Positive reinforcement: the solid principle and the nuances of the method

In 30 seconds

There is one solid principle and one nuance worth separating. The principle: a behavior followed by something the dog finds pleasant tends to be repeated, so rewarding what you want works better than punishing what you don't. This has broad veterinary support (AVSAB, ESVCE, APDT, CCPDT) and aversive methods (leash corrections, yelling, e-collar shocks) consistently come out worse: they damage the bond, raise stress, and increase the probability of aggression. The nuance: "rewarding" is not the same as "giving treats with a clicker." For our school, the highest-value reward is usually your presence, your affection, and play, plus letting the dog learn to handle situations on his own. Food and a marker are useful tools, not the core of the work. If a trainer proposes a prong or e-collar in 2026, find another.

The principle: reward what you do want

It is one of the four mechanisms of operant conditioning described by B.F. Skinner in the 1930s. The technical definition is straightforward: a behavior followed by a pleasant consequence tends to be repeated.

With a dog, this means that if he sits on your cue and receives something he enjoys, he will be more likely to do it next time. No one disputes this part: teaching by reinforcing the desired behavior is preferable to correcting by force. What does deserve some nuance is what counts as a reward and what role we give it.

"Rewarding" is not just giving treats with a clicker

It helps to separate the principle from its popular version. In recent years a narrow idea of "positive training" has spread, one that almost reduces it to a formula: appetizing food plus a marker (clicker or verbal cue) that signals the reward is coming. It works for teaching many behaviors, but it leaves out what our school considers the core of the work.

For us, the highest-value reward is usually not the treat. It is:

• Your presence. Being there, at his side, transmitting a sense of security. Many times the dog does not need a piece of chicken put in his mouth; he needs to know you are there and that everything is okay.
• Affection. A calm touch at the right moment, a quiet tone of voice, coming through an uncomfortable situation together and celebrating it with you. That reinforces.
• Play. For many dogs, a shared play session is worth more than any food reward, and it also builds the bond.
• Decision-making and self-management. Letting the dog look around, assess, and work things out on his own, accompanied and guided, rather than keeping him constantly focused on your hand and the next reward.

Food has its place, especially when introducing a new behavior or competing with strong distractions. But it is one reward, not the primary one, and it's worth not building all your work around it. A dog that only responds when he smells the treat has not learned the same thing as a dog that trusts you.

The clicker, in this framework, is an optional tool: some handlers find it useful for marking the exact instant, but it is not essential and it is not "the right way" to do things. If you use it, use it because it works for you, not because it seems mandatory.

The four quadrants (and why they are not everything)

Any interaction with your dog fits, on paper, into one of these four quadrants:

Operation	Add something	Remove something
Increases behavior	Positive reinforcement (reward)	Negative reinforcement (relief)
Decreases behavior	Positive punishment (hit, yell)	Negative punishment (withdraw attention)

Serious professionals work primarily with positive reinforcement (teaching what you do want) and set aside the classic aversives: positive punishment (pain, fear) and misapplied negative reinforcement. This is a useful map.

One school-level caution, though: the dog is not a quadrant machine. Thinking about training purely as "what consequence do I add or remove" leaves out the bond, the presence, and emotional guidance, which for us is where real learning actually happens. The chart helps you understand the mechanics; it does not replace the relationship.

What the evidence says about aversive methods

The most comprehensive review to date is Ziv (2017), which compiled 17 studies on aversive methods. The findings are consistent:

• Dogs trained with aversive methods show more stress signals (panting, lip-licking, gaze avoidance) during and after training.
• Increased likelihood of aggressive responses directed at the handler, strangers, or other dogs.
• The emotional bond measured in attachment tests with the handler deteriorates.
• Long-term efficacy is not superior to positive reinforcement. In most measurements, it is worse.

Vieira de Castro et al. (2020) reinforced this evidence by measuring salivary cortisol: dogs trained with aversive methods showed significantly higher cortisol levels, a biological marker of chronic stress.

The American Veterinary Society of Animal Behavior (AVSAB) made it clear in their 2021 position statement: punishment-based methods should not be used in dog training except in exceptional clinical circumstances under veterinary behaviorist supervision. This is the solid foundation of the article and it is not a school debate: leash corrections, prong collars, and e-collars have no place in modern training.

When someone says "positive training doesn't work for my dog"

Usually one of these things is happening. Notice that almost all of them point back to relying too much on the treat and not enough on everything else.

1. Everything has been put into food

If the only resource is a food reward and the dog is too aroused to eat, it can look like "it's not working." What's often missing is not a tastier treat, but a reduction in arousal: more distance from whatever is bothering him, your presence nearby, a calm tone. When the dog settles back into a receptive state, his ability to learn returns.

2. The behavior is being asked for in an environment that is too challenging

Learning follows a progression: first at home without distractions, then in the yard, then on a quiet street, then at the park. Skipping stages is the number-one cause of apparent failure.

3. The wrong behavior is being reinforced without realizing it

When your dog barks at the window and you yell, you are giving attention. Attention equals reinforcement. You are making the behavior more likely, not less. The same happens when, faced with fear, you pick him up and repeat "it's okay, it's okay": you may be confirming that something actually is wrong.

Positive reinforcement is not "anything goes"

This is worth clarifying because the confusion is common. Rewarding what you want does not mean an absence of boundaries or a dog that does whatever he pleases. It means boundaries are taught by accompanying and guiding, not by inflicting pain or fear.

A dog reinforced with presence and affection is not a spoiled dog: he is a dog that trusts you, has clear rules because you have established them with confidence, and knows you are there. Structure exists; what changes is how it is maintained.

If your dog jumps on guests, the answer is not a leash correction to teach him not to. It is:

• Anticipate: ask for a calm position before guests come in.
• Reinforce that calm with your attention, a touch, or a reward.
• Do not reinforce the jumping (no contact, no drama) and redirect to the behavior you do want.

The result is a dog that chooses to greet politely because it works for him, not a dog that is afraid of making a mistake.

What to check

• Whether you are relying only on food, or also on your presence, affection, and play as rewards.
• Whether your dog responds out of trust and bond, or only when he can smell the treat in your hand.
• Whether you are delivering the reward at the right moment, while the desired behavior is happening.
• Whether you have raised the environmental difficulty too quickly.
• Whether, without realizing it, you are reinforcing the behavior you do not want with attention or contact.
• Whether the trainer you are considering uses prong collars, e-collars, or leash corrections: change trainers.