The
Five Myths of Clicker Training
The myths
about clicker training and other positive training
abound. I hear them all the time. "You can't
take a clicker and treats into the ring." "My
positively trained dog refused a recall and got hit
by a car." "I tried that positive stuff
and my dog never did anything unless there was treat
in it for her."
While I bet
there are cases where every one of the above statements
is true, the fault lies, I'm afraid, with the trainer,
not the method. This article explains the truth behind
the most common clicker training myths.
Myth #1:
My dog will be food obsessed or will only work when
food - or the clicker - is present.
This is an
argument frequently made by people who train their
dogs for competition obedience. They can't take clickers
or treats into the ring, so they're afraid their dog
won't perform.
Food is a
marvelous training tool. Dogs are born needing and
wanting it - it's a natural reinforcer. But, like
any tool, it must be used properly.
Teach the
obsessed self control.
Some dogs,
often those who are unaccustomed to much variety in
their diet, become a bit too food focused. Some basic
attention exercises will help these dogs get their
minds off the food and back on the task at hand.
Start by holding
a piece of food in your closed fist. Your dog will
probably sniff, lick, and paw at your fist to get
you to open it. Instead, wait for the moment that
the dog pauses his exploration, click and offer him
the treat. Repeat until the dog learns that mugging
your hand will get him nowhere.
Then raise
your criteria. Hold the food away from you, and wait
until the dog glances your way. Just a glance - a
flick of the eyes - is enough. Click and treat. Gradually
increase your criteria until the dog maintains eye
contact with you instead of staring at the treat hand.
Be patient.
Don't call your dog's name - wait for him to offer
the behavior. No matter how persistent he is, he has
to take his eyes off the food at some point.
Keep the treats
off your body.
When you're
training in the house, keep the treats in a bowl separate
from you. It's okay if you have to walk a few steps
to get the treat after you click.
When you must
carry the treats with you, keep them in a fanny pack
or bait bag, and get them out one at a time after
the dog has performed the behavior. Resist the urge
to hold several in your hand. If you're going to have
to carry the bag of treats often, consider keeping
treats on you all the time so your dog won't associate
the presence of the treats with work.
Minimize food
lures.
Use a target
stick lure instead. Some dogs turn off their brain
and focus only on following the food if you use a
food lure. If you do use a food lure, fade it quickly,
and get the treats off your body.
Fade the clicker
and the treats once the behavior is on cue.
Once the behavior
is on cue, switch from the clicker to a verbal marker,
and begin using a variety of reinforcers: food, balls,
scratches, tug toys, and praise. Gradually increase
the number of repetitions performed solely for praise.
If your dog starts to get frustrated, back up, use
more food treats for a while, then increase the praise-only
repetitions more gradually. You want this variable
schedule to motivate your dog, not frustrate him.
Done correctly,
you can eventually save food treats entirely for new
behaviors and special rewards. However, don't quit
reinforcing your dog. Always offer praise or a smile
or some feedback that he has done what you wanted.
Myth #2:
My dog has no discipline - I don't want to punish
him, so he runs amok.
Being a clicker
trainer doesn't mean that you can't set and enforce
rules in your household. Not only CAN you set rules,
but it's an absolute must! However, you should be
a benevolent ruler rather than a malevolent dictator.
The first
step in being a benevolent ruler is to identify the
behaviors that you want to change. Every time you
interact with your dog, ask yourself, "Is my
dog doing something I want him to do?"
The second
step is to define what you want your dog to do. If
your dog is doing something you don't like, define
what you want him to do instead. It's not enough to
say "I want him to stop doing what he's doing."
He could stop doing what he's doing and choose to
do something worse - and then you'd have to stop that
as well. It's faster to define what you want him to
do from the beginning. For example:
-
I want
my dog to hold a sit-stay while I prepare his
food. (Not "I want my dog to stop jumping
on me when I prepare his food.")
-
I want
my dog to sit at the top or bottom of the stairs
when a person is walking up or down.
-
I want
my dog to lie quietly on a mat while the family
eats dinner.
-
I want
my dog to lie quietly on a mat when I have visitors.
The third
step is to manage the situation so your dog can't
do the behavior that he was doing instead of the preferred
behavior. The dog was doing the undesired behavior
because it worked, because it was somehow reinforcing.
For example,
a dog jumps on someone as a greeting, even if the
person yells and pushes him away. Why? Because the
dog wants attention. If he doesn't jump, he was likely
ignored. So he jumps, even if he is yelled at for
it. Until you can teach your dog that jumping isn't
reinforcing but sitting politely is, manage the situation
by putting him in another room when the doorbell rings.
The final
step is to train a preferred behavior. If you make
the new behavior reinforcing and simultaneously make
the old behavior unrewarding, the dog will quickly
choose to do the new behavior.
Myth #3:
Clicker trained dogs will choose to disobey because
there's no consequence for not complying.
Ignoring undesired
behavior
Because clicker
trainers often choose to ignore undesired behavior,
a myth has arisen that clicker and other positive-methods
trainers use no punishment and that there are no consequences
for not complying.
Ignoring undesired
behavior is a technique called extinction. Extinction
is an operant conditioning principle that states that
if a behavior is not reinforced, it will gradually
eliminate. It is the removal of a reward - or the
loss of an opportunity for a reward.
For example,
the trainer asks for a sit. The dog offers a down.
Instead of correcting the dog, the trainer waits five
seconds, then asks again. If the dog sits, the trainer
reinforces it. Because the down wasn't reinforced,
it is less likely that the dog will off it when "sit"
is cued.
In another
example, the same technique can be used to teach an
alternative behavior. A dog barks at passers-by. The
trainer waits for the dog to stop barking, then reinforces
the silence. To speed up the process, the trainer
sets up a scenario where a person will pass the house
several times. He reinforces the dog heavily in the
moments of silence, ignoring the barking. The dog
gradually learns that it is more reinforcing to be
quiet, so the barking extinguishes.
Negative punishment
In addition
to extinction, clicker trainers use negative punishment.
Negative punishment is the removal of something in
order to decrease the frequency of a behavior. Examples
include removing attention, giving a dog a "time
out," and feeding a treat to another dog (or
eating it yourself). When clicker trainers want to
suppress an undesired behavior, they rely more heavily
on negative punishment than positive punishment because
negative punishment is equally effective and causes
neither fear nor pain.
Although clicker
trainers try to rely most heavily on positive reinforcement,
the idea that they use it exclusively is completely
false.
Reliability
We train dogs
to make it more pleasant for our species to co-exist.
A large part of that training is teaching the dog
cues to perform certain behaviors. The unstated goal
of training is for the response to the cue to be almost
automatic -- almost a reflex.
If you can
drive a car with a manual transmission, chances are
you shift pretty much automatically. Do you remember
when you first learned to shift? Remember running
through the steps every single time you had to shift?
Think about how many times you had to shift just in
a short trip through the neighborhood. Think of all
the repetitions you have under your belt now. An approaching
stop sign, the sound of the engine straining as you
gain speed, a light changing from red to green - these
are all cues for you to shift gears when driving.
And you do so with only a cursory thought.
The conditioning
isn't 100%, however. If you're running an engine test
that requires you to make the engine work at a higher
number of revolutions, you can choose to ignore the
sound of the straining engine. Or, if you're distracted
by thoughts of your impending performance review,
you may miss that stop sign.
The same is
true in dog training. If you do the repetitions, train
for distractions, and generalize the behavior to different
locations, the cue and the behavior will be almost
inseparable. No training will 100% guarantee that
the behavior will occur. The dog always has a choice,
and someday you may find a distraction that you didn't
train for. But by building a solid reinforcement history
and doing the repetitions, you are increasing the
probability that the dog will perform reliably.
Myth #4:
Clicker trained dogs lack the precision of traditionally
trained dogs.
One of the
best techniques to emerge from clicker training is
pure shaping. In pure shaping, the behavior is reduced
to its most basic, tiny steps. For example, to shape
a spin, the trainer may first reward a glance to the
right, then a tiny head turn, then a larger head turn,
then a shifting of the weight, then the movement of
a paw, and so on until the dog is doing a complete
spin. This is how dolphin trainers and other trainers
who are unable to physically manipulate (model) their
animals train complex behaviors.
The only limitation
to shaping is the skill of the trainer. Precision
is gained by gradually increasing criteria. The most
common problem is the failure of the trainer to break
the behavior into tiny enough increments. The trainer
may expect too much too soon - wanting to start with
a step to the right, when he should start with a glance
- or may up the criteria too much at once - moving
from a head turn to a full step. Or the trainer may
not have a complete-enough picture of what the "perfect"
final behavior will be, making training less focused.
Shaping relies
on the animal's willingness to offer behaviors. If
an offered behavior doesn't earn a reward, the dog
must be willing to experiment until it figures out
what will earn a reward. This is one reason why clicker
trainers rely so heavily on negative punishment. Once
a dog understands the method, withholding a reward
motivates the dog to try something else. However,
if a wrong choice is punished using positive punishment,
the dog may become hesitant to offer another behavior
for fear of receiving another punishment.
Myth #5:
Clicker trained dogs fall apart when stressed.
To an extent,
this statement is true. If an animal has been raised
in a totally positive, stress free environment, it
will be unable to cope when faced with stress. The
same is true of a human. Negative experiences - stressful
experiences - are a part of life. These include emergency
visits to the vet, trips to the groomer, being petted
by person(s) they don't know, having their nails trimmed,
arguments between family members, and witnessing frightening
things happening to other dogs in classes or in the
neighbor's yard.
It is your
responsibility to prepare your dog to handle these
experiences. Fortunately, preparing them to handle
stress doesn't require subjecting them to horrible
experiences anymore than preparing your child to deal
with taunting classmates requires you to taunt him.
First, socialize
your dog. Expose him to as many different environments
as possible. The unknown is much more stressful than
the mundane. A dog that has met a variety of people,
seen bicycles and cars and basketballs, and visited
buildings and parks and crowded city streets is much
more likely to adapt to a new environment more quickly.
Next, handle
your dog. Clicker training encourages hands-off training,
but that doesn't mean that you shouldn't handle your
dog. In fact, you should train your dog to allow you
to clean his ears, examine his mouth, handle his feet,
and clip his nails. Your vet and groomer will be grateful.
Gently pull
his tail and his fur and reward him for allowing it.
You never know when a two year old is going to latch
on too tightly.
Teach your
dog that restraint equals relaxation. Instinctively,
your dog will fight being restrained. This can make
a bad situation worse if your dog ever slips his collar
next to a busy street or has to receive emergency
vet treatment. Wrap your arms around your dog. As
soon as he relaxes, even a tiny bit, click and release
him. Repeat this every day. With enough repetition,
restraint becomes classically conditioned to mean
relax.
Conclusion
Clicker training
is, in its simplest form, just a training method like
any other training method. Done properly it works,
just like other training methods. Clicker training's
edge come from its philosophy of rewarding the good
and ignoring the bad. It encourages the trainer to
concentrate on what the dog is doing right, not what
it's doing wrong. It gives the dog a chance to learn
how to exist in this strange, strange world. Clicker
training not only produces trained dogs, but it strengthens
that precious bond between dog and human. And isn't
that why we have dogs?
------------------------------------------------------------------------------------------------
This article
was reproduced on www.Southwestk9services.com
with the permission of the author Melissa Alexander.
www.ClickerSolutions.com ,
a site dedicated to helping pet owners improve the
relationship with their pets by teaching training
and management techniques which are understandable
and reinforcing to both human and animal
------------------------------------------------------------------------------------------------
|
