Tarheel Canine Training, Inc.
May 9, 2003
This article is copyrighted original work and may not be reproduced
in whole or in part without the expressed written consent of the author.
The electronic training collar is gaining more and more widespread
use and acceptance in both pet training circles and working dog training.
Though some criticize the electronic collar as cruel, these criticisms
are based on false analogies and appeals to emotion and have no basis
in fact. As with any training device which is designed to administer
an aversive stimulus, it can be misused. Therefore, rather than condemn
the device based on isolated misuse or arguments based on appeals to
false emotional misgivings, we should all learn and teach more about
its proper use. We should establish guidelines regarding how and when
to employ it, and with what kinds of canine temperaments. When it is
an appropriate tool to achieve the goals we have for our dogs, it is
unsurpassed in its ease of use and efficiency in meeting those goals.
Some decry its use because it is a form of aversive stimulus. But these
people do not approve of any kind of stimulus that is not positive in
nature. Their argument usually is formulated on the grounds that aversive
stimuli in training will "break the bond" between dog and
handler: cause the dog discomfort and he will associate you with that
discomfort. There is no real evidence for this, but it appeals prima
facie. This argument assumes that every dog has a soft temperament,
which is certainly not true. Further, one only needs to see a mother
bitch with her pups to see that pups learn both from positive as well
as negative consequences from mother, and after 10 years of breeding
dogs, I have yet to see puppies turn away from mother because she corrected
their behavior with a growl or alpha roll.
The benefits to using the collar to deliver aversive stimuli are many.
Primarily, it allows the trainer to teach that unacceptable behaviors
do have consequences; regardless of how close the dog is to you, or
whether the dog is wearing a leash and is tethered to the trainer. It
allows for consistent levels of correction in a wider variety of contexts.
Dogs trained to off-leash control can go to the beach, swimming, hiking
in the woods, and still remain under control should some competing motivation
Further, one can administer the aversive without becoming physical
with the dog. By necessity, a leash correction requires the trainer
to make hand and arm movements, which can cue behavior. In addition,
the e-collar allows unemotional delivery of corrections, and lessens
the association between the handler and dog with the negative stimulus,
for dogs with temperaments that are soft in nature. In fact, many people
think that soft temperaments are poor candidates for e-collar training,
and in my opinion, just the opposite is true. Dogs that are sensitive
to the handler can be corrected on very low levels of stimulation, and
the handler is not physically delivering the correction.
There are some trainers who are teaching the use of the electronic
training collar based on methods developed when electronic collars were
first employed, and labeling their training methods as "new"
or "progressive." In my opinion these claims are false claims,
as the methods used have been around for a very long time. Therefore,
buyers and seminar goers should be wary of the information being sold.
In the first part of this article we will explore how training is done
with these devices.
Electronic training collars (e-collars) are comprised of a collar receiver
that goes around the neck of the dog, and a transmitter that is held
in the hand of the trainer. The trainer can choose how to deliver the
stimulation, either by pressing the "nick" button or the "continuous"
button. In "nick" mode, the transmitter delivers a pre-timed
burst of stimulation on the order of a fraction of a second. In continuous
mode, the collar delivers a continuous string of these pre-timed bursts
as long as the button is held down. In this mode the collar usually
has a fail-safe allowing only 10 seconds maximum stimulation.
The stimulation delivered is a very low amperage electrical charge,
which stimulates the nerve endings in the neck. It feels exactly like
a static electricity shock you might get from wearing wool socks on
a carpet and then touching a doorknob. We have all done this, and felt
startled, but we know that it is impossible for it to hurt us. One feature
of getting a static electric shock is that one usually doesn't want
to repeat it. This is the essential feature of low amperage stimulation
that is useful in training a dog - the dog learns to do what we want
in order to avoid the unpleasant sensation.
To understand how an e-collar provides an unpleasant stimulation, yet
doesn't damage sensitive tissue; one needs to understand a little about
electricity and how it generates power. Electrical current has two essential
features: voltage and amperage. The power of an electrical current is
measured in Watts, and is the product of voltage multiplied by amperage.
If we think of an electrical circuit as a pipe with water flowing through
it, voltage is the force of the water pushing through the pipe, and
amperage is the volume of electrical current flowing through the pipe.
E-collars have high force but ultra low volume. Thus if the amps are
low, even if the voltage is high, you can have high force behind the
electrical current, but a very low output of power. It's like being
shot with a water pistol versus a fire hose. A water pistol has a large
force on a tiny volume of water. A fire hose has a high force on a high
volume of water.
Many people make a false analogy between e-collars and getting "shocked"
by the kind of electricity one would find in a house. House electrical
current, or that from a car battery for that matter, is high amperage.
High amperage means there is a high volume of electrical current, which
can actually do physical damage. E-collars do not carry high amperage,
and thus are unable to cause physical damage.
The manufacturers sell most modern electronic training collars as stand
alone training systems. Their instructions include procedures for employing
the collar as a teaching tool, to teach the dog new associations between
command words and behaviors. This is done through using a behavioral
consequence known as negative reinforcement. Reinforcement is anything
that increases the likelihood of a behavior. Negative reinforcement
requires removing an unpleasant consequence to increase the likelihood
of a behavior. To do this with an e-collar, one puts the collar on a
low setting and presses the continuous mode button, holding it down.
The trainer then guides the dog into the behavior (e.g. "Sit")
and when the dog places his rear on the ground, the trainer releases
the button, thus removing the unpleasant stimulation. The dog makes
an association: sit removes the unpleasant feeling, thus increasing
the likelihood of the sitting behavior.
There are, however, a few built in side effects to this approach. First,
in order to remove the unpleasant feeling when the dog achieves the
intended behavior, we must first induce the unpleasant feeling. This
also provides an association: new learning can be unpleasant. Depending
on the temperament of the dog, this can have no impact whatsoever, or
can have a dramatic impact on the dog's behavior and desire to learn.
Some of the trainers using this approach tout the fast results, and
resulting calmness of the dog. When in reality the dog is stressed,
and on the verge of shutting down, because he doesn't understand until
he has had many repetitions what he is to do to escape the unpleasant
feelings. The result is he is afraid to do anything that might bring
on more unpleasant feeling.
All learning is stressful to a dog, but by a matter of degrees. When
the dog has no idea of how to escape the unpleasant stimulation, he
likely will shift into a defensive mood. When in a defensive mood, a
dog has three options: to choose to fight against it, to choose an avoidance
strategy, or to displace (shut down). The goal of negative reinforcement
training is to have the dog figure out that he can avoid the unpleasant
feeling by performing a very specific behavior, e.g. to sit. When considering
this particular behavior out of the myriad choices of behaviors, one
can imagine that the dog will go through a number of behaviors that
don't work, since he hasn't been taught which behavior actually will
work. Some trainers call this "exploring behaviors." I see
it as a hole in the method. Why not teach the dog a set of behaviors
that are likely to come into play, in a non-stressful way, before applying
unpleasant stimulation. In fact before using the collar at all, why
not train these behaviors motivationally, then use the collar as a form
of positive punishment?
Punishment, in the animal behavior context, is anything that reduces
the likelihood of a behavior. Positive punishment means we apply an
unpleasant consequence (e-collar stimulation) to reduce the likelihood
of a given behavior. In this sense, we will positively punish all unwanted
behaviors, and positively reinforce all trained behaviors. The e-collar
then becomes a tool for what we normally refer to as correction, rather
than a teaching tool. This avoids the majority of the stress of the
old method of negative reinforcement training, and results in negative
associations only with unwanted behaviors, and positive associations
with all trained behaviors.
There may be times when negative reinforcement is a valuable approach
to teach a particular behavior in a particular way. However I still
believe that teaching the dog the route to escape the negative consequence
will make the training proceed more quickly in a less stressful way,
and consequently it is easier for the dog to choose an avoidance strategy
that will work and end the stimulation.
One training concept often trained with negative reinforcement is the
retrieve of the object exercise in any of the protection sports: schutzhund,
ring, or PSA for example. The reason we train a retrieve with negative
reinforcement is to obtain two important goals: First, the dog must
think of retrieving as work and not play. A dog that retrieves only
out of play may have many other competing motivations, especially dogs
of this caliber that do protection sports. Let's say he likes to bite
the decoy more than retrieve. All the dog needs to think, then, is there
is the promise of bite work, and he may choose not retrieve at a crucial
time. Second, play retrieving also brings with it many characteristics
that are judged negatively: fast to the object and slow to return, mouthing
and playing with the object, nosing it, pawing at it, etc. The force
tends to make the dog think of the exercise as more urgent and important
than any possible competing motivation. The force also makes it less
of a game, and diminishes the likelihood that the dog will play with
the object or drop it like he may with a play toy.
Given that there may be valid reasons for using negative reinforcement
training, then, why not make the entire exercise more easy for the dog
to learn, rather than more difficult. There must be a better way than
allowing the dog to "explore" random behaviors. Thankfully,
there is a better way. Teach the dog to hold and grab the object motivationally
(it may not be possible with some dogs that have no desire to retrieve,
but this is a big minority) first, to give the dog a sense of the exercise.
So, whether you are teaching new concepts by positive reinforcement
or negative reinforcement, there is ample reason to attempt to lay a
motivational foundation to the required work. The next issue is, how
do we do it, and then how exactly do we add the e-collar into the training
in a way that it is very clear for the dog, and minimizes stress, and
The Concept of Pairing Corrections
E-collar companies, and some of the e-collar trainers now teaching
seminars on its use, often begin their introduction to using the e-collar
by saying that right from the beginning we can teach your dog without
ever using a leash, and standard on-leash training is out of date with
the new e-collar technology. I completely disagree with this approach.
It is much easier to introduce positive punishment in the form of a
leash correction. Leashes attached to training collars have 2 components
of interest: force and direction. A leash tugged gently in the upward
direction, after a dog is taught motivationally (positive reinforcement
& negative punishment) to sit, is a very easy way to deal with non-compliance
to a learned behavior. The leash provides a reason not to choose to
ignore the command to sit, and it also provides guidance on how to correct
the inappropriate response. Each command has an associated leash correction,
on which we can vary the force, and provide feedback to the dog on what
exactly was required by the command. We teach what we refer to as guiding
corrections to introduce the dogs to leash corrections (their first
introduction to positive punishment). Guiding corrections have very
little force but apply guidance to the dog about what he did wrong and
how to correct himself in the future. He learns therefore what each
of the leash corrections mean, before they become a truly aversive stimulus.
Once this is complete we morph the guiding corrections into standard
leash corrections when commands are not properly executed. This is just
increasing the level of the force until it is enough to positively punish
the unwanted behavior in the given context.
Once the dog understands the meaning of leash corrections, we can introduce
the e-collar as a new correction. To do this we employ simple classical
conditioning to teach the dog, in a given context, that the e-collar
stimulation (an aversive stimulus) means the same thing as the leash
correction. The problem is that e-collars employ only force and not
direction, the stimulation comes from the same direction all the time.
Some trainers will move the collar box around the dog, up on the top
of the neck for the down correction, under for the sit correction, etc.
I don't do this either, as I believe it makes the dog wise to the placement
of the collar.
Before I describe the procedure, I would like to mention that for me,
the e-collar is best used as a correction for the action commands of
heel and come initially. In fact for any pet dogs we train, those are
the only corrections we give on the e-collar. This makes the context
very easy for the dog to process. And for the average pet dog owner,
if a dog breaks a stay, the dog can be called back to the handler and
the come enforced with the e-collar, and then the dog is placed back
in the stay. As working dogs progress through the heeling, we will introduce
corrections in other contexts using the collar, but they are few. Most
other training can be managed without resort to e-collar corrections.
I do use the e-collar paired up with a verbal reprimand so that I can
get more out of my verbal corrections. But to understand this we must
explore the concept of pairing corrections.
Classical or associative conditioning is what we rely on to switch
the correction from a leash correction to an e-collar correction. Classical
conditioning relies on the research of Ivan Pavlov, the Russian researcher
who discovered that an initially neutral stimulus (ringing a bell) when
paired with an unconditioned stimulus (one we don't necessarily have
to teach the dog - like salivation in the presence of food in Pavlov's
work) would elicit the same response as the unconditioned stimulus.
He discovered that the path to this result is that the conditioned stimulus
must closely precede the unconditioned stimulus in time.
We apply the same approach to the introduction of the new e-collar
correction: The correction we wish to condition, that of the e-collar,
is placed in time closely preceding the existing leash correction for
heeling. The leash correction for heeling is a 180-degree turn with
a jerk-and-release correction on the training collar (pinch or choke
- I prefer the pinch). Thus what we do is as follows: when the dog shows
an undesirable behavior (say, forges out of heel position) the trainer
makes a 180-degree turn, nicks the dog on the e-collar, and follows
that with the familiar jerk-and-release correction on the leash and
training collar. The dog will soon learn that if he forges or goes wide,
the e-collar will stimulate him, and that means get back in position.
Over a period of sessions we look for anticipation: the dog reacts to
the e-collar correction after it is administered, and before the trainer
follows up with the jerk-and-release. This anticipation clearly shows
that the dog understands that the e-collar correction now means get
back in position. The trainer can now eliminate the leash correction
and eventually just carries the leash until the dog makes few if any
errors. At that point the leash can be discarded, and the e-collar correction
controls any unwanted behavior we may get during heeling. Notice that
the heeling behavior itself is already well established before we go
to the e-collar. Thus good associations are made with correct behavior,
and the only negative associations are associated with inappropriate
It is easy to extend this simple philosophy to any correction you make
that involves a leash. The come command is a natural extension. Further
one can add the extra dimension of the verbal reprimand into the heeling
and come corrections. Just before the nick correction on the e-collar
for unwanted behaviors, use your "no" command - I usually
make a nasty grunt, it doesn't matter what it is, just that it precede
the e-collar correction, making a further classical association. This
gives power to your verbal reprimand - it is backed by the compulsive
power of the physical e-collar correction.
I use this to deal with many other situations, such as stay commands,
and corrections for sit and down, but especially if I see the dog is
about to make a mistake, the reprimand "nips it in the bud"
before a physical correction must be given. Correcting a dog before
he makes a mistake it much more effective than correcting him after
the fact. But you must be a good trainer able to anticipate the errors
before they are full-blown errors.