Do You Really Need a Shock Collar? What the Science Says About E-Collars and Dog Training

If you have a dog who bolts after squirrels, ignores you outdoors, lunges at other dogs, or appears to forget every word of English the moment you leave the apartment, a shock collar can sound tempting.

The pitch is usually straightforward: Your dog already knows what to do. You just need a way to make them listen.

And to be fair, electronic collars can change behavior. That is not really the scientific question.

The better questions are: Do they work better than reward-based training? Are they necessary? And what are the potential costs to the dog?

Those questions have actually been studied.

And the results are more interesting than either side of the dog-training culture war sometimes makes them sound.

First: What Do We Mean by a Shock Collar?

The terminology gets messy.

You may hear shock collar, electronic collar, e-collar, or remote training collar. Modern devices can include tone, vibration, and remotely delivered electrical stimulation at adjustable intensities.

This article is specifically concerned with the use of electrical stimulation as an aversive consequence or source of pressure during training.

That distinction matters. A vibration cue conditioned as a neutral signal, for example, is not automatically equivalent to electrical stimulation—although vibration itself can still be aversive to some dogs.

In behavioral terms, what matters is not what the manufacturer calls the sensation. What matters is how the dog experiences it and how it functions in learning.

Do Shock Collars Actually Work?

Yes—in the limited sense that an aversive consequence can change behavior.

That should not be controversial. Punishment and negative reinforcement are real learning processes.

If pressing a button causes an unpleasant sensation and the sensation stops when the dog performs a behavior, the dog may learn how to avoid that sensation.

The more useful question is whether adding electrical stimulation produces better training than we can achieve without it.

One of the most informative studies addressing that question was published by Lucy China, Daniel Mills, and Jonathan Cooper in Frontiers in Veterinary Science in 2020.

The researchers studied 63 dogs with off-leash behavioral problems, including poor recall. The dogs were divided into three groups of 21 dogs each:

• One group was trained by experienced, manufacturer-nominated e-collar trainers using electronic stimulation.

• A second group was trained by the same trainers without electronic stimulation.

• A third group was trained by independent professional trainers who focused primarily on positive reinforcement.

The dogs received up to 150 minutes of training over five days.

If shock collars provided a substantial training advantage, this was a pretty reasonable opportunity for that advantage to appear.

It didn't.

The Reward-Based Dogs Were More Responsive to the First Cue

Researchers examined responses to two common behaviors: Come and Sit.

For recall, dogs in the reward-focused group responded to the first command about 82.5% of the time, compared with 71.0% in the e-collar group. The difference was statistically significant (p = .015).

For Sit, the reward-focused dogs responded to the first command about 83.5% of the time, compared with 76.8% in the e-collar group (p = .036).

They also required fewer repeated commands.

Response speed showed a similar pattern.

For Come, the mean response latency was approximately 1.13 seconds in the reward-focused group versus 1.35 seconds in the e-collar group.

For Sit, it was approximately 1.36 seconds versus 1.67 seconds, respectively.

These are not gigantic real-world differences. Nobody is going to watch their dog recall 0.22 seconds faster and start popping champagne.

But that isn't the important point.

The important point is that electrical stimulation did not create the performance advantage it is often claimed to provide.

If anything, several measures favored the dogs trained primarily with positive reinforcement.

An Important Detail: The E-Collar Dogs Didn't Fail Dramatically

There is another result worth discussing because science gets less useful when we cherry-pick only the numbers we like.

There was no statistically significant difference between the groups in the percentage of outright failures to respond.

For Come, failure rates were approximately 2–3% across groups. For Sit, they were roughly 4–5%.

So the study does not show that e-collars make dogs incapable of learning or that every dog trained with one falls apart.

All three groups improved.

The more defensible interpretation is narrower:

Adding an electronic collar did not improve training effectiveness, while reward-focused training performed as well as or better on several important measures.

That is a much stronger argument than pretending the other group learned nothing.

What About Welfare?

Effectiveness is only half the equation.

Suppose two treatments produce a similar result, but one carries a greater risk of fear, stress, or unwanted associations. There would need to be a fairly compelling benefit before accepting that additional cost.

Earlier work by Cooper and colleagues investigated both effectiveness and welfare in dogs trained with remote electronic collars.

Again, the study involved 63 pet dogs divided into three groups of 21.

Dogs trained using e-collars showed some behavioral differences associated with poorer welfare. Compared with the reward-focused group, the e-collar dogs spent more time tense, yawned more, and interacted less with their environment.

Yet owners reported improvement across all three groups, with 92% reporting improvement in the behavior for which their dog had been referred.

There was no significant difference in owner-reported effectiveness between groups.

In other words: the dogs could improve without the shock.

That is an important distinction.

A Larger Study Looked Beyond Shock Collars

Electronic collars are only one form of aversive training.

In 2020, Vieira de Castro and colleagues studied 92 companion dogs from seven training schools.

The dogs were divided according to their schools' training practices:

42 dogs came from reward-based schools, 22 from schools using a mixture of reward and aversive methods, and 28 from schools using a higher proportion of aversive methods.

Researchers measured behavior during training, salivary cortisol, and—in 79 dogs who completed the later phase—a cognitive-bias task designed to examine whether dogs responded more optimistically or pessimistically to ambiguous situations.

Dogs in the more aversive training group displayed more stress-related behavior during training, were more frequently tense, showed greater post-training increases in cortisol, and demonstrated a more pessimistic cognitive bias than dogs from reward-based schools.

That does not mean every correction permanently traumatizes a dog.

It means that, at the group level, greater reliance on aversive training was associated with measurable welfare costs.

That's a much more precise—and scientifically useful—statement.

But Aren't These Studies Small?

Relatively, yes.

The direct e-collar research is nowhere near the scale of a human pharmaceutical trial with thousands of participants.

The 63-dog study had only 21 dogs per condition. The 92-dog welfare study was larger, but dogs were recruited from existing training schools rather than randomly assigned to a lifetime of reward-based or aversive training.

There are also trainer effects, differences in technique, individual dog differences, and difficulties defining broad categories such as “reward-based” and “aversive.”

The China study also deserves a methodological footnote: the two groups involving the e-collar trainers were randomized, while the reward-focused group was recruited later and matched to the earlier dogs rather than being part of the same fully randomized three-arm experiment.

Those limitations matter.

But they do not somehow reverse the findings.

A reasonable scientific conclusion is not that reward-based training has been proven superior in every dog, every behavior, and every imaginable circumstance.

It is this:

There is currently no convincing evidence that electronic collars provide an effectiveness advantage that justifies their additional welfare risks.

Why Shock Collars Can Look So Effective

This is where behavioral neuroscience becomes important.

Imagine a reactive dog sees another dog across the street.

He stiffens.

He stares.

He begins to bark.

An electrical correction occurs.

The barking stops.

From the human perspective, the problem appears fixed almost instantly.

But observable behavior is only one part of what the nervous system is doing.

The dog may have learned:

Don't bark.

But he may also have learned:

When dogs appear, something unpleasant happens.

That distinction becomes especially important when we are working with fear, anxiety, frustration, or aggression.

Suppressing a growl is not the same thing as making the dog feel safe.

Stopping a lunge is not necessarily the same thing as changing the emotional response that produced it.

And removing warning behaviors without changing the underlying emotional state can sometimes make a dog's behavior harder—not easier—to interpret.

What Does Reward-Based Training Do Differently?

Reward-based training does not mean standing around feeding chicken while your dog runs the household like a tiny furry dictator.

Good reward-based training involves structure, consequences, management, repetition, and very clear criteria.

The difference is that we deliberately teach the dog what behavior works instead.

A dog who lunges at another dog might learn to notice the dog, disengage, orient toward the handler, make a U-turn, move behind a visual barrier, or calmly continue walking.

A dog who jumps on visitors can learn that keeping four feet on the floor produces attention.

A dog with poor recall learns that turning toward the owner and returning quickly has a long reinforcement history.

For fear and reactivity, we can go even further through desensitization and counterconditioning, gradually changing what the trigger predicts.

Instead of:

Dog appears → something unpleasant happens

we build:

Dog appears → safety, distance, food, predictability, and successful behavior follow

That is not merely suppressing an output.

It is changing the learning underneath it.

What Major Veterinary Behavior Experts Recommend

The American Veterinary Society of Animal Behavior's current Humane Dog Training position recommends reward-based methods for canine training and behavior modification and advises against aversive methods including electronic collars.

That position is not based on an argument that punishment is biologically incapable of changing behavior.

Of course punishment changes behavior.

The issue is whether we need to accept its additional welfare risks when effective alternatives are available.

At present, the evidence says we generally do not.

So, Should You Use a Shock Collar?

Based on the research we currently have, I don't recommend it.

Not because every dog exposed to an electronic collar will be psychologically damaged.

Not because an electronic collar can never suppress unwanted behavior.

And not because positive reinforcement is some magical technique in which dogs never experience frustration or limits.

I don't recommend shock collars because the burden of proof should fall on the more invasive intervention.

If applying an aversive stimulus to a dog's neck were substantially more effective—particularly for dangerous behaviors—there would at least be a scientific argument for weighing that benefit against the welfare cost.

But controlled research has not demonstrated that advantage.

Reward-based training can produce excellent reliability while allowing us to teach the behavior we actually want and, when emotions such as fear are involved, address the problem underneath the behavior rather than only the visible symptom.

When the Behavior Is Serious

If you're considering an e-collar because your dog is lunging, growling, snapping, biting, guarding resources, or becoming increasingly difficult to control, that is exactly when I would be cautious about reaching for stronger punishment.

Those behaviors deserve an assessment of why they are occurring.

Pain, fear, frustration, chronic arousal, environmental stress, learned behavior, resource guarding, and underlying medical conditions can all contribute.

For dogs who bark and lunge at people or dogs, reactive dog training in NYC can focus on threshold work, management, counterconditioning, disengagement, and real-world handling skills.

For dogs with a history of growling, snapping, biting, or serious resource guarding, aggressive dog training in NYC should begin with safety, behavioral assessment, and a structured modification plan rather than simply escalating corrections.

The goal should not be to win an argument with your dog.

The goal is to change what the dog has learned—and give them a safer, more useful answer.

The Bottom Line

Shock collars can affect behavior.

But “can affect behavior” and “is the best way to train” are two very different claims.

The best controlled evidence we currently have has not found an efficacy advantage for electronic-collar training. In several measures, dogs trained primarily through positive reinforcement actually performed better.

Separate research has associated greater use of aversive methods with increased stress behavior, elevated cortisol responses, and more pessimistic cognitive bias.

The science is not enormous. It is not flawless. And there is still plenty we should study.

But we do not need perfect evidence to make a rational decision when one approach has no demonstrated performance advantage and carries additional potential costs.

If we can teach the same behavior—or teach it better—without deliberately causing discomfort, that's a pretty sensible place to start.

And your dog gets to keep wondering why sitting politely somehow keeps making chicken appear.

Seems like a fair deal.

References

China, L., Mills, D. S., & Cooper, J. J. (2020). Efficacy of dog training with and without remote electronic collars vs. a focus on positive reinforcement. Frontiers in Veterinary Science, 7, 508. DOI: 10.3389/fvets.2020.00508.

Cooper, J. J., Cracknell, N., Hardiman, J., Wright, H., & Mills, D. (2014). The welfare consequences and efficacy of training pet dogs with remote electronic training collars in comparison to reward based training. PLOS ONE, 9(9), e102722. DOI: 10.1371/journal.pone.0102722.

Vieira de Castro, A. C., Fuchs, D., Morello, G. M., Pastur, S., de Sousa, L., & Olsson, I. A. S. (2020). Does training method matter? Evidence for the negative impact of aversive-based methods on companion dog welfare. PLOS ONE, 15(12), e0225023. DOI: 10.1371/journal.pone.0225023.

Ziv, G. (2017). The effects of using aversive training methods in dogs—A review. Journal of Veterinary Behavior, 19, 50–60. DOI: 10.1016/j.jveb.2017.02.004.

American Veterinary Society of Animal Behavior. (2021). Position Statement on Humane Dog Training. Current position reaffirmed by AVSAB in 2025.

About the author: Pepe J. Hernandez, Ph.D., CPDT-KA, is a behavioral neuroscientist and professional dog trainer and the founder of PJH Dog Training in New York City. His work focuses on neuroscience-informed, reward-based training for puppies, companion dogs, and dogs with reactivity, fear, and aggression.