Boundary Training with Positive Reinforcement: Stay Out vs. Don’t Come In!

Semantics: what is the difference between training a dog to stay out, vs training a dog not to go in somewhere? It’s the difference between training a dog what To Do vs training a dog what Not To Do. With positive reinforcement we focus on What We Want. We can reinforce a dog for staying somewhere.  We can be clear about where it is ok for him to be. That is easier for him to understand than Don’t Do “x”.  The usual approach for expressing “don’t” is punishment, keeping in mind that the behavioral definition for punishment is something which makes it less likely the animal will repeat the behavior in the future.  With “don’t”, that’s the goal: don’t do that again.  So whether we chase, scold, yell or use any other method to try to stop the dog from repeating it, then by definition, we are using punishment. For those of us who prefer to come up with ways to use reinforcement whenever possible, we need to change the way we think and train.

When the service dog was visiting us, I had two opportunities to teach him to stay out of certain areas.  The first area was the kitchen.  He’s a Big Dog.  I like my dogs to stay out of the kitchen so I don’t trip over them while I’m cooking.  Big Black Dog was simply too big for me to even walk around between my island and counters! Plus, he could rest his head on the table, and with the slightest tilt of his head could rest his chin on the counters to, oh say, sniff what was up there.  He was very good about not IMG_4928taking anything, but I still thought it would be a good idea for him to learn to stay out of the kitchen.

So, where did I want him instead? I got out a mat and set it on the floor on the opposite side of the island.  It was well out of the way of foot traffic for cooking, cleaning up and serving.  I could foresee no reason he would not be comfortable and safe there at all mealtime prep and cleanup (while we were actually eating, he lay next to me and/or under the table so as to work on those skills). I used a hand target to position him onto the mat, and then asked him to lie down there, which, being a Very Good Big Black Dog, he did.  I clicked and reinforced.  I cued him to “stay” (which he already knew) and walked back to the kitchen.  There were sufficient distractions, not the least of which was a Little White Dog, who upon seeing a mat hit the floor was on it.  She knows all about the reinforcement for mat work.  Fortunately there was room for both of them and he’s a generous sort who didn’t object to her company.  Other distractions were the smells of cooking, the crinkle of food wrappers, and the movement of people. So initially I reinforced often.  As soon as I walked to the kitchen, I clicked and returned to treat (both of them).  I returned to the kitchen area again and did a few seconds of food prep, clicked and returned to them for another treat.  It was obvious pretty soon that he was fine with this arrangement.  Again, he’s had amazing foundations.  I was able to stretch the time to clicking after several minutes and each night could stretch the time further.  I did occasionally keep the duration short to keep the behavior strong, but by the end of our week together, he was ready to lie down on the mat as soon as I got it out, and stayed there reliably, along with his sidekick.

The other opportunity for teaching him to “stay out” was more complex.  One of my most important tasks for the week was to acclimate him to life around horses and this meant learning about life in the barn. I do have a dog kennel in my horse barn and when horses are about, my dogs go into the kennel to keep everyone safe.  But when horses are turned out and I am doing chores, I am fine with the dogs being loose. However, to keep lines from blurring, I decided that Big Black Dog should learn from the start to stay out of stalls. This was a serious safety issue for dogs, horses and people. I did not want to tell him he had to stay somewhere specific as he had in the kitchen example.  I wanted him to be free to explore the aisle, find a sunny spot to lie in or the best place to catch a summer breeze (come time for summer breezes). I just wanted him to stay out of the stalls.

The distractions were enormous. He thought horse manure was the best thing ever and he couldn’t quite believe it was free for the taking everywhere he looked…in the stalls. Step number one in setting him up for success was to close the stall doors unless I was ready to actively train. That prevented him from making any mistakes. This was management and I could have locked myself in each stall as I cleaned it to just use a management solution. But that wasn’t training. I wanted him to learn that even when stall doors were open, he needed to stay out of them. Therefore, I had to make the aisle an appealing place to be.

I did that with treats. Not only did he get to eat them, but he got to chase them down and sniff them out. He gets all his food as reinforcers. I had an entire cup of kibble in my treat pouch for morning chores and I used it! I began in the aisle, just tossing a piece of kibble on the floor to start the game. He ate it and as he was finishing crunching, I’d toss another.  He liked that game.  I opened a stall door and tossed another piece which he chased down.  I stood in the stall and tossed treats out into the aisle each time he looked at me. First to the left, then to the right. This was a great way for him to get a little exercise in the bitter weather, back and forth up and down the aisle.

IMG_4927A couple more tosses and I picked up my pitchfork.  This time I tossed three pieces of kibble far down the aisle, which gave me just enough time to scoop one manure pile into the wheelbarrow. I tossed another three pieces the other direction. The kibble would roll and bounce which sometimes took him a minute or two to find, giving me more time to muck. Sometimes I’d toss 5 or 7 pieces, but I always made sure that as he was approaching the stall, he’d be met with kibble to hunt. Pretty soon he was coming to the stall door and stopping himself, waiting for the next toss.  Success.  This was the critical point. I never had to say a word or stop him.  Simply expecting the next kibble toss was enough to stop him at the stall door.

Now I could wait a second or two before tossing.  I had a whole week, so I was in no hurry. Each morning we played this game and before long, he would sit outside the stall as I worked, only occasionally tossing kibble. Oftentimes, I’d toss a whole handful.  That kept him busy longer and allowed me more time to work.

After a few days, I decided the behavior was strong enough to introduce some more distraction. After cleaning one stall, I left that door open as I went to the next stall.  Granted, the stall was clean, but there were still many edible bits in there.  Big Black Dog paid no attention. He was waiting in the aisle for kibble games. I left that stall open as I moved on. No problem. He was focused on me (yay!) and ignored the temptations in the stalls. After a few more days, I left the stall across from me open even before I’d cleaned it. That way I could easily see if he slipped into it when I might otherwise think he was still hunting kibble. When he looked into it, I simply said his name and tossed kibble. Temptation averted. By the end of the week, I could do all my chores with all stall doors open and he never entered. Why would he when all the reinforcement was in the aisle?

Good Big Black Dog.





Pay It Forward: Reinforcement Is Not Bribery

IMG_2946Many times when I do an evaluation on a new dog, the owner will tell me “she’ll come if she knows I have treats”, or “if I shake the box of dog biscuits, he’ll come running”.  But, they say, the dog will not come when called, unless they know the goodies are waiting. This causes several concerns with dog owners regarding using positive reinforcement to train.  They don’t see how positive reinforcement is any different than what they are already doing and they think that if they don’t have the treats, the dog will ignore them.

This is a hard concept to break through when we do use food as a reinforcer. When I talk about using high value reinforcers depending on the difficulty of the behavior, they think that they need to show the dog that they have hot dog bits in their pouch before they step outside so the dog will pay attention.

The opposite is actually true. We do not want the dog to know ahead of time what, if any, reinforcer is available.  Shaking the treats, showing the cheese, making a show of putting on the treat pouch: these are all forms of bribery.  Effective use of positive reinforcement is not bribery.  The proof of that is a dog who has been trained with the correct use of treats as reinforcers.  These dogs come running every time, from a cue, whether it’s verbal, visual or other.  They don’t look to see if the person has worthwhile treats- they come immediately, with great speed, and directly to the handler. 

The reason positive reinforcement works so well is in the definition of the word reinforcement itself.  In behavioral terms, reinforcement is something which makes it more likely the behavior will happen in the future. So this past week I found myself pointing out to clients that the treat they deliver has no effect on what the dog just did- the behavior is over!  But it does have an effect on the next time you ask the dog for that behavior.  If you call your dog, without any bribery, and then are able to reach into your pocket and pull out a piece of cheese to give him, that may make it more likely he will come when you call in the future. He did not even know you had cheese in your pocket, but you did!  You want to be full of happy surprises like this.  Every single time your dog does something you like, you want to be able to surprise her with something she likes.  This is called building a reinforcement history.

When you have a strong reinforcement history, meaning you have a reputation with your dog of “paying well” for behaviors, your dog will begin to respond to cues more rapidly and with much greater reliability.  We refer to reinforcement history as “putting money in the bank”.  You want to have plenty in the account so that when you need to make a withdrawal, the reinforcement account is flush.  When your dog responds to your cue, for instance when he comes when you call, you are making a withdrawal from that “trust account”.  The level of difficulty of the behavior will determine how much is withdrawn.

In the house, getting ready to go for walk and you call your dog, well that might even be free because the dog wants to go out with you!  The opportunity to go outside is payment enough.  In your yard with no distractions and you call your dog, that might cost you a little bit.  Dogs love to sniff and smell things so while it may look to you as if it’s an easy thing, really he had to give something up- the pleasant occupation of sniffing around- to come to you.  So you just made a small withdrawal from your account.  But you can replenish that easily by giving him a treat when he returns.  The balance has been restored. There is enough in the account so that next time you call, he will respond.

The challenge, of course, is when the UPS man pulls in the driveway.  Or at least that’s the biggest challenge at my house.  That UPS truck is the source of so many distractions: movement, intruder on the property, friendly guy in the truck, who gives dog biscuits, and oh the smells on his tires, his clothes, the packages.  I don’t know about you but my dogs act as if the property is under attack when he pulls in and then switch to “here’s Santa Claus!” mode when he stops.  They are in that truck checking everything out.

For me to be able to recall them in that situation, I have to have an account that is full to bursting.  The only way to get it that full is to put lots and lots of little deposits in over time.  When we are outside, whether it’s to go to the barn, or to get the mail, or to go for a walk, I make sure I have low and high value treats with me.  For my dogs, examples are kibble, cheese and hot dogs.  I cut them up in tiny bits (smelly hot dogs go in a tiny plastic container with a lid), I hide them away in a pocket before we leave so the dogs never know what’s there. And then we practice.  I call them when nothing is distracting them and pay with kibble.  I call when there is a mild distraction (sniffing in the woods) and pay with string cheese.  I call when a car is coming and I pay with hot dogs.  My dogs don’t know what I have, but they know that it will be good.  Each time I give a treat, I am putting money in that bank.  Every day I put more and more in- if I’m cheap and only put in kibble each time, that might keep the account level…but will it put enough in so that when the UPS man comes, I have what I need?

I never know when the UPS man is coming, so I can’t slap a steak in my pocket for that particular occurrence.  I can, however, pay it forward.

Keep treats on hand and on you all the time.  You want to pay for the easy stuff, so that the account can cover the cost.

Using Management While Clicker Training

Beetle in his mud boots
Beetle in his mud boots

One of the first things I tell clients is that both management and training are essential to living happily with a dog.  There is no magic wand to wave, and so unless I want Beetle, pictured at left, to go into my house with that mud, I need to MANAGE the situation.

Beetle has not been trained to wipe his feet, or hose himself off, or that when his feet are muddy, he must wait for me to clean them before going into the house.  If I neglect to shut the mudroom door, and he trots his muddy feet into the house, then I need to go to that special place in my house for banging my head against the wall (which is the only appropriate response when a dog does not do as I had hoped).

It is not fair to yell at him as he trots through the house (I would argue it’s never fair to yell, but that comes later).  He doesn’t know better.  I used that phrase- “he knows better!” for a long time, but I have no idea why I thought a dog would know anything other than what I taught him.  And if I never taught him to stop and wait at the door, then why would he?

I know many people who train their dogs to wait until their feet are wiped off before going inside.  I could do that as well, but I would have to go through a training process (no wand waving) to teach that.  I would need to train that long before mud season if I expected it to be a solid behavior in mud season.  I would need to hold myself accountable if the dog did not wait; if he doesn’t wait, then I haven’t done a good enough job of training it.

I would also need to make sure I reinforced the dog (with a click and a treat) for waiting at the door even if he wasn’t muddy.  I can’t let him walk into the house without waiting on dry, sunny days and then be upset if he doesn’t stop when his feet are muddy.  I think that’s where some people get frustrated thinking that the dog “knows better”.  THEY know that they always stop the dog to dry off his feet when it’s muddy so they think the dog should figure it out as well.

Dogs pick up on many cues- but they aren’t always the cues we think they are.  “Muddy feet” is not a cue easily distinguished to a dog who has no problem with muddy feet and barely 286340_10151053062730925_368639312_onotices.

Today I worked with Eloise on a similar behavior: waiting until I asked her to get into the car.  She is very good about jumping into the car promptly as soon as I open the door.  It’s quite handy most of the time…but not when she looks like this.  So I realized it is my responsibility to teach her to wait until I tell her to jump in.  We practiced that many times today.  I approached the car and then opened the door a tiny bit and then gave her a click and a treat while she waited outside the car.  I shut the door and did it again.

Here I was managing her ability to get in the car.  Before, the door opening was a cue to jump in- a strong cue.  I didn’t want her to make a mistake and jump in- that would have been practicing behavior I didn’t want.  I want to practice behavior I do want- that’s what makes it stronger.  If I only opened the door a crack, she couldn’t jump in- she was bound to be successful and she was learning in baby steps.  The next time I opened the door a tiny bit further- click and treat for waiting.

I continued to increase the amount I opened the door each time, but by now, she realized she was getting treats for just standing outside the door, so the urge to jump in was gone.  Finally I could open the door all the way and she just waited for her treat outside. Then I said “hup”, which basically means jump up on something to my dogs, and she jumped in the car for which she got another click and treat.  I want both behaviors to be equally as strong.

After practicing this several times, I noticed that Eloise was hesitating as we approached the car, rather than dancing excitedly to get in.  That was information to me that she was starting to realize that every time we approached the car, she would be expected to wait.  Eventually, approaching the car will become the cue to wait outside the door…as long as I am consistent in remembering to ask her every time!  At that point, I will no longer need to go through the stages of opening the door many times, but I will still reinforce her for waiting- and jumping in.

Expo Inspiration: Which Clicker Training Project to Choose?

The annual Clicker Expo (East) was held last weekend in Virginia Beach (it’s in a different city each year).  I flew down Thursday and arrived in time to meet up with old friends and new at the reception.  As this is my third Expo, I found I’ve gotten a little jaded- I now know that the people there will be kind and encouraging, the dogs well behaved and happy.  My first year I was astonished by the overall positive feel of the place!  It doesn’t make it any less wonderful to be surrounded by all that reinforcement, however.

The challenge, after hearing lectures, watching labs and sharing stories with fellow trainers, is which projects to take on upon my return.  There are so many things I want to do!  I started off by attending Ken Ramirez’s lecture, “What a Concept: New Frontiers in Concept Training” followed by observing the associated lab where a dozen or so dogs and handlers got to try out some beginning concept training under Ken’s guidance.

Ken covered three basic concepts: Match to Sample (where the dog is shown an object and then must choose an identical object from a collection of varied items); Do As I Do (where the dog mimics either another dog or his handler); and Counting (in which the dog demonstrates an ability to distinguish the number of objects in a collection).   When I walked out of that lab, I wanted to attempt all three projects and I wanted to do them with both dogs and horses.

On Saturday, I started the morning by listening to Laurie Luck‘s lecture “At Your Service: Teaching Service Dog Behaviors”.  The lab for this lecture was at the end of the day.  In her lecture, Laurie stressed that the things a service dog does are “just behaviors”.  They can be trained just like any other behavior, once you break them down into manageable components.  In the lab, she coached many dog/handler teams through the beginning of teaching a dog the opposites “push” and “pull” and then scent detection, as a service dog who learns to alert a diabetic to low blood sugar would.

This little dog captured my attention in the Service Dog lab.
This little dog captured my attention in the service dog lab- she loved the “pull” exercise where her love of tugging produced quite a vigorous pull!

This lab rekindled my desire to do some nose work with Beetle (whose sight and hearing have faded with age, but whose nose works just fine!).

The topic which surprised me most was the lecture Hannah Branigan gave, “Obedience Competition: Break It Down to Build It Up!”. I have not previously been interested in obedience competition for dogs, but this talk opened my eyes to the many different behaviors required of these dogs AND I was hit with a realization of how useful the techniques Hannah explained would be for Dressage with horses.  My mind was just popping with connections and it’s one of the things I love about learning about training dogs- takes me right outside the box with the horses.   I could not attend the lab with this lecture (we are limited to one lab per day so that everyone gets a chance to get in to limited space), but the talk did inspire me to attend another lecture of Hannah’s on Sunday, “Prep School for Competition Dogs”.   By the end of this, I was thinking it would be fun to do some of these exercises with Eloise until the snow melts and I can try them with the horses.

On Sunday, I was able to be a coach in Kay Laurence‘s lab “The Craft of Fine Slicing”.   In this lab, Kay asked participants what behaviors they would like to teach, and she then showed them how to slice these behaviors down into the smallest beginning components- where do you start to teach a dog to back, to turn to left or right or to cross its paws?  She stressed the importance of a base position- a position from which you can begin to shape other behaviors.  She frequently likes to begin with a base position of “standing with stillness”.  This means the dog is simply standing and waiting for your guidance.  I decided that with all the glory of the things I had seen over the weekend, that was probably a good place to start.

Kay had demonstrated how she tossed treats behind the dog and then clicked as he returned while he was still standing. Many dogs have a “default” behavior of sitting or lying down- we train these early and they are pretty strong.  Eloise well knows that if there is any question, it’s a good idea to hit the ground.  Trying to catch a dog whose legs are only 6 inches long before she drops is a hard thing to do!  I was tossing treats as Kay had, but Eloise kept sliding right into a sit or down when she came back.  Kay, and others, have stressed the importance of observing and setting the dogs up for success so I watched and tried to figure out what I could do.  At one point, I tossed a treat higher than previously and Eloise jumped up to snatch it out of the air- that time she came back and stood in front of me, ready to leap up again.  Aha!  That was the way I could keep her on her feet.  I tossed the next ten treats up, she jumped up to catch them and then stood waiting for the next one.  I could then click as she stood, even withholding the click for a few seconds as she continued to stand, and then toss another treat.  Ten clicks, ten stands.  We have a start.

I still have plans for the other projects.  I also began Beetle on the nose work this afternoon.  But one thing at a time.  There were many other lectures I attended which addressed more nuanced topics of how to go about training- with luck, I will retain those thoughts to be woven into my daily interactions.

Standing with stillness
Standing with stillness