Have you ever considered using an electronic collar (aka e-collar or shock collar) to train your dog? Perhaps you advocate against their use or feel they can be helpful in certain situations. The use of e-collars (one form of positive punishment) in dog training has been hotly debated for the last several decades. One school of trainers uses little or no (positive) punishment in training and focuses on positive reinforcement to teach the dog the desired behaviors. Another school of trainers argues that positive reinforcement techniques are not as effective as techniques that include punishment and advocate the use of both punishment and reinforcement in training (see last week’s post for more on positive/negative punishment/reinforcement).
We can argue until we’re blue in the face, but we’ll never really settle this argument without good, impartial data from experimental studies that examine the impact and efficacy of both techniques. A recent study at the University of Lincoln (Cooper et al., 2014) in England did this by comparing the results of training using e-collars with training using only positive reinforcement. Let’s see what they found.
The published paper included results from two studies. The first, smaller study was a preliminary study, followed up by the second, larger study. The first study used nine dogs, trained with the use of an e-collar by one of four different trainers. Of those four, only one of the trainers used the industry standard for training with an electronic collar. This involves starting at the lowest setting and gradually increasing the intensity until a mild response is triggered in the dog. It also requires using pre-warning cues that signal a potential stimulation if the dog continues a behavior. The other three trainers started at the highest level of stimulation (or, for one dog, in the high end of the range) and did not use warning cues. Dogs in both studies were presented by their owners for training due to chasing livestock or wildlife and/or failure to come when called from a distance.
In the second study, there were three groups:
- dogs trained with e-collars by industry approved trainers
- dogs trained by those same trainers, but without e-collars
- dogs trained by members of the Association of Pet Dog Trainers UK (APDT UK) without e-collars. The APDT UK opposes the use of e-collars.
It’s important to note that the trainers in group 1 used e-collars according to the industry approved methods mentioned above (using lowest effective setting and pre-warning cues).
The study assessed the impact of the different training methods by using three different measures:
- Observations of the dogs’ behavior (preliminary and main studies)
- Measurements of stress using salivary cortisol (preliminary and main studies)
- Owner reports post training (main only)
The findings of the study can (and should) be used to guide our choices regarding training methods.
Dogs in the preliminary study showed clear evidence of distress. Overall, dogs displayed the following in the period of time following a stimulation:
- Abrupt changes in movement – such as running to walking or stopping. The one dog being trained on lower setting with warning, showed a subtler reaction – a change in orientation and posture.
- Increase in vocalizations, including yelps and whines
- Increase in percentage of time tail was tucked (2% of time before à 20% of time after)
- Increase in percentage of time “tense” (10% before à 50% after)
- Decrease in amount of time investigating (20% before à 5% after)
- Increase in interaction with owners (14% before à 56% after)
- Elevated cortisol levels (the “stress” hormone) after stimulation
The researchers conclude that the “results are consistent with exposure to a significant short term stressor in the form of an aversive and probably painful stimulus during training.”
Differences between dogs trained with properly used e-collar or without the use of e-collars are still present, but more subtle. Compared to dogs trained without the use of e-collars, those trained with e-collars:
- were more tense
- spent less time sniffing and exploring their environment
- were no different in terms of lip licking and yawning (common signs of stress)
- were no different in salivary cortisol levels
Training with and without e-collars was equally effective at addressing the problem. That is, e-collars did not provide an advantage over positive-reinforcement training. There was no significant difference between groups in the owners’ reports of efficacy. Overall, 88% of owners reported improvement in the dog’s overall behavior and 92% reported improvement in the specific problem for which their dog had been referred. However, owners in the e-collar group were slightly less confident in applying the methods used on their own.
A number of important conclusions can be drawn from this study:
- The specific method of training with the electronic collar matters. Poor use of the e-collar produces far more distress in dogs than proper use of the e-collar.
- Even proper training with the e-collar does not provide an advantage over training with positive reinforcement. In addition, dogs trained using positive reinforcement showed less distress than those trained using e-collars.
- Owners of dogs trained using e-collars were less confident about carrying through on training at home. This indicates that these dogs are probably at higher risk for experiencing improper use of the e-collar once the owner takes over training which, as mentioned above, can be very upsetting.
My opinion. The primary argument in favor of using e-collars is that they result in faster, more reliable learning than positive reinforcement alone. If electronic collars can cause pain and distress to dogs, do not result in faster or more reliable training and are at risk of being misused, then trainers and owners should turn to positive reinforcement instead. Perhaps you already had come to this conclusion on your own. If that’s the case, I encourage you to use the facts from this study to inform other people of the risks of e-collar use and that better and more humane options exist.
Are there other studies on dog behavior you’d like to see done? What kind of questions would you like to see addressed?
If you’d like to receive immediate notification when new articles are posted, please enter your e-mail under the subscribe box on the right. You may need to scroll up to see the box.
4 comments
I have a few questions about the study. First off, why is there only one group properly taught to use e collars? Why did trainers not ensure the owners were sufficiently schooled in using the collar so that training would remain consistent? Is it possible that the owners’ lack of confidence increased the dogs’ stress levels?
Also, many of the things you noted are normal for e collar training. A slight movement right after a stim is applied? That’s normal, it’s like scratching an itch. Vocalizing is common as well, especially if the collar wasn’t being used properly (like in the study), or the dog isn’t conditioned to the collar. Hell, it’s common even without an e collar.
And why are results based on inexperienced owners’ reports instead of that of an experienced trainer?
Hi Vex, thank you for your questions! Unfortunately, I cannot directly answer many of them as I am not the one that conducted the study. However, every study is imperfect as we cannot possibly address every single important factor in any one study. That’s why repetition is so important.
Why is there only one group properly taught to use e collars?
They state that “For this preliminary study, trainer contact details were obtained from publically available marketing (e.g. websites, magazine advertisements) or through collar manufacturers.” From this, I assume they contacted a variety of trainers and included the ones that actually agreed to participate. It sounds like they then had the participants use their normal training methods, rather than directing them how to train. However, I am only making inferences based on what’s already written in the paper. For a definite answer, you would need to contact the authors directly.
Why did trainers not ensure the owners were sufficiently schooled in using the collar so that training would remain consistent?
I don’t have enough information to answer this question.
Is it possible that the owners’ lack of confidence increased the dogs’ stress levels?
Yes, it’s possible! It’s a very interesting research question and would definitely be testable!
And why are results based on inexperienced owners’ reports instead of that of an experienced trainer?
Again, you would have to contact the researcher’s directly about this one – I don’t want to speculate about their reasoning.
Also, my apologies for the delayed reply! This popped up today when I was editing my site, but I see that it is from 2018. I am not sure why it never showed up earlier as I update the site on a regular basis!
As a professor and engineer, this study is bad and people take it as gospel. It could be horribly done, it uses base language that is not descriptive at all which does not help us understand how they reached their conclusions. It doesn’t show how the collar was actually used in practice other than “according to manufacturers directions,” they don’t take into account what elevated levels mean, nor do they show timeline and lessons learned. Of course positive reinforcement can be effective and have positive results. It’s like saying I took an aspirin and my headache went away. But how much was learned between the two training methods and how much stuck.
It also doesn’t take into account all the variations of use for the collars. I have seen professional trainers get great and positive results with e collars by not using manufacturer directions but using a lightened and more positive approach to its use so it can be weaned off after effective obedience is reached.
The other thing the study ignored, what happens when other methods are used. What are the downfalls of the other techniques. Positive reinforcement has its own issues such as instantaneous situations that call for you to hold your dog back from chasing, lunging, jumping, and any other unfavorable behaviors during training. It requires you to pull your dog from situations that could be harmful to it. What does the dog face in those moments.
What happens over time with each method. You can’t do a realistic study of an ecollar if the use being studied is of only one form. What did the dogs know in advance? What didn’t they know? Temperaments, breeds, pre mental health screenings? Age?
This study stinks of bias And the idea they can state “no need for these collars and should be banned” as a fact is downright scary for any “professional study.” Especially for so many unanswered questions regarding the use and the comparison to moments when dogs have similar reactions to even give context to what the dog may be going through. I can promise you my dog probably has more of a stress reaction from hearing a garbage truck than he would from the stim of an ecollar.
Hi Justin. I agree that there are flaws in this study, as there are in any study. No single paper can cover all of the different aspects of a particular issue. That is why replication is so, so important. The topic of punishment, in general, has been extremely well researched in a variety of species for many decades now and there is a wealth of evidence for a variety of undesirable side effects as a result of the use of aversive. Dog research is still playing catch up, but we do actually have quite a few studies that have looked at the topic of the use of punishment or aversive training methods as a whole as well as studies that have looked at the use of electronic collars in particular and the results are pretty consistent. The research generally finds no benefit of aversive methods over positive reinforcement and in many cases, there are unwanted side effects. Here is a link to a list of studies on this topic: https://www.companionanimalpsychology.com/p/resources.html?fbclid=IwAR1f6H1D2UdiJEh_QTJCNTUIobUH2hVFA1gUuLnjATLuvwMT2Wk3o90VAGQ