Humans have been walking dogs since the 19th century, and in many places, a guardian’s ability to control their dog in public is considered a critical component of ownership. However, previous research has shown that leash-pulling is a common behavioral concern for guardians. When people don’t enjoy the experience, this can influence walk frequency, length, and quality to the detriment of their dog’s development.

There are numerous devices available meant to help curb leash-pulling behavior. Some devices, like prong and choke collars, accomplish this by inflicting pain on purpose. Research has repeatedly demonstrated that aversive equipment negatively affects animal welfare and inhibits training. However, much less is known about other devices generally viewed as non-aversive, such as flat collars, harnesses, and head collars. Thus, this review examined 21 articles and two abstracts on these other devices to help guardians choose the most suitable walking equipment for their dogs.

Articles were selected through Google Scholar keyword searches for “dog harness,” “dog collar,” “dog head collar,” “dog restraint methods,” “dog walking,” and “dog collar injury” in February and March 2025. Restraint devices covered by the review included:

• Collars: Also called flat or neck collars, these consist of a strip of fabric that fits around a dog’s neck.
• Martingale collars: These are neck collars with a loop, usually made of fabric or chain, that can tighten up to a point and help prevent a dog from slipping out. Because of this, they’re sometimes called limited-slip collars.
• Head collars: Also called head halters, these fit over a dog’s muzzle and around their neck. Some tighten when pulled.
• Harnesses: These fit around a dog’s body, chest, and shoulders. They come in a variety of designs depending on their shape (e.g., H-shaped or Y-shaped) and where the leash attaches (e.g., front-clip or back-clip). Some tighten under pressure.

Collars

Studies compared collar effectiveness mainly by measuring pressure and force. Researchers found that all collars tested are capable of causing injuries, particularly if they have a smaller surface area and apply pressure in the middle of the neck. Even the cushioning in a padded collar can unexpectedly concentrate pressure and increase injury risk.

Research also showed that neck pressure is higher when walking in circles and leash tension increases on a complex course, suggesting that collars may be an issue in situations where tighter control is needed, such as on a busy sidewalk.

Harnesses

For harnesses, studies assessed movement patterns and gait changes using motion-capture systems, video recordings, and pressure-sensing mats. While all harnesses altered dogs’ gaits, no significant differences were observed to indicate an optimal harness. Researchers concluded that the individual fit of the harness and the context of the walk are key factors.

Interestingly, although the chest-strap harness is often assumed to be restrictive and the Y-shaped harness non-restrictive, one study found that the latter limited shoulder extension more than the former. Similarly, another study found that the Y-shaped harness with a front clip limited shoulder and elbow movement more than the chest-strap and H-shaped designs.

Collars Versus Harnesses

Some studies compared collars and harnesses using a variety of measures, such as intraocular pressure and behavioral signs of stress. These included lip-licking, yawning, trembling, ear and tail position, body posture, and vocalizations.

One study found that intraocular pressure was higher when dogs pulled while wearing a collar compared to a harness. Brachycephalic (flat-faced) dogs also exhibited elevated intraocular pressure in collars but not harnesses in another study.

Several studies showed slight differences in stress-related behaviors between the two types of devices, but the findings were mixed and difficult to interpret. For instance, nose-licking was observed more often in harnessed dogs when encountering an unfamiliar dog, whereas paw-lifting was seen more often in collared dogs. In another study, dogs pulled back their ears more if they had a history of wearing a collar.

There were similarly mixed results for pulling behavior. Some researchers found that dogs pulled more when wearing a chest-strap or Y-shaped harness compared to a flat collar, while others reported more pulling with a martingale collar compared to a front-clip harness. This suggests that pulling behavior was reduced when dogs found the device in question uncomfortable — whether a collar or a harness.

Overall, though, most studies didn’t report conclusive differences between collars and harnesses in terms of their impact on dog welfare.

Head Collars

Studies investigating head collars looked at physiological and behavioral responses, such as heart rate, cortisol levels, nose pawing, and face rubbing. One found that while head collars effectively controlled behavior, dogs also crouched, lowered their heads and ears, and pawed at their noses significantly more than when wearing a flat collar. Another study compared different head collar designs and reported no advantage of one type over another. Both studies found that dogs seemed to habituate to the head collars somewhat over time, as a few of their physiological and behavioral measures of stress decreased.

So Which Device Is Best?

The review concludes that flat collars shouldn’t be used on brachycephalic dogs due to health risks, or dogs with conditions like thin corneas or glaucoma, where increased intraocular pressure is a concern.

For dogs who pull, reward-based training and non-tightening front-clip harnesses are recommended. Chest-strap and Y-shaped harnesses with back clips are preferred for dogs who aren’t pullers, while flat collars are only appropriate for dogs who consistently walk on a loose leash. Future research on factors such as dog age, training strategies, and specific device features like padding, strap width, and material can inform overall welfare improvements.

Guardians can consider the benefits and drawbacks of each device presented in this review and make a decision based on their dog’s anatomy, what they want the device to accomplish, and what type of walks they go on. It’s important for guardians to recognize that even devices considered non-aversive discourage pulling because they cause dogs some level of discomfort. Ultimately, the best restraint device depends on the individual dog and their guardian.