Dog Life Balance

Would you raise a child and never teach them "no"?

Asked that way, almost nobody says yes. A "no" is a boundary, not a cruelty. And yet positive dog training has almost no "no". It keeps one: "then you won't get it". The treat disappears, the game ends. But a dog can't hold that kind of link in its head across time, so that "no" almost never reaches it. This page doesn't dismiss the positive approach. On the whole it agrees with it, and with the logic behind it, right up to the point where the logic… misses.

Common ground1 / 8

Positive or balanced, we agree on most of it

Positive reinforcement works, and it works very well.

When reward-based and mixed methods were compared on teaching a new task, the reward-based group did at least as well. Most of all training, on either side, is exactly this.

An unpleasant stimulus, a correction, produces stress in the dog.

Dogs from aversive-based schools showed more stress behaviours in class, higher cortisol after sessions and a more pessimistic bias on a judgement task. Nobody on this page disputes it.

A dog with the freedom to engage the world is a happy dog.

Sniffing, running, meeting, choosing. The more of the world a dog can safely have, the better its life. Both sides are training towards this.

It's best when our lives and our dogs' lives fit together well, instead of being run by outside constraints.

Not run by gates, leashes and timetables set by what the dog can't be trusted with. Both sides want the equipment to come off eventually.

In inexperienced hands, tools do harm.

In a French owner survey, e-collar results fell well short of what trainers achieve in studies, and 7% of the dogs had burns, most from automatic bark collars.

Where we disagree2 / 8

Where we disagree

Does a dog need to know what "no" means?

The positive approach says no. Teach what to do, pay it, and the unwanted behaviour fades for lack of payment. Its "no" is a withdrawal: the treat goes away, the game ends, the door stays shut. On paper that is a consequence. But it says "then you won't get it", and a dog can't hold "because of that, this" across time. For the withdrawal to teach anything it has to land in the same second as the choice, every time. This page thinks a dog needs a "no" that arrives in a form it can use.

Is an unpleasant stimulus a legitimate way to communicate with a dog?

By correction this page means a brief, unpleasant signal that tells the dog "not that": a sound, a short leash tug, a body block, pressure that ends the moment the dog stops. Not pain. Some tools, at some settings, are more than that, and that is a separate question about intensity. The positive approach says the unpleasant channel is never legitimate, because it costs the dog stress and the same result is reachable without it. The balanced position says it is a channel dogs use with each other every day, and closing it leaves the dog with less information, not more.

Before answering either question, look at what the research can and cannot say.

The studies3 / 8

The studies answer a question we're not asking

The studies measure the moment, not the life.

The research on training methods is real, and the best of it favours the positive approach. It measures stress markers during a session and for a few weeks after: behaviours, cortisol, a judgement task. None of it measures the dog at three years old, off leash, in an emergency, or how much of its life is still being managed. And none of it separates the skill of the trainer from the tool in their hand. The studies measure the moment, not the life.

2020

Vieira de Castro et al., PLOS ONE

What it measured
Stress signs and cortisol in 92 dogs, in class and a few weeks later.
What it did not measure
Whether that stress did any harm, or what the dogs could do a year on.
Why it doesn't decide the question
It shows an unpleasant signal is stressful, which nobody disputes.
Open the study
2014

Cooper et al., PLOS ONE

What it measured
Sixty-three dogs with recall or chasing problems, trained with or without an e-collar, over a few weeks.
What it did not measure
Whether the recall held months later, off leash, with prey in front of the dog.
Why it doesn't decide the question
All three groups improved, and the same trainers with the collar off had lower cortisol than the reward group: it measured the trainer, not the tool.
2021

Vieira de Castro et al., PLOS ONE

What it measured
How fast dogs from reward-based and mixed schools learned a new task in a room.
What it did not measure
Anything about stopping a behaviour, or whether it held in daily life.
Why it doesn't decide the question
It confirms that rewards teach well, which this page agrees with.
Open the study
2012

Blackwell et al., BMC Veterinary Research

What it measured
What owners in England said had worked for recall and chasing.
What it did not measure
Any dog. Nobody watched a dog or a trainer.
Why it doesn't decide the question
It records what owners believe, not what their dogs can do.
2018

Masson et al., Journal of Veterinary Behavior

What it measured
Burns from electronic collars reported by owners in France: 7% of dogs, mostly from automatic bark collars.
What it did not measure
A correction given by a person, timed to a behaviour.
Why it doesn't decide the question
It measures a machine firing at a dog, which this page opposes as much as you do.
2021

AVSAB, Position Statement on Humane Dog Training

What it measured
Nothing. It is a position statement recommending reward-only training.
What it did not measure
It ran no experiment; it weighs the studies above.
Why it doesn't decide the question
Policy made in good faith, resting on the same narrow studies.
Open the study
2007

Schalke et al., Applied Animal Behaviour Science

What it measured
Stress in lab beagles when an e-collar signal was predictable and controllable, versus random.
What it did not measure
Pet dogs in daily life, over a long time.
Why it doesn't decide the question
The closest to our question: when the dog could predict and end it, there was no lasting stress.
2013

Dale et al., Applied Animal Behaviour Science

What it measured
120 hunting dogs, one aversion session around kiwi, checked a year later.
What it did not measure
Everyday pet training or obedience.
Why it doesn't decide the question
One well-timed unpleasant event, and 87% of the dogs still avoided kiwi a year on.

What every one of these studies shares

Small samples, windows of weeks, owners grading themselves. No study follows a dog's life by training method, and none compares methods on severe aggression.

Parenting4 / 8

So, again: would you raise a child and never teach them "no"?

Sixty years of research on parents sorts them by two things: how warm they are, and whether their limits hold.

Diana Baumrind described three styles in 1966. Permissive: warm and responsive, few demands, rules rarely enforced. Authoritarian: demanding, with little warmth and little explanation. Authoritative: warm and demanding at once, with clear limits that are explained and then held. Maccoby and Martin later added a fourth, neglectful, low on both, which does worst of all.

Baumrind 1966, Child Development; Maccoby & Martin 1983, Handbook of Child Psychology

1,435

Across 1,435 studies, children of warm parents who held their limits had the fewest behaviour problems.

A 2017 meta-analysis pooled 1,435 studies of parenting and externalizing problems: aggression, defiance, rule-breaking. Warmth, behavioural control and the authoritative style went with fewer problems, at the time and over time. Authoritarian, permissive and neglectful parenting went with more. The effects were small to very small, and the largest associations of all were between harsh or psychological control and more problems.

Pinquart 2017, Developmental Psychology, meta-analysis of 1,435 studies

308

In 308 studies, children of warm parents who held their limits did better at school. Permissive and authoritarian homes did worse.

The effects were small to very small.

Pinquart 2016, Educational Psychology Review, meta-analysis of 308 studies

2,300

About 2,300 teenagers followed for a year: the gaps between homes held or grew, and the teens of warm, firm parents stayed the best adjusted.

Steinberg, Lamborn, Darling, Mounts & Dornbusch 1994, Child Development

160,927

Across 160,927 children, spanking went with worse outcomes on 13 of 17 measures, and better on none. A "no" that holds is not a hand.

Gershoff & Grogan-Kaylor 2016, Journal of Family Psychology, meta-analysis

This is correlation, and the effects are small.

Mostly Western samples, mostly questionnaires. The style matters, and it isn't everything.

Warmth, and a "no" that holds. Now ask the same about your dog.

Reinforcement5 / 8

Reinforcement teaches what to do

Something good follows a behaviour, and the behaviour becomes more likely. That is positive reinforcement, and it is most of all training.

It builds behaviour. A sit, a recall, a loose leash, settling on a mat: every one of these starts as something the dog offered and got paid for, and then offered again. There is no other way to put a behaviour into a dog. Both sides agree on this, and both sides spend most of their hours doing exactly it.

It is also the pleasant part, for the dog and for the owner. A dog that gets paid learns to try things, to stay in the game, to look at the person holding the food. Study after study finds reward-trained dogs with the fewest stress signs, and that is not surprising. Nothing about a boundary changes any of it.

What reinforcement cannot do is make a behaviour less likely. It only ever says "yes, more of that". So the order matters. Teach what to do first, with rewards, until the dog is fluent. Stopping what not to do is a separate job, and it comes next. Nobody should be correcting a dog for a behaviour it was never taught.

When it is the right tool

  1. Teaching a new behaviour
  2. Building a habit in place of an old one
  3. Anything the dog can be paid for

Nobody teaches a behaviour with a correction. Reinforcement is where every behaviour starts.

Correction6 / 8

What a correction is for

A correction is a brief, unpleasant signal that tells the dog "not that".

A correction is a brief, unpleasant signal that tells the dog "not that". Never pain. Its job is to stop a behaviour and make it less likely next time, and that is all. Nothing is taught with it. Some tools, at some settings, are more than a signal, and whether a device is too much is a separate question about intensity, not about whether the mechanism is sound. This beat is about the mechanism.

Dogs already live with "no"

Adult dogs correct each other, and they correct puppies. A stare, a lifted lip, a body block, a brief warning nip. The other dog adjusts and stays adjusted. This is documented ethology, in Bekoff's work on play and fair play and in Bradshaw and colleagues' work on domestic dogs, not pack theory. Pack theory's mistake was the conclusion it drew, that the human should be in charge and that problems are status-seeking. Drop the conclusion; the observation stands.

Call the mechanism what it is. Something unpleasant appears and the behaviour stops: that is positive punishment, and it is the channel dogs use with each other every day. Discomfort as information is native to how they communicate. The positive approach asks the owner to be the only being in the dog's life who never uses it. Even its own puppy advice, socialise with well-mannered adult dogs, hands the job to dogs who do it well.

The only guardrail is proportionality, not mechanism. A warning nip is not a fight. The dog's correction is proportionate, instant, and over the moment the other dog yields, and that is the standard every human tool should be held to. The dog's own ladder, a look, a lip lift, a growl, a snap, is the least-intrusive principle in fur. This page endorses that principle. Here it is, written by dogs.

You can't explain it to a dog

You cannot tell a dog why not. There is no version of "because the road is dangerous" that reaches it. A consequence has four channels: something good appears, something good disappears, something unpleasant appears, something unpleasant disappears. That is the whole vocabulary a dog can read. Dogs use "something unpleasant appears" with each other daily. The positive approach removes that channel and works with the other three.

The usual objection is that dogs aren't training each other, so the comparison doesn't hold. It answers itself. Neither is the trainer, when correcting. The purpose of a correction, between dogs and on a leash, is the same: stop a behaviour and make it less likely next time. Same operation, same function. Teaching is the previous beat, and it comes first.

First, teach what a correction means

The rule that every correction must come with a reinforced alternative is a reward-only rule of thumb, not a law of learning. No adult dog ever shows the puppy what to do instead, and the puppy learns. What matters is that the dog understands the correction, and that is taught once, at the start, before any correction is used for real.

What is taught is simple. There is a way out. The pressure ends the moment the dog takes it. Good things follow. A dog that has learned this is resilient: a correction is now complete information, "not that", and the dog chooses from everything it already knows how to do. No per-incident alternative needs building, because the dog is not confused. It understands.

This is the concrete form of "stress in the right dose". Schalke's condition with no lasting stress was exactly a dog that could predict the stimulus and control when it ended. A dog that has learned the way out is that dog. Skip this phase and you have the other condition, and the other results.

Stress in the right dose

Stress is not damage. Moderate stress the animal can predict and escape builds resilience; that is the pattern across the human and animal literature on stress and coping. Random, inescapable stress causes harm, and the experiments that established it, Seligman's learned-helplessness work, were done on dogs. Same literature, both halves.

The qualifier does all the work: predictable and controllable. Read that way, the literature is this page's definition of a good correction versus a bad one. A correction the dog saw coming and can end is the first kind. A correction that arrives at random, or that the dog cannot turn off, is the second, and the Belgian military-dog study shows what it costs: worse performance and a more suppressed dog. The tool did not decide which. The handler's timing did.

A correction the dog can predict does not produce lasting stress. A random one does.

Dogs that could see the stimulus coming and avoid it through their own behaviour showed no notable or persistent stress. Dogs corrected unpredictably showed large cortisol rises. Small laboratory sample.

Schalke et al. 2007, Applied Animal Behaviour Science

Dogs given shocks they could neither predict nor escape stopped trying to escape at all.

Dogs that could end the shock by their own action stayed active and learned. Dogs given the same shock with no way out later failed to escape even when escape was easy.

Seligman & Maier 1967, Journal of Experimental Psychology, laboratory dogs

The moment that decides it

The cat across the road. The bird that lifts out of the grass. The high-arousal moment when the dog's whole body is already moving. This is where a reliable "no" is the difference, because nothing the owner is holding competes with what the world is offering. Rewards teach a recall. The question is whether it holds when the dog is not listening at all.

The evidence here has to be read both ways. Dale's kiwi study shows a single, well-timed unpleasant event lasting a year with the collar off, in a conservation setting with a strong stimulus. Cooper's study shows skilled trainers fixing chasing with and without the collar, and the collar group more tense. Take both: a correction can hold in the moment that matters, and the skill to time it is what makes it hold.

A well-timed aversive can stop predatory behaviour, and the effect can last.

In a New Zealand conservation programme, 120 hunting dogs had one session of aversion training around kiwi with a brief, high-level stimulus. A year later 87% still avoided kiwi, in new places, with no collar on.

Dale et al. 2013, Applied Animal Behaviour Science, 120 dogs, one-year follow-up

Recall and chasing were fixed with and without the collar.

Experienced trainers who normally use e-collars got the same improvement working without one, and their dogs showed lower cortisol than the reward-only group. The e-collar group showed more tension.

Cooper et al. 2014, PLOS ONE, 63 dogs; group B, the same trainers without a collar

When not to correct

The positive side's critique that suppression just relocates the problem survives for one class of behaviour: symptoms. Fear, pain and unmet needs get treated, not corrected. Rule out medical causes and fear first. Don't correct a dog that is over threshold; the work happens below it. Refer when the case is beyond you. This is a rule about when, and it is a real one. It is not a requirement to bundle every "no" with a "yes".

Correction in five sentences

  1. Discomfort as information is the dogs' own channel, and the only one that needs no language.
  2. Its job is to stop a behaviour and make it less likely. Same function between dogs and on a leash.
  3. The dog is first taught what a correction means and how to end it. After that a correction stands alone.
  4. Proportionate, instant, over the moment the dog yields. The dog's own ladder is the standard.
  5. Not for symptoms. Fear, pain and unmet needs get treated, not corrected.
Freedom7 / 8

Can rewards alone reach reliability in daily life?

Training is not for obedience. It is for the size of the dog's life.

The reward-only route to reliability is real, and for many dogs it arrives. You teach the behaviour, pay it, raise the difficulty a step at a time, and practise it in new places. Meanwhile every situation the dog can't yet handle has to be arranged so the mistake never happens: the leash stays on, the gate stays shut, the trigger stays at a distance. That is management, and both sides start with it.

The question is whether it is a bridge or a destination. For a hard case, a careful reward-only plan tends to contain a long line, secure fencing, a muzzle that may stay on for good, and no guest greetings for months. If the behaviour doesn't shift, the equipment stays and the walk gets shorter and more planned. No study tracks how much equipment a dog wears over its life by method. This is what we see in practice, and we say so.

A dog with freedom and boundaries gets the woods off leash, the cafe, the family lunch, the beach, the new dog at the gate. The boundaries are what make the freedom possible: the dog can be trusted with the world because it knows what in the world is off limits. In children, warmth paired with limits is what the parenting research calls authoritative. In dogs, it is the trained dog in the list below.

Can rewards alone reach reliability in daily life?

Managed dog

Off-leash walk in the woods: Long line only, A cafe with friends: Left at home, Family gatherings: Muzzled, or behind a gate, Beach and open fields: Long line, quiet hours, Meeting a new dog: Avoided; cross the street, Guests at the door: Crated until they leave, Holiday with the family: Kennel or a pet-sitter, Children in the garden: Supervised and muzzled

Trained dog

All of it, on a loose leash or none.

No study measures a dog's freedom over a lifetime by training method.

This list is what we see in practice. The one measured signal in the sources cuts the other way too: in Cooper et al. 2014, the e-collar dogs explored their surroundings less. A tool, badly timed, shrinks the dog's world as well.

Management8 / 8

A dog that needs constant management is not a dog the positive approach worked on

The positive approach works.

The positive approach works. It manages situations during learning, on purpose, so the dog is never rehearsing the mistake while the new behaviour is being built, and the point of that management is reliability in real life. The leash, the gate and the distance are scaffolding. Scaffolding is meant to come down.

So the test is simple. Look at the dog at three years old. If the long line is still on, the guests are still met behind a gate, and the walk is still at quiet hours, the plan did not reach its goal. The dog is being managed, and the management is being called training.

If you are working the positive approach and the management never stops, you are not working.