you wobble your hand when your bird (who is standing on it) puts his head down to nip at you, and your bird puts his head down less in the future, then you are also using P+. You have added something unpleasant to your pet’s environment.
Negative Punishment (P-) refers to a consequence that is removing something positive from the environment to decrease the probability of a behavior. When you turn your back and walk out of the room when your dog barks in his crate or your bird screams in his cage, if the barking and screaming behaviors decrease in the future, then you have used negative punishment (assuming that your dog and bird see value in your presence).
Learning and understanding this, you realize that using positive reinforcement 100% of the time would be difficult to do; however, I make the commitment to use the most positive and humane strategies possible to modify and manage behavior.
The Humane Hierarchy is a ranking of training methodologies going from least intrusive for the learner to most intrusive with Level 1 being the most socially acceptable and giving the animal the highest amount of control. How do you think those quadrants rank in terms of most positive to the learner? Please click here to find out and learn more about the Humane Hierarchy.
I was working with one of my puppy clients the other day. And, as many young puppies do, without hesitation she grabbed my tug toy when presented, she ran after a toy I threw and she watched with her tail wagging as I showed her a toy that made strange noises. She also did not hesitate to eat any of the treats offered. Confidence definitely is not lacking in her.
One thing I think about when I teach an animal like her (and her human) is how wonderful it is that she is showing me SO many possible reinforcers for teaching her behaviors.
Her owner had a question for me about rewarding behaviors. It wasn’t anything that I hadn’t been asked before but it got me thinking, how often it is that I hear of people make reference to rewards.
People often think of the words rewards and reinforcement being interchangeable but their interpretation can be different. So, I figured I’d write a post to give some clarity to it.
Let’s look first at reinforcement. Paul Chance’s definition (Learning & Behavior) is: the procedure of providing consequences for a behavior that increase or maintain the strength of that behavior.
That is very important to understand because behavior, simply put, is a tool for an animal to get a consequence. It is feedback that the animal uses as to whether or not that behavior should continue in the future.
In training, it is also important to note that the delivery of that reinforcer can greatly affect its effectiveness. Dr. Susan Friedman taught me that contingency occurs when the presentation of the reinforcer DEPENDS on the performance of the behavior. (If behavior X occurs, then consequence Y will occur.) An example of this is *if* Sam runs to his bed when given a verbal cue, *then* he gets a piece of chicken or *if* Sam walks by my side, *then* I will run with him to the grass to sniff.
The greater the contingency, the faster the learning curve, and that occurs with consistency in pairing the behavior and consequence. Contiguity is the amount of time between the behavior and the reinforcing consequence. The shorter the delay, the faster the learning process.
This is why moment markers (marking a specific behavior with a click, verbal word or something else specifically when the behavior occurs and following it with a reinforcer such as food) are so effective because they provide the learner with such precise feedback. The moment marker serves to tell the animal that YES, that is the behavior that is earning you reinforcement.
Reinforcers can be negative or positive. Negative reinforcement are consequences that are removed, avoided or escaped in the environment while positive reinforcement are stimulus added to the environment and are consequences the animal behaves to get. What they share is their impact on the future rate of the behavior, to either increase or maintain the behavior’s strength.
What we as trainers and teachers also need to understand is that reinforcement is absolutely the study of one and it can change from moment to moment. A hungry animal may have more value for food, and especially by food that it is not part of its everyday diet. A puppy in the morning may have more value for active play. Environmental reinforcers are all around too such as opportunities to sniff.
The proof of a reinforcer’s effectiveness is measured by the future rate of the behavior. That is key.
On the other hand, a reward by definition is something given in recognition of one’s service, effort, or achievement. However, what is important to note is that rewards are ONLY reinforcers when they increase or maintain the strength of behaviors.
A common mistake is when people *reward* behavior based upon what they ‘think’ the learner should value. As an example, I often see people reward their dog for coming by bending over to fluff up the fur on their dog’s head only to have their dog move backwards. There is a high probability that the recall behavior could break down instead of increase if the dog learns that something aversive will happen when he comes. By definition then, that reward is not a reinforcer but a punisher.
If people reward their dog for a behavior with a cheerio or verbal ‘good boy’ in a soft tone, and their dog’s fluency in that behavior weakens, then by definition that reward is also not a reinforcer.
The take-a-way here is to remember, if you want to teach and strengthen a behavior that you want your pet to do, make sure you are following that behavior with a consequence that is of value TO YOUR student.
It is not uncommon for people who share homes with a dog to complain about their furry friend bursting into a barking frenzy as a response to seeing or hearing something outside the window. Understandably the noise can be really annoying to human ears, especially when it comes at inopportune times.
There are so many reasons why dogs react to stimulus by barking, panting, running side to side, and have accelerated heart beats. It could be territorial or fear or barrier frustration, or for herding dogs, it could even be due to your dog’s instincts to herd.
If this happens on a regular basis, you may want to take steps to modify your dog’s behavioral reaction.
Here are a few suggestions. Please note that behavior is always the study of one, with differing environments, stimulus and animals. These are some general considerations to think about.
Let’s look at this behavior modification plan from the standpoint of the Humane Hierarchy, a ranking of training methodologies going from least intrusive for the learner to most intrusive. For more on this, please see my post.
If you have ruled out a medical or nutritional variable, then let’s begin with antecedent arrangement. What can be done to manage the environment so that a) your pet will not have access to practice that unwanted set of behaviors and b) your pet will have less motivation to do the unwanted set of behaviors.
Remember, practice strengthens behaviors and if your pet has access to seeing and hearing those outside stimulus when you can not be in teaching mode, you will be setting your pet up to keep practicing and getting reinforcement from those behaviors.
If your dog is barking at what he sees out the window, then consider blocking access to that window. Drawing the curtains may not be the solution as curtains can be moved by a pushy nose. Some suggestions are preventing access to the window or applying a cling film (that can be easily removed) to the window (purchased at a home supply store).
If your dog reacts strongly to outdoor noises, playing white noise or a radio may help.
As a motivating operation, if you increase your dog’s mental and physical exercise, you will be making resting more valuable to your dog. Think in terms of exercising your dog’s mind and body through training, thinking toys and games.
In a controlled setting, when you are fully focused and in training mode, you can teach your dog behaviors you would like to see in him when he sees something outside.
A friend of mine saw trouble ahead when her neighbor began letting two dogs out to run and bark on the other side of the fence. Karen first saw Baxter, who is a certified therapy dog, running back and forth and she anticipating the barking that would come next. In that moment, she averted his attention quickly and then started putting together a plan. With high value treats, she began teaching Baxter that the cue ‘doggie doggie’ was for alerting (turning his head to look at them) to the other dogs, and then running to Karen for something awesome. She began teaching this inside, behind a window where Baxter could succeed before moving to outside.
What Karen was doing was very similar to Leslie McDevitt’s Look At That Game beautifully detailed in one of my favorite books, Control Unleashed. In a very simplified description, Look at That teaches your dog that *when* he looks at a stimulus, *then* something awesome happens like a pretty tasty treat getting delivered by a well liked human. As your dog’s teacher, playing this means being in a location and at a time when your dog will not be over threshold (in other words BEFORE the lunging, barking behavior begins). Begin by teaching your dog to look at something more neutral. As soon as your dog notices the stimulus, then you mark that behavior such as with a verbal Yes! or a click, and then follow it with a high value treat. As you have continued success, you can first move this game to a more distracting environment, and then a more distracting stimulus. (Leslie recommends teaching this with a cue.)
As for timing, remember, for your dog to learn that one stimulus (in Karen’s case, the barking dogs next door) predicts another stimulus (tasty food), then the dogs barking must come before the tasty food. Marking the very moment your dog sees the stimulus is very important as the quicker that consequence occurs after a behavior, the easier it is for an animal to build the association between behavior and consequence.
The management portion of this plan is important because if you allow your dog to practice reacting to his environment outside of your limited training time, you will make behavior change very difficult.
As for Karen, with enough practice, instead of barking back to the dogs next door, when he hears them outside, he runs to find his housemate.
In your house, you can practice classical conditioning without the cue as well. In a controlled learning environment and without the cue, practice having your dog see stimulus such as people walking to your door or riding a bike down the sidewalk and immediately follow that with giving your dog a super tasty treat. (beginning this at a distance from the window or with people at a distance from your house where your dog will not begin barking and progressing only at the pace at which your pet can succeed at remaining calm – having relaxed body muscles and normal heartrate). The changes you are seeking are internal, involuntary responses. As trainer Kathy Sdao says, the emotional response to the second stimulus infects the emotional response to the first stimulus. You can do this exact same process only with sounds instead of visual stimulus.
To see a fun example of the effectiveness of classical conditioning, please watch this video.
On a last note, remember, if your dog jumps and barks at the sound of your doorbell, chances are pretty likely he has learned that jumping and barking at the doorbell causes the door to open and either – incredible people or scary creatures to walk in. That is another lesson for another day.
Hundreds of dog trainers (actually probably upwards of close to 1000) from across the country converged on Covington in October. It was the very first time that the Association of Pet Dog Trainers (APDT) has held its large annual conference here and I was fortunate to be among the attendees. APDT is a professional organization of individual trainers who are committed to becoming better trainers through education.
Steve White was one of the presenters. Accredited as a master trainer by the Washington State Police Canine Association, Steve has served as vice president of the Certification Council for Pet Dog Trainers, and has been an instructor for the K-9 Academy for Law Enforcement. He specializes in teaching behavior modification, tracking, and scent work through the use of positive reinforcement-based operant conditioning.
In his talk, he reminded us about the danger of using punishment in our training. He even went so far as to compare aversive-based training with a nuclear war.
“If the blast doesn’t get you, the fallout will,” he said.
How about that for making a point?
I’ve seen it all too often…a dog whose fear issues have only heightened, who has become aggressive or disinterested in learning OR a bird that cowers in a corner of its cage or gives a nasty bite because an owner used a form of punishment to stop an unacceptable behavior.
Dr. Susan Friedman has written extensively about the topic. “We are virtually surrounded by punishing strategies to influence behavior. Fines, penalties and reprimands whirl around us like leaves in a storm,” she had written. “Unfortunately every time an animal responds to punishment by doing something less often, the person who delivered the punishment is rewarded.”
So, how do you define punishment?
Functionally speaking, it is any consequence to a behavior that reduces its frequency or intensity – but that is a very individualized. What one animal may see as aversive, another may find rewarding. Uh oh!
And punishment is just an umbrella term for a number of strategies. A mild strategy is withdrawing or removing something from the environment such as each time your bird chews on your ear, you gently remove him from your shoulder. No rough handling is ever needed – just immediate removal followed by an opportunity to do it right.
Another example of a mild punishment is to ignore a behavior – but that can be too difficult (for us not our pet) to do properly and affect change. One of the reasons so many birds have become loud and incessant screamers is because their owner ‘tried’ to ignore the behavior but couldn’t. And, it is only after the bird has increased the intensity of the behavior that the owner finally gave in with attention (remember – whether something is aversive or not is in the eyes of the behaver). Then guess what? The owner inadvertently reinforced the louder, more obnoxious scream. Ugh!
Ignoring an unwanted behavior while also hugely reinforcing another acceptable behavior instead is a much more effective approach. In fact, it is that approach that I learned over 11 years ago that not only saved my relationship with Barnaby but got me hooked on the study of behavior. (I have the story of my behavior plan in solving Barnaby’s screaming on my pet behavior blog.)
I won’t detail the more aggressive strategies of punishment here because I’d rather not focus on them. I’d rather focus on the fact that we can affect behavior change in our pets AND foster a love for learning if we switch the way we look at it. Instead of working to get your pets to stop doing something, work to teach your pet the behavior that you would like to see him do instead…then reinforce the heck out of it.