EDITORIAL
Considerations for shock and âtrainingâ collars: Concerns
from and for the working dog community
In the United States, the FBI periodically convenes sci-
entific working groups (SWGs) to establish guidelines, stan-
dards, or best practices in fields or disciplines relevant to a
law enforcement focus. Given the global events of the past
few years it is appropriate that dogs are now a focus addressed
by the Scientific Working Group on Dogs and Orthogonal
Detector Guidelines (SWGDOG) (
). As
co-chair of this SWG my task has been to emphasize roles
for canine behavioral genetics where appropriate in the
development of the guidelines, and to chair the subcommit-
tee on unification of terminology (SC1) and co-chair the
research subcommittee (SC7). This group is as unique and
diverse a congregation of people as I have ever encountered.
In addition to scientists, membership includes those from
national and international government and canine groups,
law enforcement officers, trainers, handlers, and those in-
volved in virtually all detection disciplines involving ca-
nines. All documents produced by this group are posted on
the public website in a step-wise manner and available for
public comment as the first pass of the topic is completed.
This is a truly collaborative effort, and I encourage readers
of this journal to visit the website and contribute when the
potential to be helpful exists. More information on the
mission and work product of SWGDOG will appear in
JVB:
CAR
in the future, but I wanted to discuss an issue that arose
as a result of my role as co-chair of this group.
Those of us who publish and speak frequently at collo-
quia or continuing education are often quoted or cited. In the
absence of a context that is provided by being present when
a talk is given or when a complete article is read, we often
have statements or polices attributed to us that do not
completely or accurately represent our views. I experienced
this phenomenon recently as a result, in part, of publishing
my editorial on shock (
). Because the issue is
a common one I thought that we might be able to engender
an expanded discussion on the facets that are part of this
complex discussion.
The issue raised with me involved a post on a listserv for
canine handlers that stated: âThe co-chair of the committee
is a behaviorist/vet, Karen Overall. Sheâs a sworn enemy of
such devices as Ecollars, pinch collars and any other devices
that can cause pain to a dog.â There are multiple issues
implied in this statement, and I shall go through them
stepwise. I have included some references in this article, but
the original list of references can be found in the editorial on
shock. Here is how I explained my thought process to those
on the listserv.
As an academician, I publish extensively, primarily in
the peer-reviewed and textbook literature. This means my
âviewsâ are a matter of public record, and that I based my
arguments on results found in the published literature.
As a specialist in veterinary behavioral medicine and a
researcher who focuses on the development and treatment
of anxieties in dogs, behavioral genetics, and performance
outcomes I am asked to evaluate and comment on a wide
population of dogs. Although most of the dogs on whom I
now focus are selected because of research interests (the
genetics of aggressions, anxiety, and noise reactivity), many
of these dogs are actually patients who come to me because
of problematic behaviors. Additionally, I continue to see a
subset of dogs who are not the focus of any behavioral
investigation but who are troubled dogs with distressed
humans. This means that one population for whom I eval-
uate tools are patients . . . problematic and distressed dogs.
Absolutely, without exception, I oppose, will not recom-
mend, and generally spend large amounts of time telling
people why I oppose the use of shock collars, prong collars,
choke collars, and any other type of device that is rooted in
an adversarial, confrontational interaction with the dog.
Without exception, such devices will make my anxious
patients worse and allow the anger level of my clients to
reach levels that are not helpful and may be dangerous. The
link between dog abuse and spousal/child abuse is now
well-established (
Ascione and Arkow, 1999; Lockwood and
). I educate people about this and about
breaking the cycle.
Simply put, when these adversarial methods are used on
my patients they become more anxious, more pathologic,
and potentially more aggressive and dangerous, depending
on their problem. I am currently dealing with a poodle who
is now biting the owner moreânot lessâ often than before
because the owner has begun to use a prong collar in her
1558-7878/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jveb.2007.07.001
Journal of Veterinary Behavior (2007) 2, 103-107
obedience training classes. This is a woman who has shown
dogs in obedience for years and thinks she is doing well
because she has rejected the calls from her fellow trainers to
use a shock collar.
The reason these devices make my patients worse are the
same reasons that I would oppose their useâwere I askedâ
for working dogs. They do not work the way people think
they do, and there are better tools and understandings for
accomplishing what people who use these seek to accom-
plish.
Such tools âworkâ by engendering fear, pain, and distrust,
and in doing so they cause long-term damage that make
dogs more reactive, less trusting, and less able to reach their
full potential in their partnership with humans, no matter
what form that partnership takes. These are not my opin-
ions: these are the findings from the scientific literature, and
this is an essential point.
As a scientist I do not have the luxury of claiming that I
am right because I have done something one way for years,
nor do I have the ability to assert that a technique does not
injure dogs when the evidence indicates otherwise. Science
is about understanding patterns in how the world works and,
as such, provides a methodology by which we can test
claims and assertions. When these adversarial methods have
been tested rigorously, they have been found wanting.
In the past decade, scientists have evaluated the effects of
shock, sudden loud noises, the force exerted by tight neck
collars versus harnesses, and numerous other control and
treatment issues that are now falling under the growing
domain of welfare issues in animals. The United States lags
behind most European countries, Australia, and New Zeal-
and, and, to a lesser extent, Canada in consideration of
animal welfare issues, but these issues affect many working
dogs and pertinent welfare information should be available
and accessible to those handling working dogs.
Those who know me can have no doubt that one of my
missions is to ensure that scientific articles and the infor-
mation they contain pertaining to canine behavior and be-
havioral genetics are available to the working dog and dog
fancier community. In fact, the citation and compilation of
such sources is one of the missions of SC7 of SWGDOG
and our database now contains greater than 900 research
articles and chapters in scientific publications.
A brief review of 2 of the main pointsâroles for immo-
bility and obedience and the effect of shock on molecular
and cellular âlearningââfrom the original editorial is war-
ranted here.
Roles for immobility and obedience
We have known for decades that shock works to teach
avoidance and cessation of behavior, which in the extreme
form often examined in the psychologic literature, is re-
ferred to as âimmobility.â It is this criteria of âimmobilityâ
by which learned helplessness is accessed (
No one who is recommending shock for treatment of be-
havioral problems has evaluated scientifically the extent to
which they may be inducing learned helplessness. In none
of the website sources supporting the use of shock that I
read did any of the authors realize that cessation of one
behavior did not mean that the dog was normal, or that he
or she was rationally complying with a program designed to
eliminate the reason for the behavior. In fact, it is often
claimed that clients find that the dog becomes âobedient.â
Obedient dogs can be quite distressed, and suffer from
profound anxiety while complying with a request.
We must ask ourselves 2 questions with respect to ces-
sation of a behavior and the potential for subsequent immo-
bility: (1) is immobility what we want and is cessation of
one behavior about which a client has a complaint suffi-
cient; and (2) what other behaviors or behavioral processes
are being affected when one is exposed to shock?
Cessation is insufficient for 2 reasons.
First, if the behavior stops we must realize that a âstopâ
here is only a halt in the process or signal and that the dog
must then be directed toward and rewarded for an appro-
priate behavior if we wish for him to be able to make such
a decision himself as a result of learning.
Second, and more important, the canine behavior for
which the animal is receiving a shock is not analogous to the
level-pressing behaviors so often cited in the rodent litera-
ture for which shock has been used as an assay for âmoti-
vation.â The behaviors for which people wish to use shock
in dogs are those that annoy humans. These behaviors are
either signals or non-specific signs of underlying distress. It
is clear from the above example that such distress is neither
considered nor addressed.
Effect of shock on molecular and cellular
âlearningâ
If shock and pain are profound, it is possible to induce
almost immediate long-term potentiation (LTP), the molec-
ular changes associated with hippocampal memory that will
lead to a strong aversion or phobia. The hippocampus is the
primary region where fears and anxieties associated with
fearful stimuli are thought to originate, so a logical sequela
to a stressful, painful stimulus may be fear, phobia, or
withdrawal. At the cellular level any kind of repeated rein-
forcement ensures better, more numerous and more efficient
connections between neurons. When stimulation continues,
we know that activity dependent plasticity at synapses (e.g.,
learning) occurs in the lateral amygdala. This is one modal-
ity postulated to be involved in learning of contextual fear
(
We may also be changing other behaviors or processes
when we expose an animal to shock (
In a landmark study published in 2004,
showed, using guard dog training of Ger-
man shepherd dogs, that there were untoward, negative,
long-term effects of training with shock. Dogs that were
shocked in training, but not when the evaluations were
104
Journal of Veterinary Behavior, Vol 2, No 4, July/August 2007
made, showed a lower ear posture in free-walking, and more
stress-related behaviors than did dogs who had not been
shocked in training. These differences were also found
when these dogs participated in obedience training and
manwork. In addition to the noted behavioral responses
associated with stress and distress found in dogs that had
been trained with shock, the researchers also found physi-
ologic differences in the HPA axis that were most profound
when the person associated with the shock (the owner or
handler) was present. Their conclusions were that: (1) this
type of training, in general, is stressful; (2) receiving shocks
is painful for the dogs; and (3) dogs learn a context-depen-
dent concern: the presence of the owner/handler and his or
her commands announces the reception of shocks. This is
still the most rigorous study on the responses of dogs to
shock, to date, and it shows that, although shocked dogs can
excel as guard dogs, their behaviors toward humans and
work circumstances changed, often indicating heightened
uncertainty and reactivity. We need to ask ourselves if this
is the response we truly desire or if we could do better.
Even if the response to a shock in a training situation
is to âstop,â a response, it in no way represents what the
dog will do in real-world situation. I am sure that most
readers of this post are familiar with the You Tube video
of the French police dogs, one of whom bites the handler
when the handler tries to stop him from going further (the
dog is wearing a shock collar) and ONLY responds to the
handler when the handler punches the dog in the face.
Simply, the dog is cognitively and physiologically past
any response to shock except to become more aroused by
it. To rely on shock to stop a dog in a real-world situation
is neither safe nor rational, based on what we know from
the literature of aroused states and physical violence. Any
officer who has responded to a domestic violence call
knows what force or pain will do to the situation. The
same logic applies to dogs. The president of one of the
larger, regional detection dog groups in the United States
told me that he believed that any handler who hits the
streets with a dog wearing a shock collar did not have a
well-trained or reliably trained dog. The handler, unfor-
tunately, does not understand how much he or she con-
tributes to the dogâs unreliability. In short, we need the
best partners available for K9 officers and the data indi-
cate that shock interferes with achieving this goal.
There are also serious physical and medical concerns for
using choke or prong collars on dogs. For years there have
been data implicating these tools in cervical (neck) insta-
bility and degenerative arthritis in dogs and in recurrent
laryngeal nerve paralysis, which can affect voice, swallow-
ing ability, etc. Such concerns are not rare in working dogs.
We now have data (
) showing that the
effects of neck pressure by collars increases intraocular
pressureâthe pressure in the eyes of dogsâin a manner
that is injurious to the vision of many dogs over the long
term, and in the short-term in a way that puts dogs with thin
corneas, glaucoma, eye injuries including corneal lacer-
ations, etc. at serious risk, especially during exercise or
activity. It should be noted that German shepherds have a
relatively high incidence of some of these eye conditions.
None of these effects were found when the dogs wore a
harness. The pressure changes noted in the eye were the
result of increased pressure on the jugular vein and all veins
to the head and eyes. Furthermore, the effects were more
profound with age, meaning the effect was larger for older
dogs.
The final issue I wish to address is the one that says you
should use aversive stimuli to teach dogs appropriate avoid-
ance. There are 2 parts to this issue: timing and learning
theory, and practical tests.
I have observed dogs on whom shock is being used to
stop the dog from pursuing the wrong target. For this to
have any chance of working one basic tenet of learning
theory MUST be complied with: dogs learn by association
if the stimulus is delivered within the first few seconds of
the behavior or its onset. This means that by the time the
dog is off-course, the onset of that behavior in the dogâs
mind may have been minutes ago. It is no surprise that when
shocked, many of these dogs just stand there and yelp.
Simply, the timing was wrong, the stimulus was greater than
it needed to be just to teach âno,â no alternative was offered,
and the dog learned something unintended. Many, many
working dogs have been shocked, but they may work as
well as they do, despite this finding, not because of it. That
statement should allow us to raise the question of whether
they would have worked even better were other methods
used.
âAversion therapyâ has been used to teach dogs to
âavoidâ classes of individuals like snakes, in regions of the
world where there are poisonous snakes, and stock, in parts
of the world where people are concerned that dogs will
worry or kill stock. There are no scientific studies on
whether shock teaches dogs to avoid snakes, in part, be-
cause the population data on the range of ânormalâ canine
responses to snakes are lacking completely.
There are, however, some population level data for the
shock and stock issue, within a restricted set of circum-
stances. Mort Bakken and colleagues in Norway have
looked at using electric shock to prevent dogs who might
do so from worrying/attacking sheep (
). In the Norwegian summer approximately 2.2
million sheep and 80,000 goats graze freely. In one
county in Norway, 66 sheep were killed by dogs between
1991 and 1998 (
âŹ
10 sheep/year for this county). This
study, conducted over 2 response years, was designed to
measure the effects of electric shock on behaviors
thought to be associated with sheep attacks. It has some
interesting implications, not all of which are noted by the
authors. First, these are contrived experimental situations
and there were no actual quantitative or qualitative data
on the dogsâ behaviors before the experimental test. Sec-
ond, to receive the 1-second, 3,000 V, 0.4 Amp electric
shock (manufacturer specifications, no validation data),
105
Overall
Editorial
the dogs had to exhibit a specific set of behaviors when
in the presence of sheep. In the first year of the study
87.7% of the 138 hunting dogs chosen for study received
no shocks because they did not exhibit the needed be-
haviors (!). There were individual and breed differences
in response to sheep and these differences in the way or
rate the dogs withdrew from sheep was not independent
of behavioral patterns exhibited in the presence of other
dogs and humans reported previously. Third, younger
dogs (
âŹ
3 years) exhibited more interest in sheep and
chasing of them, suggesting that social development was
occurring and an understanding of developmental stage-
associated specific behavioral suites may have been help-
ful in understanding how many behaviors labeled as
associated with sheep attacks actually reflected such be-
haviors. Fourth, there were some interesting behaviors
noted in the second year of the study when shocked
versus unshocked dogs were compared. Interest in sheep
decreased between years regardless of group, although
reactivity (measured by latency to respond and discovery
distance) increased in shocked dogs. In the end, one dog
who was persistent with sheep the first year and was
shocked for it consistently, still attacked sheep in the
second year. This is important because it hints at variance
in rewards. One interpretation is that the cost of the shock
was worth the reward gained by ignoring it. Fifth,
shocked dogs displayed various reactions to shock, in
addition to withdrawal, including head shaking, vocaliza-
tion, and jumping. There were also changes in their
behaviors toward humans suggesting increased vigilance
pertained to humans. Fourteen dogs studied the second
year who approached sheep and were shocked in this
experimental context during the first year, did not do so
or receive shocks the second year. However, these dogs
exhibited different behaviors than did the unshocked
group, most of which are only hinted at here, but suggest
a response to distress. Given the behavioral patterns
discussed in this article, and the overall incidence of the
problem, one is left wondering whether an aversion train-
ing system that alters normal behaviors for the worse is
worth risking given that truly dedicated animals donât
necessarily show the desired response. In the absence of
a comparable, but non-antagonist method (e.g., a control)
to discourage such behaviors, we have to acknowledge
that the data are incomplete.
A more recent study comparing the response of gold-
fish and trout to shock intended to teach avoidance (
) shows why our investigation of specific
behavioral responses to shock should be careful, com-
plete, and nuanced. In short, social patterns between
species affected responses: trout, a social species, would
withstand the highest level of shock (30 V) administered
in the experiment if their conspecifics were on the other
side of the shock. Otherwise, they learned to avoid 3 V
shocks. For goldfish, the presence of conspecifics did not
have the same effect. No fish were unchanged by the
shock when before and after behaviors and physiology
were examined. Cortisol indicated the presence of a
stress response even when a âmildâ shock was applied,
and the development of behaviors indicative of fear were
noted.
Given what I do for a living, my focus is always going to
be the dogs. That said, whether or not anyone understands
or accepts what I have written here, there is one aspect, not
yet discussed, that must be considered. A blind reliance on
these methods is preventing handlers and trainers from
having the partnership with science and scientists that
would benefit both of them.
We are in a time period where canine cognition is the
focus of much important research, often with surprising
results (
Hare et al., 2002; Kaminski et al., 2004)
. An un-
questioning commitment to old methods is going to hurt the
canine-handler team. Regardless of how I feel about these
methods, I think the people and dogs who work so hard to
make the world a safer place are heroic. They deserve the
best methods and data available to help them do their jobs.
A large part of my research is committed exactly to this
effort. Putting to use the knowledge we have regarding
canine cognition and learning depends on doing something
that was antithetical years ago: working with the dog as a
cognitive, reasoning individual in a partnership based on the
best use of everyoneâs skill sets. This is a heck of a lot
harder than treating dogs simply as another tool that cannot
reason. The historic use of adversarial, coercive techniques
no longer makes sense given what we now know about dog
cognition and learning.
We can do better. The canine handler teams that work
best are those that best understand and trust each other.
Anything that interferes with that trust and understanding is
hurting the team. When I work with individuals or groups
who use dogs in work I spend 90% of my time just trans-
lating for the dog.
There are alternatives to aversive devices. I recently
watched a Schutzhund dog work just as well on a
Scruffy-Guider (Misty Pines Dog Park, Sewickley, PA)
as he did on a choke collar, but he breathed better. I have
seen military dogs learn almost instantly using head col-
lars (Gentle Leader; Premier Pet Products, Midlothian,
VA) because the target of their focus was clear. And I
have seen my own dog, Flash, recover from being hung
from a choke chain until he passed out, after which time
he put the trainer in intensive care. That is how he
became my dog. . .he was my patient first. Some people
reading this may have met him, and so know what an
amazing dog he is. Flash is the individual who first
opened my eyes to learning to think in a different way
simply because any forceful interaction with him would
have resulted in injury to those exhibiting the force. No
exceptions. His lessons have benefited many.
Finally, SWGDOG is committed to using science to
make canine teams more effective, and to ensure that the
data from such teams meets standards that will allow those
106
Journal of Veterinary Behavior, Vol 2, No 4, July/August 2007
data to be upheld in court and other related situations. I am
co-chair of this group, along with another academician, a
chemist, Dr. Ken Furton, Florida International University,
because canine behavior and genetics and analytical chem-
istry are both growing fields and at the cornerstone of what
detector dog teams do. When we created SWGDOG we all
realized that this was a chance for the operational and
scientific community to talk and work together, and I remain
committed to this effort. We live in a time when we simply
have no other choice.
I think that the same collaborative sentiment must hold
true for all of us who work with animals. We have run out
of time to hide behind our own myths and in ivory towers.
If not us, who; if not now, when?
Karen L. Overall
Philadelphia, PA
References
Ascione, F.R., Arkow P., 1999. Child Abuse, Domestic Violence, and
Animal Abuse: Linking the Circles of Compassion for Prevention and
Intervention. Purdue University Press, West Lafayette, IN. p. 479.
Beerda, B., Schilder, M.B.H., van Hooff, J.A.R.A.M., de Vries, H.W., Mol
J., 1997. Behavioural, saliva cortisol, and heart rate responses to dif-
ferent types of stimuli in dogs. Appl. Anim. Behav. Sci. 58, 365-381.
Christiansen, F.O., Bakken, M., Braastad, B.O., 2001a. Behavioural dif-
ferences between three breed groups of hunting dogs confronted with
domestic sheep. Appl. Anim. Behav. Sci. 72, 115-119.
Christiansen, F.O., Bakken, M., Braastad, B.O., 2001b. Behavioural
changes and aversive conditioning in hunting dogs by the second year
confrontation with domestic sheep. Appl. Anim. Behav. Sci. 72, 131-
143.
Dunlop, R., Millsopp, S., Laming, P., 2006. Avoidance learning in goldfish
(
Carassius auratus
) and trout (
Oncorhynchus mykiss
) and implications
for pain perception. Appl. Anim. Behav. Sci. 97, 255-271.
Hare, B., Brown, M., Williamson, C., Tomasello, M., 2002. The domes-
tication of social cognition in dogs. Science 298, 1634.
Kaminski, J., Call, J., Fischer, J., 2004. Word learning in a domestic dog:
Evidence for âfast mapping.â Science 304, 1682-1683.
Lockwood, R., Ascione, F.R., 1998. Cruelty to Animals and Interpersonal
Violence: Readings in Research and Application. Purdue University
Press, West Lafayette, IN. p. 452.
Overall, K.L., 2007. Why electric shock is not behavior modification. J.
Vet. Behav.: Clin. Appl. Res. 2, 1-4.
Pauli, A.M., Bentley, E., Diehl, K.A., Miller, P.E., 2006. Effects of the
application of neck pressure by a collar or harness on intraocular
pressure in dogs. J. Am. Anim. Hosp. Assoc. 42, 207-211.
Schafe, G.E., Nader, K., Blair, H.T., LeDoux, J.E., 2001. Memory consol-
idation of Pavlovian fear conditioning: A cellular and molecular per-
spective. Trends Neurosci. 24, 540-546.
Schilder, M.B.H., van der Borg, J.A.M., 2004. Training dogs with the help
of the shock collar: Short and long term behavioural effects. Appl.
Anim. Behav. Sci. 85, 319-334.
Seligman, M.E., 1971. Phobias and preparedness. Behav. Ther. 2, 307-320.
107
Overall
Editorial