Beyond Treats and Praise: The Harmful Results of All-Positive Dog Training
When our puppy, Skye, was about three months old we sat down on the living room sofa with a tantalizing plate of bacon and eggs and placed it on the coffee table - eye level for a golden retriever. Having never learned the rule, we suspected she would go for it, and of course she did. Then, at the very moment that Skye rested her chin on the edge of the table and reached out her tongue for a curious lick, Ovidiu gently tapped her nose while saying “phooey” in low, firm, displeased voice. Having spent the last month learning how to communicate clearly with each other, Skye immediately understood that we were establishing a rule. She backed off, relaxed, and played with a toy while we chatted, and that was the last time she tried to eat food off the table.
According to many leading voices in animal welfare and training, we had just committed an act of aversive training, indistinguishable from an entire class of abuse that includes kicking, hitting, and hanging from a choke collar. Let’s back up a bit.
A small but growing body of studies have used an operant conditioning framework to examine the science of dog training. Operant conditioning1 is a learning process most commonly associated with behaviorist B.F. Skinner where rewards and aversive stimuli are used to modify voluntary behaviors. There are four main techniques: positive reinforcement (reward for doing something right), positive punishment (an unpleasant consequence for doing something wrong), negative reinforcement (the withdrawal of an unpleasant stimulus for doing something right – for example the release of tension from the collar), and negative punishment (take away something good, like a toy, when a dog does something unwanted).
In the context of dog training, these four actions are usually grouped into two categories: positive reinforcement and everything else. The everything else are called “aversive techniques,” and the few studies that have been done – the most frequently cited is a 2019 study2 of 92 dogs from either aversive, mixed, or positive dog training schools in Portugal - show that dogs trained with aversive techniques are more stressed and less successful at learning new things during testing than dogs trained only with positive reinforcement. Imperfect though the studies are, they have confirmed our anecdotal experience: positive reinforcement is the best tool for teaching a dog to do something, and developing a trusting relationship between dog and handler is key to success in training. But if we stop here, the simple and natural act of communicating displeasure to a dog is categorized as aversive. Setting boundaries with voice, body language, or a mild tug of the collar become actions that will lead to abuse and must be categorically frowned upon.
One might now wonder why we shouldn’t just stop here. Why complicate things instead of keeping it all positive? For starters, because it isn’t working. As owners of an NYC based dog training and services company for over twenty years, we have a pretty good sense of overall behavioral trends in the urban dogs we work with. We are seeing more and more pups who spend their days in crates because they haven’t been taught the rules of the house, who aren’t allowed off leash because they can’t be trusted to stay nearby, and who are isolated from new experiences because they have never learned to make good choices. These dogs require extensive control and supervision, and their surroundings must be fully managed and curated.
As anyone with a child knows (and the science here is extensive), safely establishing rules and boundaries with an authoritative parenting style is the best way to foster secure attachment, emotional regulation, and curiosity in learning. For some perspective, we know that kids learn better when given encouragement (positive reinforcement) and that beating (positive punishment/aversive) or neglecting (negative reinforcement/aversive) a child is not only morally wrong but leads to bad outcomes. Imagine if we interpreted this information to suggest that parents should not set rules because, by definition, rules require a consequence to be enforceable. Not only is this a logical leap, but it is also contrary to parenting studies science, and as it turns out, to studies of dog training as well. Scientists at the Human-Animal Interaction Laboratory at Oregon State University recently published a report3 in the journal of Animal Cognition confirming that dogs learn better, are more persistent, and are more resistant to stress when their parents adopt an authoritative style, setting high expectations of their dogs.
So how can this study be reconciled with the results of studies that found aversive training to be less effective? Does the cult of all-positive dog training preclude rule-setting because we are afraid of our own ability to walk the line between abuse and clear communication? Does this fear warrant that we sacrifice boundaries all together? And at what cost? An entire generation of dogs are restricted and restrained from enjoying fundamental life experiences because they have never been taught the rules.
Carving out space and language for clear communication allows us to see that rule-setting need not be linked to abusive consequences. From this perspective, an authoritative parenting style and a system of positive reinforcement can co-exist. In reality, many dog trainers already take this approach even when they claim to be using positive reinforcement only. If you watch closely and ask the right questions, those with the best results use their firm voice and leash to get a dog’s attention and communicate that the behavior is unacceptable. Parents, however, aren’t getting this message.
It might sound obvious, but when the loudest voices in pet training tell parents to use an all-positive approach, they listen. Dogs are kept on a tight leash, both literally and metaphorically, because their parents have no other way to keep them safe and prevent bad behavior. They wear uncomfortable harnesses that inhibit their movement instead of being taught to walk on a leash without pulling. Exploring the park off-leash is totally unimaginable. In the home, it is now “common knowledge” that puppies should be crate-trained to keep them and the home safe instead of establishing a few rules from the get-go and teaching them to make good choices. If we step back and look at the big picture, this all-positive approach starts to look less warm and fuzzy.
Skye is an integral part of our family, and she loves going on adventures with her two human sisters, ages six and eight. All three of our “kids” are given the freedom to explore on our hikes, go for walks in town, and have human and canine friends over every day because the expectations for safe behavior and good listening are clear. Tonight, we are going to get ice cream, and Skye will come along to have a pup cup. She’ll have a lightweight leash attached to her collar, but we won’t hold onto it. She’ll hang out with us and enjoy family time. She won’t run into the road next to the creamery (neither will the kids), and she won’t go after the other littles’ cones. Everyone knows the rules, and we all enjoy a richer life because of them.
https://en.wikipedia.org/wiki/Operant_conditioning
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0225023#pone.0225023.ref005
Trainers and commentators point to the study in Portugal to draw simple and sweeping conclusions where it should merely be the starting point for asking a new set of questions. What level of stress in learning is productive and healthy, and what level becomes harmful? Is the use of crates, gentle-leaders and no-pull harnesses correctly categorized as aversive and what are their impacts on a dog’s well-being? How effective is the all-positive training method at extinguishing unwanted behaviors?
https://link.springer.com/article/10.1007/s10071-022-01694-6
These findings suggest that the quality and style of individual dog-human relationships, including Pet Parenting Style, may be relevant when evaluating and interpreting outcomes of canine cognition research.