A look at the evidence

Dog training has never been more accessible, and the results have never been more inconsistent. There are more trainers, more methods, more YouTube channels, and more conflicting advice than at any point in history. In this saturated landscape, one of the most robust and consistent bodies of scientific research in applied animal behaviour is frequently ignored simply out of habit, tradition, and an attachment to ideas that simply do not hold up to scrutiny.

The science of how dogs learn is not new, but it’s definitely worth a refresher.

The Dominance Theory problem

Let us begin with the most persistent myth in dog training: the idea that dogs are constantly attempting to assert dominance over humans and each other, and that the solution is for the owner to establish themselves as the ‘alpha.’

This framework, rooted in wolf studies conducted in the mid-20th century on captive, unrelated wolves forced into artificial groups, was debunked by the researcher who originally promoted it, Dr L. David Mech, decades ago. Mech has spent years actively trying to undo the damage his early work caused. Natural wolf packs, he clarified, are family units led by parents, not aggressive hierarchies maintained through constant challenge and submission.

More importantly, dogs are not wolves. They diverged from wolves thousands of years ago, and domestication has profoundly altered their social cognition, their relationship with humans, and the way they navigate the world. Applying outdated wolf hierarchy theory to the dog-human relationship actively leads to training approaches that damage the bond between dog and owner and, in many cases, produce worse behavioural outcomes.

Yet alpha rolls, dominance-based corrections, and the idea that a dog walking ahead of its owner on a lead is ‘trying to be the boss’ remain deeply embedded in popular dog training culture. We need to let them go.

What the research actually shows

The evidence on training methods is extensive and consistent. A 2020 study published in PLOS One compared dogs trained in reward-based, mixed-method, and aversive-based schools. Dogs from aversive training environments showed significantly more stress indicators, like panting, yawning and lip-licking, and had measurably higher cortisol levels after training sessions. A study in Scientific Reports found that dogs trained with two or more aversive techniques were slower and less willing to engage with ambiguous tasks – a well-validated indicator of a negative emotional state.

A 2025 study found that dogs whose owners were willing to use aversive training methods were more likely to display reactive and aggressive behaviours overall. Multiple reviews of the literature have reached the same conclusion: reward-based training produces dogs that are more obedient, less fearful, less aggressive, and more emotionally resilient than those trained with punishment or intimidation. Critically, there is no evidence that aversive methods are more effective than reward-based approaches, and considerable evidence that they are less so, while carrying significant welfare costs.

The American Veterinary Society of Animal Behavior, the Royal Society for the Prevention of Cruelty to Animals, the Association of Professional Dog Trainers, and virtually every major veterinary and animal welfare organisation in the world now endorses reward-based training and recommends against the use of aversive methods. It’s certainly not ‘fringe’ anymore to insist on positive reinforcement methods; it’s following scientific consensus.

The ‘balance’ argument

A common response to the evidence above is to advocate for ‘balanced’ training, the use of both rewards and corrections, supposedly giving the best of both worlds. The argument sounds intuitive, but in practice, the research does not support it.

Studies consistently show that mixing reward and punishment confuses dogs, increases stress, and produces more problem behaviours than reward-only approaches. One study found that dogs whose owners used both rewards and punishment showed higher rates of aggression than dogs trained with either method alone, meaning that the supposedly balanced approach produced the worst outcomes

The training method gap

Perhaps the most sobering finding in the research landscape is this: despite the scientific consensus, surveys consistently show that only 16 to 20% of dog owners use exclusively positive reinforcement. The majority continue to use punishment in some form, with roughly half using it more often than rewards.

There is clearly a communication problem between the scientific community and the dog-owning public, and that gap is being actively filled by trainers, social media personalities, and television formats that prioritise dramatic transformations and mythologised authority over the slower, quieter, genuinely effective work of positive training.

What good training actually looks like

Reward-based training is often mischaracterised as permissive, unstructured, or suitable only for ‘easy dogs’. This is false. Force-free training is strategic, systematic, and effective for dogs with significant behavioural challenges, including reactivity, aggression rooted in fear, and complex anxiety disorders.

Good training is built on a clear understanding of how animals learn. It uses high-value reinforcers to build behaviours the dog is motivated to repeat. It sets dogs up to succeed rather than waiting for failure to punish. It reads the dog’s emotional state and adjusts accordingly. It is patient, consistent, and genuinely attentive to what the dog is experiencing – not just what the owner wants to see.

This is not soft. It is skilled and it produces results that last, because they are built on understanding and trust rather than suppressed fear.

A note on regulation

Dog training in most countries is almost entirely unregulated. Anyone can call themselves a dog trainer or behaviourist without any qualifications, any scientific knowledge, or any accountability. This is a serious welfare problem, and it is one reason why harmful methods persist in the marketplace.

When choosing a trainer, look for professional qualifications from respected bodies such as the Association of Pet Dog Trainers (APDT), the Institute of Modern Dog Trainers (IMDT), or the Animal Behaviour and Training Council (ABTC). Ask the prospective trainer explicitly whether they use punishment-based techniques and if the answer is yes, or unclear, walk away.

The bottom line

The science is clear, but the gap remains in application. Closing that gap, between what the research tells us about how dogs learn and what is actually happening in training classes, homes, and online tutorials worldwide, is one of the most important things the dog behaviour community can do for canine welfare.

Your dog is not trying to dominate you. They are trying to understand you. The question is whether the methods we use help them do that, or make it harder.

Sources

This article reflects the scientific consensus on canine learning and training methods, as established through peer-reviewed research in applied animal behaviour science. Key studies referenced include work published in PLOS One (2020), Scientific Reports (2021), and Applied Animal Behaviour Science (2025).