Collar or Harness: Which Is Better for Your Dog?

In brief

The question “collar or harness?” is often answered with great conviction, but the evidence is thinner and more nuanced than many people think. A review published in May 2026 concludes that there is little evidence on the best designs and best practices for standard dog equipment. Four things matter. First: pulling on a collar raised intraocular pressure in several studies, while pulling on a harness did not; in brachycephalic dogs, the rise was already seen while standing when the leash was held taut for ten seconds. Second: in a neck model, all the collars tested produced pressures during pulling that the authors did not consider low enough to rule out a risk of injury; the review, however, found no evidence of damage from occasional leash tension in live dogs. Third: in the published gait studies to date, every harness type tested altered at least one measured gait variable, including models described as “non-restrictive.” Fourth: in one study, dogs actually pulled harder in a harness than on a collar. The real solution for a dog that pulls is therefore training, not equipment.

This guide shows what research says about collars and harnesses, where the limits of the studies lie, and how to make a sensible choice for your dog.

We deliberately don’t cover prong collars, choke collars or tightening anti-pull harnesses in this guide. These tools are designed to inhibit behavior through targeted pulling or pressure, and in some cases through pain. We don’t consider this aversive approach sound from a learning-theory perspective and advise against using it.

1. What the newest review says

In May 2026, a review was published that brings together the known and alleged physical risks of collars, harnesses, muzzles, vests and leashes (Ridgway, 2026). Its conclusion is sobering for anyone looking for a simple answer: there is little evidence on which equipment is best designed and how best to use it.

On collars, the review notes that important structures in the neck, such as the carotid arteries, jugular veins, larynx and trachea, are only lightly protected. Case reports describe severe injuries, such as embedded collars or strangulation. At the same time, the author found no documentation in the literature that intermittent pressure from leash tension on a collar directly damages tissue. According to the review, claims of damage to the thyroid gland, trachea, jugular veins or carotid arteries from normal collar use are not supported. Whether repeated lower-level pressure has measurable long-term effects is simply unknown.

On harnesses, the review notes that in all published gait studies to date, every harness type tested altered at least one measured gait variable. It ends by calling for more studies rather than reliance on tradition and anecdote.

2. What happens at the neck

2.1 Intraocular pressure

Pauli et al. (2006) measured in 26 sled dogs how intraocular pressure changes when tension equal to each dog’s pulling force is applied via a collar or a harness. With the collar, intraocular pressure rose by 7.4 mmHg on average; with the harness, by only 2.3 mmHg, which was not statistically significant. After the tension was released, the value returned to normal within one minute. The authors recommend a harness for dogs with glaucoma, corneal damage or other conditions in which a rise in intraocular pressure could be harmful.

Bailey, Packer and Wills (2025) repeated this in 20 dogs, ten short-nosed (brachycephalic) and ten long-nosed. In the brachycephalic dogs, the collar raised intraocular pressure already while standing when the leash was held taut for ten seconds, and also during walking. In the long-nosed dogs, it rose only during walking on the collar. With the harness, there was no significant rise in either group.

👉 Brachycephaly in dogs

2.2 Pressure on the neck

Carter, McNally and Roshier (2020) tested seven collars and a slip lead on a neck model fitted with pressure sensors. With a firm pull (40 newtons), a strong pull (70 newtons) and jerking (141 newtons on average), pressures ranged from 83 to 832 kilopascals. The authors concluded that none of the collars tested produced a pressure low enough to rule out a risk of injury when the dog pulls. Important: this is a plastic model, not a living dog.

Hunter, Blake and de Godoy (2019) measured neck pressure in eight dogs walking on three collars. A double-layer collar that looks padded produced higher peak forces and contact pressures than two single-layer collars. Changes of direction, especially counterclockwise, led to higher forces than walking straight. So wider or padded doesn’t automatically mean gentler.

2.3 What this means in practice

For a dog that walks on a loose leash, a collar is not a documented health risk on current evidence. With repeated leash tension, a well-fitting harness is the more cautious option, especially for dogs with particular health vulnerabilities:

• Short-nosed dogs such as pugs or French bulldogs, in which ten seconds of taut leash tension while standing already raised intraocular pressure.
• Dogs with eye conditions such as glaucoma or corneal damage.
• Dogs with a cough or airway disease, such as tracheal collapse. This is a veterinary recommendation from practice, for which there is no comparative study.
• Dogs with problems of the neck or cervical spine, after consulting your veterinarian.

Avoid jerking the leash in any case. In the neck model, jerking produced the highest forces (Carter et al., 2020).

👉 Why is my dog coughing?

3. What a harness does to gait

Lafuente, Provis and Schmalz (2019) measured shoulder extension in nine dogs at walk and trot, without a harness, with a Y-shaped harness considered “non-restrictive,” and with a harness with a horizontal chest strap considered “restrictive.” Both tested harnesses reduced shoulder extension. In this small study, the tested Y-harness reduced extension more, by just over 9 degrees at a trot compared with just under 5 degrees for the tested horizontal chest-strap harness. That does not show that Y-harnesses in general are more restrictive.

Williams et al. (2025) studied three commercially available harnesses in 66 dogs on a loose leash. Almost all measured values differed between the harnesses, such as stride length, weight distribution between front and hind legs and shoulder angles. How much varied by breed. The authors advise choosing and fitting harnesses to the individual dog rather than recommending one design for all.

In practice, this means a harness is not automatically neutral for the musculoskeletal system. The labels “restrictive” and “non-restrictive” say little about how a particular harness works on a particular dog. Whether the measured changes have long-term health effects has not been studied.

4. Does a dog pull harder in a harness?

Bailey, Discepolo, Baker and Perry (2025) measured pulling force in 28 dogs outdoors on a flat collar or a padded back-clip harness when they encountered stimuli such as unfamiliar dogs, food or toys. In the harness, the dogs pulled with 60.5 newtons on average; on the collar, with 37.8 newtons, so considerably harder. Small dogs pulled particularly hard relative to their body weight.

Another study with 23 shelter dogs and different equipment came to a different result: there, dogs pulled harder on a martingale collar than on a front-clip harness (Johnson & Wynne, 2024). However, every dog tested the equipment in the same order, collar first and harness second, so order or habituation effects are possible. The walks also lasted only five minutes.

Above all, this shows that how hard a dog pulls depends on the equipment, the dog, the stimulus and the person holding the leash, and that no equipment replaces training. A harness shifts leash tension away from the collar and onto the chest and trunk, but it doesn’t teach a dog to walk on a loose leash.

👉 Loose-leash walking

👉 Tightening anti-pull harnesses: risks

5. Do collars or harnesses affect well-being?

Grainger, Wills and Montrose (2016) walked 30 dogs for 20 minutes each, once on a collar and once in a harness, and recorded signs of stress. For most behaviors there were no differences. Dogs used to a collar more often showed low ear positions, which the authors cautiously interpret as a possible sign of stress. Overall, they found no indication that either of the two models tested caused stress. They did not study long-term effects.

6. How to choose

In practice, based on the studies above:

• Does your dog pull regularly, is he short-nosed, or does he have eye, airway or neck problems? Then a well-fitting harness is often the more cautious choice, while you train loose-leash walking alongside.
• Does your dog reliably walk on a loose leash? Then, on current evidence, nothing speaks against a flat, well-fitting collar.
• Check the fit. A harness shouldn’t rub in the armpits, shouldn’t press on the windpipe and should let the shoulder move as freely as possible. Watch your dog at a trot to see whether he moves differently with the harness than without.
• Usually take it off at home. It generally serves no purpose there, and removal reduces rubbing and snagging risks.
• ID tag. If a collar is worn routinely for identification, it should be light, well fitted and removed where snagging is a concern; the leash can be attached to the harness.
• No jerking, whatever the equipment.

If your dog suddenly pulls differently than usual, moves unevenly or avoids certain movements, pain may also be behind it. That should be checked by a veterinarian.

👉 Recognizing pain in dogs

👉 Retractable leashes: freedom or risk?

7. A note on the evidence

The studies on collars and harnesses are almost all small: nine dogs for shoulder extension (Lafuente et al., 2019), eight for neck pressure (Hunter et al., 2019), 20 for intraocular pressure (Bailey, Packer & Wills, 2025), and 26 sled dogs in the older study on that topic (Pauli et al., 2006). The neck model (Carter et al., 2020) is a plastic cylinder, not a living dog. The study with 66 dogs (Williams et al., 2025) examined only a loose leash, not pulling. The two studies on pulling force (Bailey, Discepolo, Baker & Perry, 2025; Johnson & Wynne, 2024) used different equipment and reached different results. Johnson & Wynne also tested equipment in the same order for every dog, so order or habituation effects cannot be excluded. The study on well-being (Grainger et al., 2016) involved two walks per dog.

None of the studies examined long-term effects, either for collars or for harnesses. The 2026 review (Ridgway, 2026) is a narrative review, not a systematic analysis with predefined search criteria. The recommendations on choosing equipment in this guide are therefore marked as practical experience.

8. Conclusion

The honest answer to “collar or harness?” is: it depends on the dog. When repeated leash tension occurs, and especially in short-nosed dogs or dogs with eye, airway or neck problems, a well-fitting harness is often the more cautious option. For dogs that walk on a loose leash, there are currently no documented indications that normal use of a well-fitting flat collar causes tissue damage, although long-term data are lacking. A harness is not automatically neutral either: in the published gait studies to date, the harness types tested altered at least one measured gait variable and need to be fitted to the individual dog.

The three things that make the difference:

Choose by the dog. Pulling behavior, head shape and health decide, not marketing.

Pay attention to fit. A poorly fitting harness is not a good alternative to a collar.

Training, not equipment. No equipment teaches a dog to walk on a loose leash.

Key takeaways on collars and harnesses

Little solid evidence. A 2026 review found little evidence on the best design and use of standard equipment (Ridgway, 2026).

Collar tension raises intraocular pressure. With tension on a collar, it rose by 7.4 mmHg on average; on a harness, not significantly (Pauli et al., 2006).

Brachycephalic dogs responded even while standing. In brachycephalic dogs, intraocular pressure rose while the leash was held taut for ten seconds in the standing condition (Bailey, Packer & Wills, 2025).

High pressure in the neck model. All the collars tested produced pressures between 83 and 832 kilopascals during pulling (Carter et al., 2020).

Padded isn’t automatically gentler. A double-layer collar produced higher peak pressures than single-layer ones (Hunter et al., 2019).

All harness types tested so far have affected at least one gait variable. In one small study, the tested Y-harness reduced shoulder extension more than the tested horizontal chest-strap harness; this does not establish a general ranking of harness designs (Lafuente et al., 2019; Williams et al., 2025).

Some dogs pulled harder in a harness. In one study, dogs pulled with 60.5 newtons in a harness and 37.8 newtons on a collar (Bailey, Discepolo, Baker & Perry, 2025).

References

• Bailey, J., Discepolo, D., Baker, J., & Perry, E. (2025). An investigation of force potential against the companion dog neck associated with collar use. Journal of Veterinary Behavior, 78, 19–24. https://doi.org/10.1016/j.jveb.2024.10.007
• Bailey, M. E., Packer, M. J., & Wills, A. P. (2025). Effect of a collar and harness on intraocular pressure and respiration rate of brachycephalic and dolichocephalic dogs. Veterinary Medicine and Science, 11(3), Article e70384. https://doi.org/10.1002/vms3.70384
• Carter, A., McNally, D., & Roshier, A. (2020). Canine collars: An investigation of collar type and the forces applied to a simulated neck model. Veterinary Record, 187(7), e52. https://doi.org/10.1136/vr.105681
• Grainger, J., Wills, A. P., & Montrose, V. T. (2016). The behavioral effects of walking on a collar and harness in domestic dogs (Canis familiaris). Journal of Veterinary Behavior, 14, 60–64. https://doi.org/10.1016/j.jveb.2016.06.002
• Hunter, A., Blake, S., & de Godoy, R. F. (2019). Pressure and force on the canine neck when exercised using a collar and leash. Veterinary and Animal Science, 8, Article 100082. https://doi.org/10.1016/j.vas.2019.100082
• Johnson, C. M., & Wynne, C. D. L. (2024). Comparing efficacy in reducing pulling and welfare impacts of four types of leash walking equipment. PeerJ, 12, Article e18131. https://doi.org/10.7717/peerj.18131
• Lafuente, M. P., Provis, L., & Schmalz, E. A. (2019). Effects of restrictive and non-restrictive harnesses on shoulder extension in dogs at walk and trot. Veterinary Record, 184(2). https://doi.org/10.1136/vr.104946
• Pauli, A. M., Bentley, E., Diehl, K. A., & Miller, P. E. (2006). Effects of the application of neck pressure by a collar or harness on intraocular pressure in dogs. Journal of the American Animal Hospital Association, 42(3), 207–211. https://doi.org/10.5326/0420207
• Ridgway, M. (2026). Health implications of dog-worn equipment: A review of known and alleged physical risks. Frontiers in Veterinary Science, 13, Article 1711781. https://doi.org/10.3389/fvets.2026.1711781
• Williams, E., Hunton, V., Boyd, J., & Carter, A. (2025). Effect of harness design on the biomechanics of domestic dogs (Canis lupus familiaris). Journal of Applied Animal Welfare Science, 28(2), 301–317. https://doi.org/10.1080/10888705.2023.2259796

Note: This guide does not replace individual advice. If your dog coughs, has eye problems or moves unevenly, have him examined by a veterinarian before changing equipment.

Frequently asked questions about collars and harnesses

Choice, health and pulling