Are Shock Collars Painful or Just Annoying to Dogs? A 2004 Study Reveals Some Answers

46 | Posted:

By Sophia Yin, DVM, MS

R.I.P. 1966-2014

Trainers often debate about the use of electronic shock collars. Some trainers find these collars unethical and unsafe. The pro-collar camp takes a different stance. Some say it just distracts the dog, calling it “tap technology” and others say it may be painful at the instant but then the dog learns to behave and there are no lasting negative effects.

In 2003, researchers from the Netherlands, Matthijs Schilder and Joanne van der Borg, assessed the short and long term behavioral effects of dog training with the help of shock collars. They wanted to know three things:

  1. Do shock collars cause pain and fear or just cessation of a behavior? This could be evaluated by looking for signs of fear and pain when dogs receive a shock.
  2. If the shock collars cause pain and fear, do the signs of fear fade afterwards such that the dog is completely normal or do signs of fear and anxiety persist? For instance, if dogs have received shock on the training ground do they show more signs of fear during non-training times in the same area when compared to dogs that have not been shocked?
  3. And lastly, the researchers wondered if they could distinguish shocked from non-shocked dogs by fear/anxiety responses outside the training grounds. That is, are dogs who have been shocked more fearful in non-training locations? If so, it indicates they associate the handler or being given commands with the reception of shocks.

The Study Group

Schilder and van der Borg used Malinois, Malinois crosses, German Shepherds and one Rottweiler from a group of dogs being trained for their official (IPO ) certificate as police dogs as well as dogs being trained for standard watchdog training for a comparable (VH3) certificate, which is the highest possible in this type of training. Because these were working dogs they differ from the general population of dogs in that they are higher energy, higher drive, and have a higher tolerance for the correction-based training for which they are bred.

The 32 shock-collar group dogs (S-dogs) received shocks during training. The control group received no shocks but did receive other harsh methods including choke chain corrections, pinch collar corrections, other physical corrections (C-dogs). The researchers had no influence upon the methods and aids used, rather they just observed the trainers during the routine training sessions and “free walking” sessions in which the dog was not being trained or given corrections.

Overall they observed 32 shock collar-group dogs receiving 107 shocks and 16 control dogs who received other types of corrections instead. They evaluated control and experimental dogs in three situations:

  1. First a free walk on the training grounds in which the dog was walked on leash but no orders were given to the dog. This was to see if there was a behavioral difference between the non-shocked vs the shock collar dogs and whether the type of correction had a lasting effect outside of the correction-situation.
  2. An obedience work session on the training ground which included the following commands—sit and down in motion, heeling in slow, normal and fast walking speed with changes of direction, and recall to the handler. This situation was to determine whether the S-dogs showed signs of fear or pain when corrected.
  3. A protection work session on the training ground in which the dog performed a number of exercises such as search for criminal, hold and bark at criminal, escape and defense, followed by attack by the criminal, and finally transport back.
  4. They also filmed the dogs during a “free-walk” session at a park (a new location) and then an obedience session at the park. This was to see whether there was a difference between control dogs and S-dogs and whether S-dog associated the shock correction with the handler.

The Effects of Shock-Collar Corrections on Body Posture

The study found that in the 32 dogs that received a total of 107 shocks, there was an immediate direct effect in which the dogs most commonly:

  • Lowered their of body posture (22 of 32 dogs)
  • Gave high-pitched yelps (17 of 32 dogs)
  • Gave tongue flicks (18 of 32 dogs)
  • Lowered their tail (13 of 32 dogs)
  • Squealed (13 of 32 dogs)
  • Turned their head down and to the side to avoid the shock (7 of 32 dogs)
  • Moved away (avoidance) (14 of 32 dogs)
  • Gave a barking scream (5 of 32 dogs)
  • Crouched  (6 of 32 dogs)

Dogs also lifted their front paw, lowered their back, jumped, licked their lips, circled, trembled, and sniffed the ground. All of the listed behavioral responses are signs of fear, pain, or anxiety and stress. Seven dogs showed no reaction.

The Effects of Previous Shock-Collar Corrections on Behavior at the Training Ground

Dogs that had been shocked previously showed more signs of anxiety and fear then the control dogs during free-walking on the training grounds as well as when they were being trained. During the free-walking and obedience work, S-dogs exhibited significantly more lip licking and lower ear positions indicating lasting effects of shock on overall fear and anxiety. During the protection work they showed more paw-raising.

The Effects of Previous Shock-Collar Corrections on Behavior in a New Setting (The Park)

Dogs that had been shocked previously showed more signs of fear an anxiety in the park situation than the control dogs. They showed a higher frequency of low ear position during the free walk than the control dogs and lower ear position and tongue flicking during obedience exercises in the park.

Behavior on the Training Ground Vs the Park and When Being Trained Vs on Free Walk.

Dogs that had previously been shocked were more frightened on the training ground than in the park. They carried their tails lower on the training ground than in the park and lifted their paw more. They were also more frightened during training than when being walked—ears and tail position were lower when being trained. However, non-shocked dogs also showed more signs of fear when being trained than when being walked.

The Take Home Messages

Overall the researchers concluded that even when compared to working dogs trained using choke chain and pinch collar corrections, dogs trained with electronic shock collars showed more fear and anxiety behaviors than those trained by other traditional police dog and watchdog methods. They concluded that:

  • Avoidance behavior and fear postures during the shocks indicated that the shock elicited both pain and fear and, therefore, were not just a distraction or nuisance.
  • The fact that the dogs showed more fear than control dogs both in the non-training situations in the familiar training grounds as well as in the park indicates that dogs are learning to associate the shock, not just with the unwanted behavior, but also with the location/environment as well as the trainer. The researchers found some evidence that some dogs had also learned to associate commands with shock. For example they state that one dog, shocked immediately after getting a “heel’ command, yelped after getting the next “heel” command without being shocked. The authors point out that the dog was not given a chance to respond after given the “heel” command. Rather, the command was immediately followed by the correction, hence, increasing the likelihood that this type of aversion association would be made.
  • The researchers state that in the presence of the handler, the dog has learned to expect something aversive. “The enormous rewards the dogs experience during training, i.e., chasing down, catching a criminal and winning the sleeve, do not counter the negative effects of getting shocked. This is in spite of the fact that handlers of non-shocked dogs admitted that they use prong collars and that their dogs experienced beatings and other harsh punishment, such as kicks or choke collar corrections.”
  • Both dogs trained using electronic shock collars and those trained with other traditional coercive methods (choke chain, pinch collar, physical punishment) showed more signs of fear and anxiety when being trained than when on a free walk.

Interestingly, the results did show that 7 dogs out of 32 (22%)  showed no signs of fear or pain while actually receiving the electronic collar shock which indicates that some dogs bred for high drive and to withstand the demands of the coercive-type training appear to have no pain or fear of the shock. The study does not indicate whether these 7 dogs failed to show fear and anxiety in the other test situations though.

Their final thoughts—it would be interesting to see whether the shocked dogs also show more signs of fear with a different handler and the next step is to compare protection and guard dogs in a more “friendly” way.

Sources: Schilder, M., Van der Borg, J., 2004. Training dogs with the help of the shock collar: short and long term behavioural effects. Appl Anim Beh Sci, 85, 319-344.

For more information on dominance-based training, read Experts Say Dominance-Based Training Techniques Made Popular by Television Can Contribute to Bites and Trainers with Jackhammers Need Not Apply

You can also read Perfect Puppy in 7 Days: How to Start Your Puppy Off Right for examples of positive reinforcement methods.

Tags: , , , , , , , , , , , , , , , ,

46 responses to “Are Shock Collars Painful or Just Annoying to Dogs? A 2004 Study Reveals Some Answers

  1. Interesting article but I don’t understand why positive reinforcement wasn’t used also. I think they would have seen the effects positive punishment really has on dogs.

  2. Hm, interesting study. But if training with a shock collar and after the shock the dog showed the effects you listed, you have the shock collar on way too high of a level.

    The tapping system (which really should be the only system used) is not designed to get those effects. It should only be enough to modify the behavior, or to irritate the dog. Not to make him yelp, bark, cringe, avoid, etc.. So I don’t see the results of this study as ‘dogs trained on shock collars are more fearful’ I see it as ‘incorrect use of shock collars produce fearful dogs’.

    I also find it interesting the study couldn’t show weather or not the 7 dogs that didn’t show signs of fear, even through such high level shocks, their behavior when not being trained. Why would the study not collect that data? A bit suspicious to me. But an interesting read, none the less.

    1. Exactly. Virtually any tool or object for that matter can be used to inflict pain. We should ban pens as some idiot could buy one and jab it into the eye of a dog. I use an ecollar to allow my scent hound some off lead freedom. She is an ex racing dog who was bred and trained to follow a scent cross country while ignoring dogs and people. Not an easy animal to teach recall to. Using the lowest level she could feel she was taught to look to me when she felt it. I have then used more traditional recall training which she follows once I have her attention. She also recalls to the tone and vibrate settings if its a low distraction situation.

      1. Actually dr John Bradshaw demonstrat s in DOG SENSE how even the use of beep or vibration caused immense fear and subsequent issues in some dogs with an especially bad effect on dogs with anxiety to begin with. After reading the fear it instills (which is hey so effective short term) it is not likely you will continue to use unless your dog’s safety is in question which very well may be. I have no judgement, I am just sharing the disturbing research I read in his book. This information should be spread so we all have the info we need to make educated decisions.

      2. Was a pen designed to be jabbed into a dog’s eye? But a shock collar is designed to inflict pain or at the very least significant enough discomfort to change or stop a behavior. The intention of the tool is equally as important as how it was used when considering banning. A hammer is used to pound or remove a nail but it can be used to cause harm. But that’s not what it was designed and intended. To misuse a tool that isn’t intended for that purpose could be considered abuse of the tool, however when you use a tool for the purpose to which it was designed that’s called using it. see the difference?

  3. I would really like to see some research on a particular group of dogs-in North America, there is a really large group of dogs regularly trained with shock at a distance from the handler-the retrievers andgun dogs. And what I generally see in these dogs is confidence, not fear. But I see these dogs after training, not before.


  4. Thanks for sharing the results of these studies. As a trainer we often describe the fallout from shock collar use to our clients but it is helpful to be able to point to empirical studies to support these arguments especially when positive training methods are a much safer and more viable alternative.

  5. Sophia:

    You dismiss the cognitive aspect of the relationship and its connection as it links with the device. This is a behaviorist approach. I am about to release a very detailed study which will connect the dots. It brings the cognitive communicative aspect of the relationship into full focus. Much of this information is humanized and out of context as it links with nature. Dominance and influence as also completely taken out of context.

  6. Since collars can be set at different levels, were the trainers questioned as to how high or low the collar was set?

  7. Very interesting and informative study! And I see one of the first questions I asked, will the dogs show the same fear response with a different handler, is being addressed. Also wondering more about the overall personalities of the 7 dogs that showed no response. And of course, the ultimate: Do these dogs who receive harsh treatment during training make better guard dogs than those trained with non-aversive methods, weighted heavily with positive reinforcement?

    As an aside, in the past I’ve had my dogs trained to invisible fence. One day I inadvertently walked through the electric field holding one of their collars, by the terminals. What was I thinking? I got zapped, big time, and I can only describe the sensation as not really painful, at least not like any pain I’ve ever felt, but very profound, and my involuntary reaction was to swear, jump, and toss the collar away from me, all in one instant. It was like a big “knock” in my palm, with maybe some tingling sensation. Now, I knew what it was, understood not to do it again, and didn’t have a fear of walking in that area. But for an animal that can’t rationalize, that gets this sensation from out of nowhere, and who knows if they would describe this the way I did or if they felt true pain, it’s a completely different story.

  8. Thanks for posting this article – Its great to see reference to the peer reviewed research. All to often I see people default to the “I read it in a book, it must be true” mentality.

    To be forthright, I am not a proponent of punishment based training, however my experience is limited to working with herding dogs (border collies and australian shepherds) in training for freestyle frisbee, so I am admittedly biased.

    The only thing I wanted to point out is a flaw I see in the study procedure (going from your description). For the observers to watch the dogs receive the shock corrections, and then later assess them on their anxiety in the absence of corrections, completely violates sound experimental protocol. The researchers have an a priori knowledge of which dogs received shocks, and therefore have an expectation of what they will see. This is an extremely biased situation, and unfortunately really invalidates results. While I’m certain their general findings are accurate, I would be curious to see this replicated with a double blind protocol, in which the researchers have no prior knowledge of the type of corrective method used in training.

  9. Sharon-
    Yes they state that they wish they had tested that too. Plus I was not able to tell from the article whether the 7 dogs NEVER showed fear response or they just did not show fear/anxiety response when they got shocked with the collar. It’s possible that they did. And it is interesting that they did not respond to the shock with signs of startle or fear. Because even when I have used “tap” level some dogs are more likely to show startle response regardless of the level. (e.g. if they don’t, then they also don’t respond with a change in behavior). I don’t know if authors asked about shock level.


  10. Laurie-
    I don’t think they asked about level since the trainers may have been changing the level via rheostat during the sessions. There’s only so much you can get non-scientists to cooperate with in terms of info they provide:-). I do know that for some of the dogs I have worked with they either don’t respond to the collar at a low level, or when I find that lowest level they will respond to, they have a startle, fear response (e.g. walking up the levels). So for those readers who might argue that it must be that the trainers are using the collars at too high a level, I would say that no matter how careful you are, because the level you use varies with the dog’s arousal state and the contact of the collar with the skin, you can never hit that just right level 100% of the time. If you have a Labrador it’s much easier in general, but a more sensitive dog…. not. And as long as you can’t do that it can be a scary experience.


  11. Dale-
    Regarding the study: This was not my study. It was a study by the researchers I listed above. They are studying one aspect at a time as any good scientist does. It’s studies like these that build the foundation for our understanding of how our actions affect our dogs. for decades trainers have been working with dogs but never observing and coding the actual responses of the dogs in a scientific manner. In science, it’s not just one study that defines the field or our knowledge of animals. Its a series of studies over a decade or two that paint an clearer picture for us. And often is the work of multiple scientists and labs that come together to build the clearer picture. I have an example in “The Medium is in the Message.”

    Almost forgot to mention that the authors do not talk about dominance because this article has nothing to do with dominance (which is defined as a relationships established by force, aggression and submission in order to establish priority access to resources—see for more information on this.

    If your study makes it past peer scientific review and gets into the Journal of Applied Animal Behavior Science or the Journal of Veterinary Behavior, or Animal Behavior then perhaps we will have the pleasure of reviewing it here. As you probably know, that acceptance and publication process generally takes over a year in the animal behavior field.


  12. Thanks for your comments, Sophia. I’ve also seen differences in how dogs respond to the collar, and some dogs are definitely more vocal than others.

    While there can certainly be differences in how individual dogs respond, I’d like to see a study where all the dogs were handled by one trainer so you take intensity, timing and human emotion/frustration out of the equation as much as possible.

    And I’d also like to know what type of collar conditioning the dogs received before being observed. Less or no collar conditioning, in my opinion, would probably mean more of a negative reaction from the dog. To be a “true” study, I think the same experienced trainer should initially introduce the dogs to the collar.

    And I agree about the observers not being able to tell which method of correction was used, although that may be easier said than done. Interesting article. Thanks!

  13. Jordan-
    They actually videotaped the events and used 1-0 sampling (making it pretty easy to catch and code the behaviors) and then did inter-observer reliability tests So they did not take the measurements in the field. An obvious flaw like that likely would not have made it past peer review.

    But I don’t know if off-hand whether they observed the film in order (e.g. they may have observed the freewalking part first and later then correction part since it’s on film) and I don’t know whether the observers even knew what the experiment was about.I think these things were not mentioned in the journal article. perhaps we can get the researchers to comment.


  14. Thank you for posting this. I am always at a loss as to what to say to someone who claims that the shock of an e-collar is just a benign annoyance. We annoy our dogs all day long when we nag at them, and that has never been very effective for changing behavior. And the argument that we might become better naggers to more effectively influence behavior doesn’t ring true to me.

    So, for punishment to work, it must be punishing. We cannot console ourselves that our dogs do not “mind” being punished. If that were true, nothing would change.

  15. Laurie-
    I agree that a follow-up study with one trainer using different methods would be great.
    The current study was a field study vs a lab study and comes with all of the issues that one sees with a real field study. I”m guessing /hoping their future studies will answer other questions in a systematic manner.
    The benefit of the field study is that it is looking at what people actually do, rather than what we ideally wish they would do though. So it does make it applicable to what happens in real life grin. And the trainers in the study were “professionals” in their field and the certifications they are working for were well-established. (which doesn’t really necessarily say anything about how “good” the trainers are regardless of how “successful” they are at getting the certifications.Successful trainers could just select for dogs that are suitable to their training style).

    I was surprised that they would see a difference between electronic collar and the other “harsh” methods that were used.


  16. It would be very interesting to see a comparison study using the electronic collar that emits a citronella spray used in place of the shock collar. The spray is distracting for the dog because of the sound and strong smell, but it is not painful. What is effective as a training aid, distraction from an incorrect response or punishment via shock? The pain caused by a shock collar has the potential for so many unwanted behavioral side effects.

  17. I’d like to see the study repeated with a single trainer also. My big concern with this study is that “the researchers had no influence on the methods and aids used.” We have 2 groups of dogs being trained to do different things by 2 groups of humans. For all we know, the only significant difference between the groups may have been that the handlers using leash corrections were more skilled than the shock collar handlers at making their corrections contingent on the dogs’ behavior. Dogs who quickly learn to avoid punishment will almost certainly be less stressed than dogs who have trouble learning to avoid it.

    The idea that shock collars don’t cause pain has never passed the laugh test with me, but I don’t see how this study gets us closer to demonstrating that scientifically.

  18. Interesting study, although it seemed as if there were too many variables to get definite answers. From what I noticed in general, there are a lot of people out there who don’t know how to use e-collars properly. Were the handlers experienced e-collar users? What were the dog’s personalities? Soft? Hard? It seemed like the dogs who yelped and screamed are on a level that’s way too high, or the dog were too soft for e-collars. I would rather see the study repeated by trained e-collar instructors, than just people they chose out of an obedience class. When I see handlers who use e-collars properly, I see confident dogs who do not show fear or cower, so perhaps the study was using inexperienced handlers?

  19. Talk about observer bias.

    This study told me more about the expectations of those conducting it than it did their subjects. They went in with an agenda and found data to support it, amazing.

    yet another junk science study. Strangely the Netherlands seems to churn out a lot of these……

  20. Dr Yin, Thanks for posting this research paper. As with any paper there needs to be follow-up studies with similar findings before we can draw definitive conclusions. We’ve all seen studies where the results were due to errors in research design or methodology. Other papers have been published on the subject and it is my hope that you will post them on your blog for more debate.

    I don’t think Lou will like those papers either. Lou says this study is no good because of a quote he provides from Steve Lindsay. Steve Lindsay has written 3 books but I can’t find any citations for work of his that appeared in peer-reviewed journals – where rigorous standards are set for publication. In the 1990s, I was at the initial meeting American Humane Assoc convened to establish humane dog training techniques. The facilitator asked if the group assembled could agree on training techniques that everyone could agree were inhumane. Mr Lindsay actually objected to saying kicking a dog was an inhumane training technique. For that reason, the quote Lou provides means nothing to me. It is simply an opinion from a person I know I have fundamental differences with regarding what constitutes dog training.

    I know that most people have not had the opportunity to actually read this paper unless they have a subscription to Applied Animal Behaviour Science or are willing to pay $35.00 to download it from Elsevier. It’s clear there are some misconceptions about the methodology. The researchers did not decide the level at which the shock collars were set. This was left to the dog trainers who are considered experts in their field. One of the things we hear repeatedly about using punishment in dog training is that in the hands of professionals, you can see good results. Although using multiple trainers in this study introduced many variables, it raises questions to about the cost to the dog of using this aversive, the lasting effects of its use and in the end if it is counterproductive to producing a well-trained effective working dog.

  21. This study is hardly “new” as you’ve called it. It was published in 2004, some 7 years ago. It’s one of the worst studies ever done on the Ecollar. It used completely subjective measures by the researchers, allowing them to draw whatever conclusion they wanted. There are many other reasons that dogs will do the things they attributed to stress, yet they failed to allow for any of them.

    Stephen Lindsay, the author of the “Handbook of Applied Dog Behavior and Training” said this about the study cited, “This was an intentionally deceptive study, designed to reach predetermined conclusions. And that couldn’t have supported the findings that it reached even if it had been competently done.”

  22. Lou-
    Actually I do state in the article that is was published in 2004. But while the article is not new to scientists, judging by the thousands of people who have already read this blog article apparently it is new to the general public. When I saw the article posted, I did think about taking the “new” out of the title, but by then it had already been read by many people who would have already read the date of the research publication anyway.

    On another note, people who consider behavioral measures to be subjective must think that most behavior research is no good. The methods they used are standards methods in which they have an ethogram, video tape the situation and measure (record) when particular behaviors occur. They chose 1 zero sampling which is not my personal favorite, but which is used commonly in field-type studies.

    In addition, they did two-tailed statistical cases in most instances meaning they were looking for a different in either direction, not just e collar is better than choke/chain and pinch collar. That in itself says something.

    Realistically, there are a lot of other studies that should be done. One study in a limited situation does not generalize to all situations. Hey comparing electronic collars to training in which the trainers admitted beating/hitting their dogs is not the best comparison. But it’s probably a good representation of what happens in “professional training” situations in some countries. And I’m sure all of these trainers think they are the best their country has to offer:-).

    So, do I think this study is perfect? No. There is no perfect study in animal behavior. But the findings are useful and they spark a lot more questions which is what a good study does. It’s just one study, hopefully the first of many.
    In future blogs I will be giving examples of historically important studies that turned out to be incomplete or even flawed and that required a body of additional studies to provide the entire picture. It’s good that this study has sparked so much interest and even controversy. It’s good for people to read and think critically rather than just agreeing with the last thing they read or the things that only match with their current beliefs.

  23. Dr.Yin the study did not answer the question that you asked in the title of the blog entry. It’s impossible for it to do so. Those researchers were not interested in levels of stim that are used with modern methods, the ones that dogs find only annoying. Rather they were interested ONLY in the highest level of stims used in dog training. They purposefully studied dogs that are trained by people who are famous the world over for using the highest levels of stim. They purposefully used (but somehow avoiding naming) an obsolete Ecollar that is no longer made. It provides higher levels of stim than any of today’s collars AND it has its contact points on either side of the dog’s neck (rather than today’s units where the contact points are only about 1 1/4″ apart). This means that for this study MUCH MORE tissue was involved that is involved with today’s units and today’s methods.

    This study has nothing to do with modern method of using modern versions of the tool under discussion. It’s simply another case of GIGO … Garbage in – garbage out.

    As to your comment that it’s “new to the public;” it’s been discussed quite a bit on many K−9 forums and emails lists around the world since it was published. On just one site it’s come up 7 times in the past couple of years. On another forum, out of the UK, it’s been discussed in seventeen threads. And those are just the conversations that I know about.

    Stephen Lindsay also said this about this study, “Although they offer no substantive evidence of trauma or harm to dogs, they provide loads of speculation, anecdotes, insinuations of gender and educational inadequacies, and derogatory comments regarding the motivation and competence of IPO trainers in its place.”

    And this, “Instead of instilling social aversion and anxiety as suggested by the authors, competent electronic training may actually promote social attachment, reward, and safety, With the behavior- contingent cessation or avoidance of ES, dogs experience immediate emotional relief that subsequently merges into a state of progressive relaxation incompatible with social aversion and fear – a sequence of opponent emotional effects contrary to those alleged to occur n the case of working dogs exposed to ES in the context of training.”

    And this, “The authors made no attempt to blind themselves to the experimental and control groups. Unlike the Tuskegee studies, they weren’t measuring cortisol levels in the saliva and/or blood – they were subjectively interpreting behavior. And they knew which dogs were being subjected to which training protocol, when they were doing the interpretation.”

    The Tuskegee study that he refers to is still out for peer review but preliminary results show that electronic bark collars were not only effective in controlling excessive barking, but that they also did not cause any lingering adverse physiological effects.

    This mirrors the finding of another Ecollar study, done by Schalke (available online in 2006) where it was found “… animals, which were able to clearly associate the electric stimulus with their action … did not show considerable or persistent stress indicators.”

    In both of these studies, once the dogs learned what was causing the discomfort (as any modern training program will do) the dogs were not stressed. Rather than the subjective observations to determine if the dogs were stressed, as was the case in Schilder, (where it’s easy to make a study say whatever you want it to say) these studies measured cortisol levels and so bias was not possible.

    I suggest you spend some time investigating what’s really going on in the dog training world rather than accept nonsense like this study from Schilder. It’s been widely discredited, and with good reason.

  24. The Schalke study is interesting and partly what I would expect, but also had some findings that surprised me. The study involved 3 groups of beagles in a highly controlled lab environment. The study was conducted over a 7 month period and and the dogs had daily 1.5 hour training sessions even before any aversive at all was used. And then the study found that dogs shocked with good timing in one type of situation had only low cortisol elevation. Those shocked with poor timing had high level cortisol elevations (and face it, the general public has poor timing). Those shocked in a 2nd situation with good timing also had high cortisol elevations but not as high as the randomly shocked group. I plan to summarize this paper more fully in a future blog post.

    The authors conclude,

    The results of this study suggest that poor timing in the application of high level electric
    pulses, such as those used in this study, means there is a high risk that dogs will show severe andpersistent stress symptoms.
    We recommend that the use of these devices should be restricted with proof of theoretical and
    practical qualification required and then the use of these devices should only be allowed in
    strictly specified situations.

  25. Regarding quotes that are from Steven Lindsay’s book (Since 2 volumes of his book are dated before the Shilder article I have to assume it’s’ from the 3rd book? However the third book is not that long after Schilder and books generally are in production for a year after they are written–so it’s cutting it close). The quotes appear to be inflammatory and exaggerate what the Schilder and van der Borg actually said.

    I think it’s best to go back to the original source and draw conclusions based on what was actually written and the context in which it was written. The article can be downloaded at the Journal of Applied Animal Behavior web site for a fee. Or if you are affiliated with a university that subscribes to the journal, you can download it for free.

  26. The researchers of this paper have concluded that “Being trained is stressful, that receiving shocks is painful experience to the dogs, and that the S-dogs (the dogs who received shock) evidently have learned the presence of their owner (or his commands) announces the reception of shocks even outside the normal training context.” So yes the question in the title of this blog was answered by the researchers. How was this determined? The researchers give this explanation for their conclusions:

    “Inspection of Table 3, depicting immediate responses of dogs to shocks, shows a number
    of behaviours, that in the literature are connected to pain, fear and/or submission. Lowering
    of the components that make up the posture of the dog (ear and tail position and position
    of head and body), are related to submission and fear (Fox, 1974; van Hooff and Wensing,
    1987; Beerda et al., 1997, 1999, 2000) and harsh training (Schwizgebel, 1982). Beerda et al.
    have shown, that, even in the absence of a person or dog, dogs lower their posture when
    confronted with an unexpected aversive stimulus. This shows that lowering of the posture
    is not an expression of submission per se, but certainly is connected to fear. They also have
    shown, that certain behaviours (e.g. lifting a front paw, tongue flicking, licking lips and
    vocalisations) are connected to either chronic or acute stress.
    Vocalisations are also indicative of pain (Hellyer, 1999; Noonan et al., 1996; Conzemius
    et al., 1997), especially the higher frequency squeals, yelps and barks. Biting attempts can
    be interpreted as pain-induced aggression (Light et al., 1993; Ulrich, 1966; Polski, 1998).
    A characteristic, swift head movement sidewards and downwards often follows a shock as
    does a swift avoidance action. Both these reactions also indicate that reception of a shock is
    unpleasant. All in all these responses show that shocks elicit fear and pain responses. This
    means that shocks are not just a nuisance, but are really painful. In spite of the enormously
    high arousal of the dogs in this type of training, that very likely implies an increase of
    analgesia, receiving a shock may sometimes be perceived as a traumatic event by a dog.
    One of our study dogs still behaved as though it received shocks during protection work
    although the last shock was delivered 1.5 years before!”

    It is my sincere hope that anyone who uses aversives in training can recognize signs of pain and distress in dogs.

    As Lou notes in his post, the type of shock collars in use were not mentioned. So how does he know which collars were used? The researchers did not purchase any of the shock collars that were worn by the dogs observed in this study. The shock collars that were used were the ones that the individual handlers had purchased and were routinely using to train their dogs. Did they all use the same collar? That is unknown. This study was conducted in 2003 – less than 10 years ago – so I would assume that these collars are still available on e-bay if nowhere else.

    I believe the “Tuskegee study” referred to is the one by Janet Steiss, Caroline Schaffer, Hafiz Ahmad, and Victoria Voith and was published in Applied Animal Behaviour Science in 2007, entitled, Evaluation of plasma cortisol levels and behavior in dogs wearing bark control collars. As the title suggests, this study evaluated bark activated shock and spray collars. Not the remote shock collars evaluated in the Schilder et al. study.

    There is a huge difference between shock collars that are triggered by barking and shock collars that are triggered by humans using a remote control. That huge difference – humans. With bark generated shock or spray collars the dogs can learn quickly to avoid the shock or spray by not barking. Unfortunately, with humans in control of the shock that is not always the case as was noted in the Schilder study.

    “Secondly, we have some evidence that getting an order, which previously was immediately followed by a shock or shocks, had obtained a negative connotation: for example one dog, shocked immediately after getting a “heel” command, yelped after getting the next “heel” commands without being shocked.”

    This may be the reason the authors offered in their conclusions,

    “Trainers and handlers should study learning theory far better and review the structure of the training in order to teach the let go command in an earlier phase and to reduce the number of mistakes. They should incorporate more rewards during exercises. Also, less temperamental and less forceful dogs should be bred. This also would decrease the chance that dogs make mistakes for which they could receive punishment.”

    This is reported in Steve Lindsay’s quote as,

    “Although they offer no substantive evidence of trauma or harm to dogs, they provide loads of speculation, anecdotes, insinuations of gender and educational inadequacies, and derogatory comments regarding the motivation and competence of IPO trainers in its place.”

    That seems to be an over-reaction to what was written.

    Throughout the quotes attributed to Steve Lindsay posted above, there is not one reference listed. There may be flaws in the research but the authors took the time to design a study, collect data and analyze their results. The only thing Lou has provided is Steve Lindsay’s opinion. These opinions are offered without specific reference to what was actually printed in the study. Since this paper is not available to the general public without payment, I think these opinions expressed need to be side by side with the actual statements made by Schilder et al. People forming their opinion of the study by the quotes of Mr Lindsey are not getting an accurate reading of the content.

    For me – I’ll go with the science, know that no experiment is perfect and look for additional studies to shed more light on this controversial subject.

  27. How about a new study? One done by someone who does not have an agenda? One done with a collar on the market, available to JQP?

    My company trains between 200 – 300 pet dogs per year with the use of remote training collars. Some of the dogs are trained by myself or my trainers, most are trained through coaching the average pet owner to do it themselves.
    I would love to find a scientist without an agenda to come here and measure what is or is not happening.
    Personally I don’t believe for one second that these dogs are undergoing significant stress or pain. If they were I doubt I would still be in business after all these years and people would not be knocking down my doors to learn more about how to use a remote collar to train their dog.

    However, If a study done on these dogs proved me wrong, I would close up my doors and find another career.

    It’s an open invitation and someone should do us all a favor and prove once and for all whether or not “shock collars” used via the modern methods with modern tools on the “average pet” cause pain or long lasting behavioral problems or not.

    These are the dogs that we really all have the main concerns about right? The ones that JQP owns and is going to run out to the store and purchase a product and try to use it to solve a frustration they have with their dog.

    Perhaps I am way off base but seems it is more enjoyable for the debate to continue to rage in the professional community rather than provide real and usable evidence one way or the other for the average dog owner.

    Robin MacFarlane

  28. Seems to me the planners of this study assumed all of the subjects used the electric collar in the same way (and by “same way” it sounds like many of them used the old-school, hurt-the-dog for non-compliance method). The amount of variables in this study render it almost completely unscientific.
    It’s kind of like handing a set of tools to 32 different mechanics, and observing them work on cars of in varied state of disrepair, different makes, models, years and “studying” how the tools work. This would be ridiculous, no? Not an effective way to prove or disprove a theory.

  29. If any company is really interested in further study on electronic collars you should donate collars AND study funding. I’m sure there are a number of veterinary behavior residents or applied animal behavior graduate students who would be interested. I can any funding group in contact with potential study groups if you are interested. Or you could approach the Schalke group. You want to bear in mind that the $$ would be up front and you could not suppress results you didn’t like. And most likely while you might like some portions of the results you might not like others. It seems like what you want is a blanket statement that says that your “negative reinforcement” low level shock technique is good for everything and when used by all people and on all dogs all the time. That would be like saying. “This drug works for everyone with this disease no matter what and has no side effects at all.” It’s the American way of thinking but not reality.

    Also note that studies are not cheap. The planning stages and practice runs/pilot studies alone probably cost a couple of months’ pay for the researchers working on the project. The Schalke study ran 7 months —that’s 7 months of boarding for the lab dogs, 1.5 hours per day of training. Then data collection hours + the 100s of hours analyzing the video and then testing for reliability between observers (or training the observers)

  30. A number of electronic collar users have referred to ‘modern techniques.’ I just want to point out that these techniques are not new at all. I learned the low level shock collar use as negative reinforcement /now called tap technology, 15- 20 years ago when I took some lessons and seminars from a trainer in my area (who does travel nationally). I guess he must have been progressive. 98% of people I see using the collar do not use it that way or do not use it “correctly.” Or they still get yelps out of their dog during some situations.

    There are at least 2 Dog Whisperer episodes where the collar is used and in one the dog yelps and redirects aggression to the owners and then hides. In the other the dog does not yelp but it runs away and tries to hide. Do you expect the general public and most dog trainers to do better than the Dog Whisperer?

  31. Charles:

    No the variables do not render it unusable. increased variability generally provides results that are just less likely to be significant. That is with more variability it means that you are less likely to see differences between the two methods.

    The main issue is that the study is not answering the question that YOU are interested in. It’s asking/answering a question regarding what happens in a field situation. And for what they are asking, the way they tested it is ok.(and from what I see when I see general public, schutzhund trainers, etc using the collars this is how I see them using it–same as in the study).

    Electronic collar companies and other groups should put up more funding money to answer the many other questions regarding electronic collars instead so that more studies can be done instead of complaining about one study that other scientists view as a good “start.” I will be covering other research on shock and shock collars. Perhaps some of those findings will be more to your liking.

  32. Rob,

    You and all shock collar companies already have the ability to educate. You could require that owners only are able to use the collar once they have gone through a course (not just a video but some type of course where skills could be evaluated) before using the collar. And your collar could come with video or a warning showing what happens when they are used “incorrectly.”

    That is, rather than interpreting the article as a blanket “bad” another interpretation is that electronic collar companies need to band together an make it clear that you know that shock levels can cause pain and that here are very instructions on how to use the collar correctly. Do you do this already? I have not seen any collars that do this but I have only looked at some of the most common/popular brands. The videos I have seen show pretty crude training technique.

    Realistically the instructions must be geared towards knowing what the owners are likely to do incorrectly and give them timing exercises to practice on without their dog first. After having videotaped many many owners learning new techniques and analyzed what they have trouble with, it’s clear that most owners have very bad timing and technique (yes, even just pressing a button) at first. They need to be shown what not to do and the consequences to the pet when they perform the technique incorrectly. They also need top instruction in order for the technique to improve.

  33. Sophia,
    I think you are addressing me, but not certain of that…my name is Robin not Rob. Assuming that you meant the response for me…

    I do educate. I teach my clients here locally, (they are not allowed to buy collars from me without education) and I also teach workshops internationally.
    Do I wish the collar companies would provide more education, yes, I do. I have pushed for it for many years. Instead of continuing to expend energy in that direction, I am doing what I can to educate and have used my own resources to create videos and online help in addition to my local and international teachings.
    Should collar companies (by the way, I am not a paid employee of any of them, however I do consult with a few) require users to first attend a course before being free to use the equipment? In a perfect world, perhaps so. It would certainly require MANY, MANY more professionals to be educated on how to use them so there were adequate resources to seek out worldwide.

    But then should the same standard apply to all tools? Is there not recipe for misuse with most any tool? (I cringe when I’ve seen head halters attached to extendable leads) I’ve seen people click/treat growling, thinking they are distracting the dogs attention from the trigger…What about other aspects of dog care and handling? Should owners be required to attend a course before feeding raw or before adding suppliments to their dog’s diet? Before administering pharmaceuticals or self vaccinating?
    Heck, in a perfect world people sould be required to attend a course before getting a dog. I don’t know where we draw the line. It is what it is at this point and human nature perhaps to sit on thrones of self-righteousness coming up with reasons for rules regarding what is or is not acceptable. I would like to see our activities directed much more at providing massive education to dog owners about all their options. I personally think most are wise enough to make good decisions when presented with all information, rather than slanted bias.

    I am only guessing…but the sentiment surrounding e-collars comes from a belief that they carry much greater possibility for misuse in JQP hands than other tools. But if this is an actuality, how come, given the huge number of electronic training aids sold in this country, we don’t see proportional numbers of messed up dogs due to their use?
    I have interacted with literally 1000’s of people who are either currently using e-collars or have used them in the past. Of those who first *tried it on their own* the over-whelming majority have told me variations of the following stories: 1. that it did not work, the dog showed no response* (which is generally due to improper fit) 2.It worked fine and their dog happily wears the collar 3. if they did use a higher aversive level, (not what I am advocating, they never understood collar conditioning in the first place) they only had to push the button a time or two and the behavior they were frustrated by stopped

    I certainly wish everyone understood a variety of ways to train their dog..however studies like the one discussed here seem to only fire the fuel of groups wishing to ban e-collars or put pressure on retailers not to sell the products. In the end this does not help the consumer solve the problems. Frustration will be taken out one way or the other. Or the dog will be relinquished. At least that has been my experience through the tales I hear.
    I am in no way suggesting that e-collars are never painful or that someone could not create long term negative behavioral consequences by using one. However, I believe it is the rarer case. I do not think the study went to any extent to point out the variables that should of been taking into consideration. It is my opinion that the study was created to suit an agenda and because it did so so effectively it is used more as propaganda than to aid learning amongst professionals.

  34. It is true this study has come under scrutiny – as any do. I also agree that it will be hard to find a study to suit everyone. “Conflict of interest” will always be cited, if nothing else, because the people most likely to study such a thing are those with an actual interest in the outcome.

    That being said, I have a question. If we can train without pain, why shouldn’t we? If we can train killer whales, elephants, big cats, dolphins, chickens, horses and virtually any other animal without force, why can we not agree to do so for dogs? What is unique about dogs that would make them “learn better” and “feel calm and secure” by being trained through avoidance of pain?

    I would argue even the simplest behaviors taught for husbandry of large mammals are more complex than what an average dog owner requires of a pet.

    Can any item be misused in the wrong hands? Yes. Is that a valid argument to justify knowingly endorsing it? I am not convinced.

    Respectfully submitted.

    1. Training Killer Whales
      Have you seen the documentary ‘Black Fish’
      Yes, they ‘train’ them, but with dire and lethal results, and the whales are virtually ‘crazy’ afterwards…. I am enjoying this thread and relating a lot of what is said to things I have seen… One friend uses E collars on her dogs and they respond in a positive manner and altogether I think it works well. I have another friend who has a young Standard Poodle who’s way of showing affection to his mistress is to rush this older woman and bite her… I think an E collar could be very helpful, but know they would not consider this. a few years ago I saw a young greyhound ‘bump’ a competitor while at a Lure Course, This dog disappeared for 18 months, then was entered again. When she ran with my dog she did not follow behind the lure, she paralleled my dog, about 10 or 15 feet to the right, and half way through the run she veered away and jumped into her owner’s vehicle Obviously this dog had experienced an E Collar training session that was very negative. She no longer interfered with the competitor, but she was not able to perform the requested task,,, to chase the lure cleanly and competitively . Ann

  35. I know there are shock collar users who have tried the collar on themselves and feel confident it is not painful or abusive. However, I wonder if any of them have ever tried learning a skill by wearing one.

    The experiment would be to have someone else they felt to be a skilled trainer teach them how to do some new, reasonably complex task using shock to train as they would a dog. I wonder how the trainee would feel about the experience and how he/she would feel about the trainer at the end.

    It would be a variation on The Training Game that clicker trainers often play to remind themselves how difficult learning can be when we are not clear.

  36. Dr Yin, I’ve been reading through the comments and see some that disparage the researchers for trying to test their beliefs about electronic collars. I believe this attitude shows a lack of understanding of the scientific method. Basically there are 5 or 6 steps to the scientific method and they include 1) ask a question, 2) do background research (search the pertinent literature), 3) CONSTRUCT A HYPOTHESIS, 4) test the hypothesis with experimentation, 5) analyze the data collected and draw conclusion, 6) share what you learned – usually in scientific journals. describes simply how one goes about constructing a hypothesis. As the scientists read through all the previous studies that were done on their topic they will start to form opinions regarding the answer to the question asked in step one. This opinion IS the hypothesis! So forming an opinion about what the experiments will prove is an integral part of the scientific process. It is NOT a sign of bias on the part of the researchers. The opinion is then tested in step 4 with experiments specifically designed to provide answers to the questions. Donna

  37. Hi Susan,

    That is a great question and one that I have participated in. I taught a workshop in Athens, Greece a few years back. Most of the people in attendance were bilingual. However, we had two attendees who did not speak a word of English. I did not speak a word of Greek. I had an interpreter for the entire 2 day workshop but in the beginning I did one exercise without the interpreter’s help where I trained one of the non-english speaking attendees to touch his toe. I used a Dogtra collar, determined a level he could feel, then taught him to touch his toe to the cue word touch. The exercise took about 60 seconds and he knew that “touch” meant to touch his toe and keep his finger on his toe until I “realeased” him.
    It was a very simple exercise of staying in a position. A huge smile came across his face because not only had we communicated and crossed a language barrier…he understood the concepts beautifully because he had experienced it exactly the way his dog would.
    He went on through the entirety of that workshop to be one of the best handler teams present and got amazing results with his dog in 2 days time. The grin never came off his face and the wag never left his dog’s tail..he totally understood how the e-collar could communicate without pain.
    It was a very cool thing to experience.

    An e-collar used at non painful level is similar in many ways to using a clicker. But not only do you have the ability to mark “yes” very clearly, you also are marking “wrong answer” very clearly.
    Think of it as the game of hot and cold many of us played as a kid. You give the dog information through pressure on, pressure off, sensation on, sensation off.
    One of the things the skilled trainer does is to not let the dog *guess* about how to turn off pressure. We guide (leash, lure, body language etc) The frustration level is very low because the communication is so clear, thus the dog learns rapidly. Add in rewards that matter to the dog in question and you make incredible strides in training in very short time frames, particularly for simple tasks like a recall. Most of our average pet owning clients have an off leash, outside, with fairly heavy distraction recall in about 2-3 weeks.

    all the best,

  38. There is much criticism that this study is imperfect because it doesn’t address (amongst many other things) the dogs’ personalities, the appropriateness of the shock level, and handler skill. Yet, even when you’re studying something “objective” like cortisol or blood serum levels, there is ALWAYS inherent bias. It can manifest itself in the group from which you base your study, HOW you acquire participants, what words you use to inform them of their rights/desired actions, the specific data that you collect, and the parameters that you set. There are scientists whose entire life work is literally to sit and think of ways to eliminate bias in experimental studies – not just the “subjective” criteria for which people have been criticizing this study, but data/collection bias, as well. Statistics is a rigorous study of data, and peer-reviewed journals aim to keep that rigor. Moreover, a rigorous behavioral or social-cultural study is one that uses statistics to control for the effects of individual variability. And, honestly, think about it: biology is the study of the VARIABILITY of life. This is the heart and soul of biology! Good biological studies simply control those variables through data collection and statistics that determine whether results are significant or due to *chance*. This experiment (ALL experiments) are not out to “prove” a point. They simply aim to show that the findings of an experiment imply that the manipulation of a variable (in this case, shocking an animal) produces negative effects that are likely NOT due to chance. Thus, we can interpret this as evidence that using shock collars to mitigate behaviors can induce fear and stressed, displacement behaviors. After all, is this not a tenet of learning theory??? I’m honestly surprised that we need to publish studies that show, “Hey, if I shock you, you will develop fear.” FYI, in laboratories that deal with creating fear in ‘simpler’ organisms like snails or mice, they use one tool to do so and it only takes several tries: they shock the subject. Does it truly take a leap of imagination to suggest that the same holds true for domestic canines?

    I highly suggest that everyone pay or find access to the actual study and read the ACTUAL journal article before eating up the regurgitated interpretations of other people. It is available via Journal of Applied Animal Behaviour. College, graduate, medical students or anyone who is involved with a higher education organization should have access.

    On that note, I’ve read the actual journal article, and I’ve noticed some points of contention that are easily answered within the article. Regarding the “subjective” approach of observation: The observers did NOT know when or how the handlers would shock the dogs. Directly from the article: “We had no influence upon the methods and aids the trainers used during the previous training sessions we observed. During the previous training term sessions, we only asked the handlers to walk their dogs and to perform some standard obedience exercises, as explained below…Afterwards, we checked with the trainers to accurately determine the number of previous shocks term that had been given.” Regarding variability of handler skill: Some of the control dogs were being handled by the same person that handled shocked dogs, as well. Regarding the variability of the dogs’ temperaments, these dogs were all police service dogs trained for the IPO/VH3 certificate. To survive protection training, I’m pretty sure these dogs all had fairly “hard” temperaments, as one previous poster liked to put it. I think a more scientifically accurate way to put it is: These were all highly trained dogs with high tolerances to pain/punishment all handled by very skilled handlers. And, YET, they showed distinct fear behaviors upon being shocked.

    Now, put a shock collar in the hands of an inept handler (much of the American public!) and place the collar on a dog who is less tolerant of pain, and it doesn’t take a statistician to guess what might happen next.

  39. Years ago, I was on a Search and Rescue team. Specifically, I was on the dog unit. I was training my first dog in SAR. I am a lure reward trainer. I use as little force as possible to train dogs. This dog was chasing critters, rabbits usually. Out of frustration and coaxing by team members I bought a top of the line shock system by Innotech. They have me convinced if I do it right, its safe for the dog and will not cause any problems. I read the instructions and I set it up so Jester would only get enough shock to get his attention. I think he was on level 2 of 7 levels. No pain, just like a tap to get his attention. I’m actually feeling pretty good about it. Jester would start sniffing a critter and I would tap him. He’d look back at me and I would direct him to “find”. Handlers learn the difference between following human scent and critter scent. Alls good right?

    Until I tap Jester and he yelps loud and even jumps off the ground. What freakin was that? I hurt my dog. Best I can figure, the control has a boost button and I hit the boost button by mistake. That means Jester got hit with level 4, instead of level 2. Tape a penny over the boost button and that won’t happen.
    Another time, the same thing. I hit the button and Jester yelps and jumps. Except this time, somehow the level has increased. So, tape cardboard over the level buttons. I’m frustrated, I love my dog and this stupid thing has hurt my dog twice now. Yea, it helps me keep him off the critters most of the time.

    The last straw. The controller is attached to a string so you can wrap it around your neck. Jester gives me hugs sometimes. He jumps up when I am down and puts one paw on each side of my head. Except that one of his paws lands right in the middle of my chest and he gets tapped for giving me a hug. THAT IS ENOUGH FOR ME! I have never used that peice of crap since. If Jester even sees it he cowers.

    The next time I tap, that collar will be around a human neck. Any tap trainers willing to put one of these things around their neck? I’ll set it on 1-2, I promise. I’ve yet to have anyone agree.

    I started training Jester in air scent with a 30′ long lead. He knows if a rabbit runs, and I say off, he checks in with me for a reward and then we continue our search.

  40. Just a funny story about inappropriate use of e-collar. My husband trained bird dogs, and bird dog trainers use shock collars often. Sometimes they use them to call the dogs in if they have strayed. So this guy can’t find his dog, and is getting frustrated, so he calls the dog and continues to turn up the intensity on the collar. Finally he finds the dog, laying on the ground next to a cow. After that, every time he used the collar on his dog and the dog saw a cow, he’d fall over on his side.

  41. I would have to agree that the stim level was much too high if the dogs responded this way.
    Even in doing studies scientists should know and uderstand the correct use of the tool they are studying.
    There is also nothing to say how they determined what the comparable choke or prong level was. How was that determined. I imagine a few dogs would have been choked to death or serioulsy injusred to have a comparable effect.

    That is the trouble with often sets to prove a certain theory and loads the test to give the result it wants.

    This is much like sweetner studies saying they cause cancer in rats (yes at 600 times the usual or typical ingestion)..well give me a friggin break..
    You call this study “science”??

    I dare you to do a FAIR study and not load the test.

Leave a Reply

Your email address will not be published. Required fields are marked *