CLICKER CONFUSION
By Deborah Jones, Ph.D., 2002
A recent discussion on the Clean Run e-mail
list concerned problems that people were having with the application
of clicker training. It developed into a thread on some of the
more common areas of confusion and misunderstanding regarding
this type of training. While clicker training is not a ‘new’
idea (B.F. Skinner wrote an article on ‘How to Train Animals’
in Scientific American in 1951) it does seem radical and unusual
to those more familiar with traditional dog training methods.
In making the change to clicker training, people are starting
to become frustrated and confused, especially when different trainers
have their own unique versions of these methods. In thinking about
my own evolution as a clicker trainer, I would like to offer a
bit of background about my journey to this point. It wasn’t
quick or easy, but it has been fun! I would also like to address
some of the myths, misconceptions, and unusual clicker training
methods that have developed.
When I was an undergraduate years ago I first
learned about the principles of classical and operant conditioning.
I spent a few years studying these principles in an academic setting,
and had a fairly good understanding of them before I ever tried
to apply them to real live animals. Then, the first animals I
worked with were lab rats. The behaviors we trained were fairly
simple, more to test out the principles than with the goal of
teaching a specific target behavior. I was excited by my success
with rats and went home to teach Foxy, my old All American girl,
how to target the door where we kept the treats. Foxy surprised
me by learning not only where to go for treats, but how to open
the door herself!
Later, I had the opportunity to work with a
number of different types of animals including a retired racehorse
and a chicken. I observed trainers using basic learning principles
on elephants and tigers with great success. My first two clicker-trained
dogs taught me an enormous amount, and I made a fair number of
mistakes with both of them. But even so, I had enough success
to be encouraged to continue. Finally, when my first Papillon
puppy came along, I had it pretty well figured out and wasted
little time with ineffective and inefficient clicker training.
But I made a whole lot of mistakes getting to that point!
I started writing books and giving seminars
because I know how few resources are out there for people who
are interested in learning about clicker training. Through my
own learning stages I saw many people who claimed to be ‘experts’
in the area of animal learning and training. Some were; many were
not. My goal was to help people understand the underlying scientific
principles so that the techniques and methods would make sense.
You can think of the theories, concepts and principles as the
bigger picture. They are the laws of learning. When you develop
specific ways to practice hands-on training, you’re using
techniques or methods that come from the theories, concepts and
principles (if you’re doing it right). For example, a basic
learning principle is that behavior that is reinforced will increase.
As a technique, I might apply that by reinforcing a jumping dog
when he has all four feet on the floor. Because I know (from the
theory) that the reinforced behavior will increase, I know that
I can fix the jumping problem without ever addressing it directly.
Instead, I simply reinforce what I want and get more of it. In
a very short time the non-reinforced behavior, jumping, decreases
drastically and is replaced by the reinforced behavior, four on
the floor.
My goal is to help people truly grasp the concepts,
not just follow directions. If you follow directions they might
work fine, but if something goes wrong then you have no idea how
to fix it. If you understand the underlying principles, you can
problem-solve. For example, I can drive my car just fine, but
if it breaks down I have no clear understanding of how to diagnose
and fix the problem, even a relatively simple problem. That’s
why I pay for AAA. The same is true for training techniques. If
you understand why and how a certain technique works, then you
can make adjustments to be more successful. Also, if you understand
the principles, then you will be able to see the weaknesses and
errors in Trainer X’s version of clicker training. People
add their own little twists to the techniques, and pretty soon
they have some totally unique method that isn’t true to
the underlying theories.
Someone once said “clicker training is
simple, but it’s not easy.” (That statement has been
attributed to Bob Bailey). This seems to me a very truthful statement.
I find it amazingly simple (after 15 years or so of practice!)
However, even if you understand the basics, there is still the
issue of putting it into practice effectively. If you want a quick
solution to a specific problem, but know very little about clicker
training, it will be difficult for you to develop an effective
way to deal with the issue. You may ask your friends and other
trainers for advice, which they will probably give quite freely.
However, if you don’t understand the basic concepts, you
won’t know which advice will work, which won’t, and
why. So, the first step is to educate yourself. You don’t
need a Ph.D. in behavioral psychology. However, a good grounding
in some simple concepts can go a long way towards helping you
make sense of clicker training techniques. The two books that
I recommend most highly for this background information are ‘Don’t
Shoot the Dog’ by Karen Pryor and ‘The Culture Clash’
by Jean Donaldson. Both of these books go into discussions of
the theories and principles, not just the actual methods.
Hopefully, my story will help you to see that
learning about clicker training is a process for both the trainer
and the dog. In many cases, the trainer is the harder one to teach!
Also, it is very difficult for trainers to move into unknown territory,
especially if they have been successful with more traditional
methods. Trying something completely new is guaranteed to be scary.
It is also likely to result in making mistakes at first. However,
the rewards (to the trainer) of learning to use clicker training
effectively are huge. It is definitely worth the effort!
Below are some common clicker training misconceptions:
“I can use my voice more effectively than
a clicker”.
In my experience I have never found this statement
to be true. People have many arguments about why they feel voice
is more effective, but the opposite is usually the case. The clicker
is better, particularly in the early stages of training, or when
problem solving, as it is clearer and more precise. With just
a bit of practice, people can click faster than they can talk.
By the time you see the behavior, form the word, and say it, the
moment has passed. Once you have learned to observe and click,
the process of observing the behavior and moving your thumb will
always be quicker. This is not to say that a verbal conditioned
reinforcer isn’t useful, but that it isn’t better.
I use my verbal when I don’t have a clicker handy, or when
I’m maintaining an already learned behavior. However, I
will always choose the clicker if it’s possible.
“I don’t have enough patience to
wait for clicker training to work”.
I have never quite been sure what people mean
with this statement. I think that this idea is based on another
fallacy, that we only get behaviors by waiting for the dog to
spontaneously perform them, which is not true. In general, people
seem to be referring to the process of shaping here, which does
involve waiting for approximations of the target behavior. However,
far from being slower, clicker training will more quickly lead
to a finished, reliable behavior. Because the dog is actively
involved in the learning process, he is likely to move to a completed
behavior at an accelerated pace.
“If I show the dog what I want he will
learn faster”.
By showing, most people mean that they will
be using some combination of luring and modeling (physical manipulation).
What this accomplishes is to give you the illusion of the behavior.
You are getting the dog to move in the desired way, but the dog
actually has no ability to perform the behavior on his own without
quite a bit of repetition. Usually, the dog is waiting for you
to show him what to do, and has become very good at being passive,
rather than active, during training sessions. A passive dog is
much less likely to try new things or make the effort to figure
things out on his own, which actually slows the training process.
“Clicker trainers never tell the dog to
stop doing something”.
This may be true for some trainers, but not
for good ones. Allowing the dog to behave in a wild and free manner
all the time is not the point of clicker training at all. Clicker
training is about developing control over the dog, but in a gentle,
hands-off way. It should not be about allowing dogs to practice
unacceptable behaviors over and over again, waiting until they
stop, then clicking. That’s just bad training. Good trainers
understand that it is important to manage the dog’s environment
and to prevent unwanted behaviors while at the same time reinforcing
desirable ones.
This is especially true if the undesired behavior
is self-reinforcing, such as chewing, digging, or barking. These
behaviors are enjoyable in and of themselves, and will continue
to increase even when ignored. In these cases, the environment
must be changed so that the unwanted behaviors are prevented.
“Clicker trained dogs are not as reliable
as traditionally trained dogs”.
There is absolutely no scientific evidence to
support this idea. Reliability comes from a strong reinforcement
history for a particular behavior. By heavily reinforcing desired
behaviors, then moving to a partial schedule of reinforcement,
most behaviors can be maintained with minimal effort. One problem
is that many trainers don’t understand how to move to partial
reinforcement and how to use it to its best advantage. Another
issue is that trainers who use luring never fade the lure effectively,
and the presence of the lure is required for the dog to perform.
Both of these issues are trainer errors, not fatal flaws of the
methodology.
“Clicker training only works for tricks,
not for serious things like obedience”.
Behavior is behavior. A sit, a spin, a 2 on
2 off contact performance, a stay, waving bye-bye, giving a kiss,
heeling, whatever, they’re all of the same level of importance
to the dog. The trainer is the only one who knows that one behavior
is a trick, the other serious business (only to the trainer).
To the dog, it’s all tricks! If more people approached their
training sessions with this thought in mind “it’s
all tricks!” then attitude and enthusiasm problems in performance
dogs would be a thing of the past.
Below are some unusual and usually ineffective
applications of clicker training:
The remote click: This involves pointing the
clicker at the dog, as if it were a television remote, then clicking.
First, there is no earthly reason to point the clicker at the
dog. It does not have an infrared ray that operates the dog’s
brain (don’t we wish!) This may startle dogs who are fearful
or noise sensitive. In many cases, keeping the clicker in your
pocket or behind your back when you click will work just fine.
The treatless click: This involves clicking
to mark a desired behavior, but not offering a treat or any other
primary reinforcer after the click. If your goal is to make the
sound of the click confusing, and eventually meaningless, this
is what you would do. The click gains its ‘power’
by being a predictable signal that a primary reinforcer is coming.
If you start varying that relationship, the click loses its predictability,
and its meaning to the dog. People who click but don’t treat
sometimes think they are using partial or variable reinforcement,
but this is not the case. When we move to a partial schedule,
we require more than one correct behavior before we reinforce,
we still keep the click/treat connection constant.
The jackpot click: When the dog does something
extremely well, or makes a breakthrough, the trainer, in his or
her enthusiasm, gives lots of clicks, but just one treat. The
idea seems to be that if one click is good, two are better, and
ten are fantastic! One of the most important aspects of the click
is that it marks a specific moment of behavior for the dog. Multiple
clicks can be imprecise and confusing. It is much better to use
one, well-timed click, then give multiple treats in order to provide
a jackpot.
The attention click (also known as the recall
click): Because clicker trained dogs have learned that the sound
of the click is important, people have begun using the clicker
to get the dog’s attention, or to get the dog to come to
them. I had a client once who had a conformation Rough Collie.
Upkeep on the dog’s coat took an enormous amount of time.
The client also had a pond in the fenced-in part of her yard,
and wanted to keep the dog out of the pond to avoid extra bathing
and grooming. At her second lesson the client told me that the
clicker was working well for recalls, but that the dog seemed
obsessed with getting into the pond whenever possible. When I
asked for further details it turns out that the client would see
the dog heading to the pond and then click, rather than call the
dog. The dog would instantly turn, head back to the house, and
get a treat. What the client didn’t realize was that she
was clicking the dog for moving toward the pond by clicking then,
so the dog was definitely going to repeat the behavior whenever
given the opportunity. The client was trying to use the click
to cause the recall to occur, but she was still reinforcing what
the dog was doing at the time of the click (heading for the pond).
The same would be true in using the click to gain attention. It
will lead the dog to look at you, but will also increase the behavior
that was occurring when you clicked.
The keep going click: This is a variation of
the treatless click. The handler clicks over and over for desired
behavior, often in a behavior chain, but only treats at the end
of the behavior. For example, a handler clicks for every correct
weave, but expects the dog to continue and finish the set before
giving a treat. I find it much clearer for the dog to use the
click not only as a behavioral marker, but also as a release.
I teach my dogs that the click means “that’s right”,
“a primary reinforcer is on the way”, and “you’re
finished”. If you click once, then allow the dog to continue
and the dog makes a mistake, do you treat for the click, or not
treat because of the error? I think that this could lead to major
confusion for the dog, so I choose to release after the click.
If I want the dog to continue, yet also want to let him know that
I am pleased with what he’s doing, I use a calm and quiet
verbal keep going signal (KGS). The KGS needs to be clearly different
from your verbal secondary reinforcer (which should be used exactly
as you would a click). People use verbal KGSs such as “gooood”
or “niiice” (drawing out and softening the vowels)
and they seem to work well.
The good boy! click: This is a click at the end of a training
session or series of behaviors that doesn’t mark any specific
action. It is used simply as another form of praise, because the
trainer is pleased with the dog’s progress/behavior in general.
Clicks that aren’t tied to specific actions tend to decrease
the value of the clicker as a conditioned reinforcer.
In Conclusion
These are just some of the more common
misconceptions and misuses of clicker training. I’m sure
that ‘creative’ trainers are developing more as we
go along! My suggestion would be to go beyond a basic knowledge
of the methods and techniques, to an understanding of the underlying
principles and theories. That way, you can judge for yourself
whether or not a specific application makes sense and will work.
Clicker training techniques can be tremendously powerful and effective,
when used properly. As with any method however, improper use will
lead to failure. Good trainers strive to understand not only the
techniques (how to do it) but the theories (why it works) as well.
As they say, knowledge is power.
Back
to Articles Index
© 2005 K9inFocus.com
Built and Maintained by Pet
Web Designs
Home
| About | Books
& Videos | Reviews | Workshops
| Calendar | Seminar
Contact Form | Photos | Articles
