< A version of this article was originally published in the APDT Newsletter NOV/DEC 1999>
I knew of Marian and Bob Bailey but had never heard them speak until the Association of Pet Dog Trainers (APDT) Conference in Valley Forge in the fall of 1998. After spending an hour with them and their chickens, I realized there was a great deal that the Bailey’s could teach me. I knew they were having a 5-Day workshop in the summer, and like an eager retriever waiting for dinner, could not wait to sign-up.
When I told my family, staff, friends and clients that I was going to Arkansas to learn to train chickens, everybody laughed; “You can’t be serious,” they said. I was serious and off I went. In the process of learning how to train chickens, I became a better dog trainer!
Between the two of them, Marian and Bob Bailey have over 100 years of training experience, with more than 15,000 animals, across 140 species, from cockroaches to killer whales. I doubt any living person has more knowledge of using operant conditioning to train animals.
With all of their expertise, the Bailey’s could easily be prima donnas. They could be animal training gurus but all they desire is to share their knowledge and experience with those who wish to learn. Those of us who have experienced the Bailey’s workshops are fortunate that Bob and Marian would rather be teaching than enjoying an idyllic life in their lakeshore home. I believe anyone who trains animals, no matter how skilled or experienced, would learn a great deal from Marian and Bob, about training animals and about teaching people.
So what do you learn when you go to the Bailey’s Operant Conditioning workshop, or as I call it “Chicken Camp?”
- How to wrangle chickens (something new for us city folk, and necessary if you want to train a chicken).
- Training is a mechanical skill. Training is a mechanical skill. Training is a mechanical skill, (something Bob and the chickens never let you forget, ever).
- The importance of Timing, Criteria and Rate of Reinforcement (this is the key to it all!).
- Training is simple but not easy, and is most effective when planned and executed in a scientific manner (most trainers make things more difficult than they need to be).
- Animals can be trained to reliably perform complex behaviors, exclusively with positive reinforcement. In the 15,000+ animals trained by the Bailey’s, they used positive punishment no more than a dozen times.
Having only handled a chicken once before, I was a bit tentative about reaching into that cage to get my chicken. Feathers do not feel anything like fur, and while puppies tend to lick, the chickens tend to peck. Especially if they know you are intimidated. Actually, I was lucky compared to my friend Carolyn Clark who was affectionately calling her chicken Fang by the end of the first day. Fortunately for me, I became good friends with my chickens by the end of the week.
Bob kept telling us “You are bigger, stronger and smarter than the chicken” and eventually we started to believe him. (Note he never said we were faster). We were handling “experienced” birds, ones that had been used at several chicken camps, as well as “rookie” birds that had no prior chicken camp experience. These chickens were raised in a barnyard and were a bit wild. We were well matched with the rookie birds because on that first day we did not know any more than they did. Our first goal with the rookies was socialization and getting them accustomed to being handled. When you think about it, they really are not that different from puppies.
The experienced chickens, however, came with baggage. Not Samsonite, but training mistakes, inadvertent or advertent, which had been made at previous camps. Bob explained that they do chicken camps at universities and you can only imagine what students will train a chicken to do. Working with these experienced chickens was not unlike many of the shelter dogs we all see. We could train them new things, but also had to do some un-training.
Training Is A Mechanical Skill
The workshop was a mixture of lectures and hands-on chicken training experience. The first couple of days were spent on developing the timing and mechanical aptitude necessary to be an effective chicken trainer. Chickens are extremely fast and you must become almost as fast as the chicken.
To be the most effective you can be, a trainer needs to be able to observe the animal for behavior, mark the behavior precisely, deliver the reinforcement in a timely manner and be ready to do it again and again, rapidly. The concept is very simple, but it is not as easy as it seems.
You develop these skills like any other mechanical skills, with practice. When you are rewarding chickens with feed from a cup, you must have a nice, fast fluid motion that does not scare the chicken and does not spill feed all over the table. The act of clicking and delivering the reward must become an automatic response on your part so that you can concentrate on observing behavior and monitoring the animal’s response. It is difficult to count pecks per minute when you are concentrating on clicking and delivering the food.
Bob helped us develop these skills much like we might train a dog, shaping in small increments. We first practiced just moving the cup to present the food, no clicking and no chicken at this point. Then we added food to the cup which added a whole new dimension. If you move the cup to fast, you spill food all over the table. Next we added the chicken and started presenting food but still no clicking. Finally, by the end of the first day, we combined all of the steps and clicked and presented food to our experienced chicken.
While most of us who train dogs already have good mechanical skills, it is something lacking in many of our students. Spending some time working on these skills, in addition to having them work with their dogs, would be beneficial. It is something I believe we need to spend more time on in our classes. The use of a video camera can be a great aid in developing timing skills.
Timing, Rate of Reinforcement and Criteria
The Bailey’s believe that most problems in training, (where the trainer believes the animal is just “not getting it”), are actually trainer problems related to timing, the rate of reinforcement or the criteria being used. I think we all understand the importance of timing so I will not go into details. As indicated above, the best way to improve your timing is to practice. It is a mechanical skill. It also involves understanding the animal so you can anticipate the behavior and be ready.
Based on their experiences, the Bailey’s believe it is essential to keep the animal’s rate of reinforcement at a high level when training. The animal will match your pace, and if you work quickly, so will the animal. If you work slowly, the animal will slow down and be at increased risk for being distracted. Training is all about repetitions, and the faster you get the repetitions, the more efficient you will become.
Bob emphasized that timing of the delivery of the reward is also very important. While the clicker buys you some time, delivery of the treat needs to be prompt, particularly in the early stages of training. For example, when teaching a dog to heel, the treat needs to be given while the dog is still in a heel position. If the delivery of the reinforcement is delayed too long, the animal may exhibit an undesired behavior and may associate the reinforcement with the wrong behavior, even if you clicked at the right time.
The Bailey’s emphasize that training works better when you have a clear idea of the behavior you want and the steps it will take to get there. It is important to break a behavior down into small incremental responses, splitting, something many people find difficult as opposed to grouping several response together, lumping. Bob also advocates planning a training session and committing that plan to paper before you start, but more on that later.
When you specify a criterion for a training session, make sure it is broad enough so that the animal can be successful. For example if we are working on the heel position we might start by clicking and treating every time the dog is within 3 feet of our left leg. When 8 out of 10 training trials are successful, it is time to raise the criteria for the next training session. You want to keep the range of acceptable criteria sufficiently large so that you can make a steady progression forward. When the animal only meets the criteria twice out of 10 trials, learning is not taking place. If you have to backup and lower the criteria because you moved too fast, you are wasting time.
Training Is A Science
The Bailey’s approach to training animals is as a science and as a business. The methods they used in their business, Animal Behavior Enterprises, had to be effective, efficient and reliable. They strongly advocate planning and maintaining data on each session. I will be the first to admit that I do not document each and every training session. I have a clear goal when I start, but then tend to “go with the flow.” I believe my style works, but am convinced that I need to give Bob’s way a try. I do believe that keeping track of my chickens response rate did make us progress faster. I am hesitant to ask students to do it, because I think maintaining records may be too overwhelming, but it is something I will be trying with my own dogs.
UPDATE: I still don’t write a training plan as often as I should; however, in the fall of 1999, just a few months after my first chicken camp, we rescued a Cairn Terrier that we would name Dulcie. It was my intention to take her through our training classes like I would with any other dog. On the first night, it was clear that she was fine in a room full of dogs and people but she was not going to be lured into a sit like the other dogs. It turned out Dulcie was incredibly hand shy and any attempt to get hear her head with my hand, necessary when luring, resulted in her rapidly scrambling backwards to safety.
My next thought was I’ll just capture her sits when we’re hanging around in my office or the house. It turns out Dulcie was not a “sitter.” She was either in motion, standing, or lying down comfortably.
Dulcie presented me with my first opportunity to develop a training plan to shape her to understand the “sit” behavior. I started by using the clicker to teach her a simple touch behavior so that she could learn the meaning of the “click” and the reinforcement. When I was able to put that on a cue, I started shaping a sit. Within two months’ time and a total of 47 sixty-second training trials (we did not work every day) Dulcie had a reliable sit cue in multiple environments, with and without distractions with me in any position. My total training time to get a really reliable sit behavior, less than one hour. Developing and following a plan, keeping records or of progress or lack of progress, and adjusting the plan according, worked extremely well. – djh, 19JUL15
Bob also believes we need to avoid “trying to get into the animals head.” Since we cannot really know what they are thinking, there is no point in wasting time considering it.
Schedules of Reinforcement
The Bailey’s firmly believe, and they base this on many years of experience, that the use of anything other than continuous reinforcement is not necessary for most pet dog training. This runs contrary to what many other clicker based dog trainers say and do. The position many take is that variable reinforcement is necessary in order to get a strong, reliable behavior. I must admit, we used continuous reinforcement to train our chickens and when it came time to extinguish those behaviors, it was not easy. Continuous reinforcement can build very strong behaviors.
I need to give this some more thought before I decide how to handle it in my classes. Reinforcement schedules are not an easy subject for many students to grasp, and not talking about them may make training easier. I chatted with other trainers at the workshop about how they handle this and the consensus was that when students try to do variable reinforcement most do not do it very well. However, when they are not trying, nature takes over and they start doing it without realizing it. Perhaps that alone is sufficient.
UPDATE: Over the years we have learned that if we teach continuous reinforcement to our students they will eventually move to a somewhat schedule of random variable reinforcement on their own. Helping students to understand the value of reinforcement and that their dog will not do things “just to please them” is still one of the biggest obstacles in teaching. For so many years people have heard that “dogs should work for praise” which essentially means work for free, that they are very hesitant to reward at all, much less frequently enough or with a high enough value reinforcer. Fortunately, they’re starting to get it. – djh, 19JUL15
R+, P-, P+ and R-
No workshop on operant conditioning could be complete without a discussion on the various aspects of reinforcement and punishment. Over the past couple of years we have seen the definition of a clicker trainer evolve to mean a trainer who uses the positive reinforcement and negative punishment aspects of operant conditioning. Many of us specifically choose not to use positive punishment or negative reinforcement. While Bob freely admits he has used all aspects of operant conditioning, he clearly believes that the positive punishment should only be used as a last resort, and must be applied carefully. If you have to do it more than twice, you are not doing it properly.
Bob told us that of the 15,000+ animals he and Marian have trained they have only used positive punishment a dozen times or less, and that was at the insistence of the client. They felt the use of punishment was not necessary. I think the Bailey’s record should convince the most obstinate doubters that positive punishment is rarely necessary. Bob served as the first Director of Training for the U.S. Navy Marine Mammal program and accomplished some absolutely amazing feats of training, many of them still classified. If a dolphin can be trained to do a 30+ mile directed retrieve, it should not be necessary to use punishment to train a dog as a pet or even for competition.
So, would I go to chicken camp again? You, bet I would! I think the most seasoned, experienced dog trainers in the world could learn a great deal from the Bailey’s, as long as they are willing to learn.
UPDATE: Between 1996 and 2014 I’ve attended over 1447 hours/ 180 days of training on animal behavior and training. Out of all those hours the most valuable were: 1) my two weeks at Chicken Camp with Bob and Marian Bailey, 2) my one week at Wolf Park with Dr. Erich Klinghammer, Dr. Ray Coppinger, Terry Ryan, Ken McCort and Pat Goodmann, 3) my several sessions with Turid Rugaas, 4) my week at the Shedd Aquarium with Ken Ramirez, and 5) my many days of training with Heather Simpson at the Natural Animal Centre. I mention these opportunities here with the advice, if you are on the fence about chicken camp…. DO IT! – djh, 19JUL15
©2015, Donald J. Hanson, All Rights Reserved <Click for Copyright and Use Policy>