Motivation in Dog Training: How Dogs Learn Through Rewards

In brief

Three things explain why your dog sometimes works with intense focus and sometimes seems not to be working with you at all. First, motivation is not a character trait but a state – it fluctuates with the environment, the dog's condition on the day, the difficulty of the task and how the dog feels physically. Second, what matters is not the reward as such but its value for this dog in this situation: what works in the living room may do nothing outdoors. And third, a sudden loss of motivation is a signal that deserves to be taken seriously – often it reflects a task that is too hard, and more often than people assume, it reflects pain.

For context, right from the start: that reward-based training is the better choice for welfare has been studied in dogs and is well supported. That it also gets results faster or more reliably for every training goal is not – the evidence on that is mixed.

Why is one exercise mastered after three repetitions while another just won't stick? The answer lies in the interplay of learning history, emotional state and, above all, motivation. Dogs bring a remarkable willingness to cooperate with people, but it cannot be taken for granted. Once we understand what drives a dog, we can build and stabilize the behavior we want in a targeted way.

1. What motivation is and what it is not

1.1 A state, not a character trait

Motivation describes an internal state that influences how likely a particular behavior is to occur. A motivated dog works with more focus, stays with a task longer and copes better with distractions. The key point is in the word state: it can change within a single training session, and it changes with the location, the time of day and whatever happened beforehand. "My dog isn't food-motivated" is therefore rarely a description of the dog and usually a description of a situation.

1.2 How to read the evidence

The statements in this guide rest on very different foundations. Some come from studies in dogs, others from general learning research, and some are well-founded practical experience without dedicated studies behind them. Where a statement rests only on the last of these, this is said.

1.3 What this guide does not claim

It does not claim that motivation can be engineered once you find the right reward. And it does not claim that every drop in motivation is a training problem – a considerable share of them are precisely not.

2. Extrinsic and intrinsic motivation

2.1 The two forms

Learning psychology distinguishes two forms:

• Extrinsic motivation – The dog shows a behavior because an external reward follows it. This is the most common form in training: the dog sits because he gets a treat for it.
• Intrinsic motivation – The activity itself is already attractive. For example, when a dog starts sniffing on his own or retrieves enthusiastically without any additional reward being needed.

2.2 How they work together

The two forms complement each other. Ideally, behavior built up extrinsically becomes stabilized over time by intrinsic components – the dog then works not only for the reward but because working together has itself become pleasant for him. How robust this distinction is in dogs, however, is hard to test: whether a behavior is motivated "from within" or by a long reinforcement history that we can no longer trace is almost impossible to tell from the outside. The distinction is a useful thinking tool, not a measured quantity.

3. The neurobiological side

3.1 Dopamine is not a happiness hormone

Dopamine plays a central role – but differently than it is often portrayed. Dopamine is involved mainly in motivation, anticipation and learning processes, not in the experience of pleasure itself. The distinction is not hairsplitting: it explains why a dog can put real effort into getting something that doesn't mean all that much to him once he has it.

3.2 Prediction error

The principle of prediction error is particularly revealing: it is not the reward itself that drives the dopamine system most strongly, but the deviation from expectation. An unexpectedly good reward produces a much stronger signal than one that was anticipated anyway; if an expected reward fails to appear, the signal drops below baseline. This is exactly why training that never changes becomes dull over time. More on this in Prediction Error in Dogs.

3.3 Where these findings come from

Precision is needed here. The key review of reward signals in the brain (Schultz, 2015) relies at its core on single-cell recordings in monkeys (10.1152/physrev.00023.2014). The principle is considered well supported across mammalian species, but when we talk about prediction error in dogs, we are transferring a finding from another species. That doesn't make it wrong – it just means nobody has measured it in dogs at this resolution. The bigger picture is described in The Neurology of Dog Behavior.

3.4 What happens under stress

When aversive methods are used, the picture shifts: via the sympathetic nervous system, adrenaline is released within seconds, and cortisol follows more slowly via the HPA axis. A dog under significant stress can lose the learning capacity he would have in a relaxed state; attention narrows and behavior becomes more rigid. How cortisol in dogs should actually be interpreted, and where measuring it gets tricky, is covered in Chronic Stress in Dogs.

Our position on this: punishment can interrupt behavior in the short term – nobody seriously disputes that. The question is what was learned in the process and at what cost. A behavior that has been stopped is not a problem that has been solved, and with an insecure dog, punishment often intensifies exactly the state that drives the behavior. For more on aversive methods and on why we don't use punishment in training, see Aversive Training Methods in Dogs.

4. What influences motivation

4.1 The factors at a glance

• Reward for the behavior – A behavior is shown more often when a positive consequence follows.
• Expectation – Dogs associate cues, context and their own actions with possible rewards. This expectation increases their readiness to become active.
• Level of difficulty – Motivation drops when a task is too hard or the dog doesn't understand what is being asked.
• Environment and situation – Stress, fear, unfamiliar places or strong distractions can reduce motivation considerably. Stressors often add up over the course of a day rather than acting one at a time; see Trigger Stacking in Dogs.
• Arousal level – Too little and too much arousal both get in the way. How to regulate it is covered in Arousal Regulation in Dogs.
• Physical condition – Pain, fatigue, a full stomach or illness have a direct effect on willingness to learn.

4.2 When motivation suddenly collapses

The last point deserves special attention, and here there are solid numbers. Mills and colleagues (2020) reviewed the case records of 100 dogs referred for behavioral medicine, drawn from several authors' caseloads. Their conservative estimate: in about a third of the referred cases, a painful condition was involved, and in individual caseloads the proportion was close to 80 percent. The authors recommend addressing suspected pain early rather than only checking for it once behavior therapy fails to work (10.3390/ani10020318).

How far this goes: these are clinical cases, not a random sample of the dog population. Anyone who consults a veterinary behaviorist has usually been through quite a bit already. For everyday life, it still means: if a dog suddenly and for no apparent reason cooperates much less than usual, have it checked by a veterinarian. Before you tinker with the training, physical causes should be ruled out. How to recognize pain is covered in the guide on recognizing pain in dogs; for the link with aggression, see Chronic Pain and Behavior.

4.3 In practice

Anton, a male dog, had been an enthusiastic training partner for years – then, week by week, he became more sluggish, refused to sit and broke off exercises. His owner suspected stubbornness. The veterinary examination revealed a painful change in his back. After treatment, Anton was his old self again. A dog who "doesn't want to anymore" may, in reality, be a dog who "can't anymore".

5. Rewards in dog training

5.1 Primary rewards

Primary rewards work without any prior learning because they address basic needs: food, water, rest, movement and play.

Food rewards are especially practical because they can be dosed precisely and delivered quickly. Adjust the size to the situation: tiny pieces for many quick repetitions, something special for real breakthroughs.

5.2 Secondary rewards

Secondary rewards only acquire their meaning through learning – they take on a positive value because they have repeatedly been paired with something pleasant. This mechanism is based on classical conditioning, which goes back to the work of Ivan Pavlov.

Example: Your dog gets a treat every time you say "Good!". After a number of repetitions, the word itself already triggers attention and anticipation.

5.3 What preference tests actually show

It is worth taking a close look here, because the usual list – "praise, play, attention, petting" – glosses over a ranking that has been studied in dogs.

In several series of experiments, Feuerbacher and Wynne compared what dogs will actually work for. Brief social interaction – four seconds of petting with praise – maintained a behavior much less well than food and, in parts of the experiments, was barely distinguishable from no reward at all (10.1901/jeab.2012.98-105). Given a direct choice between the two, most pet dogs chose food over petting (10.1002/jeab.81). And in a choice between petting and vocal praise, petting won clearly (10.1016/j.beproc.2014.08.019).

So the ranking, roughly speaking, is: food over petting over praise. That doesn't mean praise is worthless – as a marker signal that announces a real reward, it can work well. But it does mean that a friendly "Good job" on its own is, for most dogs, not a reward that will sustain a behavior. How far this goes: these were laboratory situations with short sequences, some with shelter dogs, some with family dogs, and individual differences were considerable. For your dog, the rule still holds: watch what he actually accepts. And not every dog finds physical contact pleasant, least of all in the middle of a phase of concentration.

5.4 In practice

Ruby, a female dog, was considered "not food-motivated". In fact, the dry standard treat was simply too boring for her – and when training outdoors, it was also uninteresting compared with anything that moved. With a short game of tug as the reward, Ruby was suddenly fully focused. It wasn't the dog who lacked motivation; the reward was the wrong one.

5.5 Self-reinforcing behavior

Behaviors that are rewarding in themselves for the dog are a particular challenge: chasing wildlife, joggers or bicycles, digging, extensive sniffing or rolling in smells.

Such behaviors are highly attractive, and the reward a human can offer often can't compete – especially when whole segments of the predatory motor sequence are engaged.

Approach: Instead of just forbidding the behavior, it is worth offering an alternative that the dog also finds attractive. A dog who loves to chase can learn to focus on a search game or scent work instead. Two things are crucial here: the cue has to come early, before the dog is already caught up in the sequence internally. And as long as the alternative isn't yet reliable, you need management – a long line, distance, choice of route. Management is not failure; it is the precondition for training being possible at all.

6. Using motivation deliberately in training

6.1 Timing

The reward should follow the desired behavior immediately, ideally within one to two seconds. Otherwise, there is a risk of accidentally reinforcing something else – a glance away, a random movement.

6.2 The marker signal, realistically assessed

This is where a marker signal helps: a short word or a clicker that pinpoints the right moment exactly. The reward itself can then come a moment later without the information being lost. This makes precise rewarding possible even at a distance.

The widespread assumption that the clicker is superior to a marker word, however, does not hold up to scrutiny. Chiandetti and colleagues (2016) divided 51 dogs into three groups of 17 each: teaching a new behavior with a clicker, with the marker word "Bravo", or with food alone. They found no differences between the groups, either in acquisition or in transfer to a modified setup – a result the researchers themselves had not expected (10.1016/j.applanim.2016.08.006).

In practice, this means: use the marker you yourself can work with cleanly. If you need both hands for the leash and handling, this evidence says you can stick with a word without a guilty conscience. How to introduce a clicker properly is explained in the guide on clicker training for dogs.

6.3 Individual adjustment

Not every dog is motivated by the same reward – and the same dog is not motivated equally in every situation. Watch what really interests your dog in everyday life and build a small ranking: what works at home, what you need outdoors, what you save for the difficult moments. If food isn't available for once, you'll find ideas in the guide on dog training without treats.

6.4 Adjusting the level of difficulty

Successful training works with small, achievable learning steps. Each step should be chosen so that your dog is very likely to succeed. Gradually shaping a behavior ("shaping") is a proven method for this.

Also expect that what has been learned will at first work less well in new places. That is not a setback but normal generalization – see also the guide on trick training, where the same principle applies. What happens when the task is permanently beyond what the dog can manage is described in The Neurobiology of Frustration in Dogs.

6.5 Variable reinforcement, put in context

A reward that is always identical can lose its effect. That is why many trainers work with variable reinforcement: not every repetition is rewarded the same way, and now and then there is something especially high-value.

The right order matters here. Variable reinforcement is suited to stabilizing behavior that has already been built – it does not replace clean, small-step training. If you start too early, while the dog doesn't yet know for sure what is being asked, you are more likely to create frustration than motivation. Consolidate first, then vary. Which reinforcement schedules exist and how they affect resistance to extinction is covered in Reinforcement Schedules in Dogs.

6.6 Breaks and satiation

Short, frequent sessions usually work better than one long hour. The optimal length depends on the dog, his age, his training level and the difficulty of the task. If your dog is already full before training, his motivation to work for food drops – count the training portion as part of his daily ration.

And: stop while things are still going well. A dog who leaves a session with a sense of success starts the next one more motivated.

7. Common mistakes

7.1 What it usually comes down to

• Asking too much – expectations that are too high or progress that is too fast can lead to frustration.
• Poor timing – if the reward comes too early or too late, the dog can learn the wrong associations.
• Satiation – the same reward every time can lose its value.
• The wrong reward – what thrills one dog leaves another cold.
• Pressure or impatience – tension in the human can carry over to the dog.
• Sessions that are too long – waning concentration produces mistakes and frustration.
• Fading out too early – if you reduce the reward before the behavior is truly solid, you risk it disappearing again. How behavior responds when reinforcement stops is described in Extinction in Dog Behavior.

7.2 In practice

Selma, a female dog, had a reliable recall – until her owner stopped giving treats because "she knows it now". Within a few weeks, the recall became patchy. Not disobedience, but very likely the result of missing reinforcement. Once rewards were reintroduced, though on an irregular basis, the behavior stabilized again.

8. Motivation and the human–dog bond

Motivation and the bond between you and your dog influence each other. A dog who has a reliable relationship with his person is generally more willing to cooperate – and training together with positive reinforcement can in turn strengthen the bond between human and dog.

When your dog experiences that cooperating pays off and is predictable, he develops a positive expectation of training situations. Instead of feeling pressure, he actively seeks out the interaction. The learning-theory framework for this is provided in the guide on modern learning theory.

9. Where the evidence comes from

The evidence on reward-based training rests on three pillars of differing strength.

Owner surveys. Hiby and colleagues (2004) surveyed 364 dog owners about seven everyday exercises. Rated obedience was positively associated with the number of tasks trained with reward-based methods, not with punishment; conversely, the number of reported problem behaviors was associated with the use of punishment, not with reward (10.1017/S0962728600026683). Surveys like this capture what owners remember and report, and they show associations, not causes.

Field study with objective measures. Vieira de Castro and colleagues (2020) observed 92 family dogs from seven Portuguese dog training schools. Instead of asking questions, they filmed the training sessions, analyzed stress behavior and measured cortisol in saliva; 79 of the dogs also completed a cognitive bias test in an unfamiliar location. Dogs from schools using aversive methods showed more stress behavior and greater increases in cortisol, and responded more cautiously to the ambiguous situation (10.1371/journal.pone.0225023). This is the strongest of the three pillars, but it is not a randomized trial either: the owners had chosen their school themselves.

Foundational work. Skinner's account of operant conditioning (1953) provides the theoretical framework – behavior is shaped by its consequences, and punishment usually suppresses behavior without building stable alternatives. This is laboratory work with other species, not a study in dogs.

What stands out in this list: all three concern welfare and side effects; none concerns effectiveness as such. Vieira de Castro and colleagues (2021) published a study protocol to address exactly this gap and reviewed the existing literature in it: five studies point to greater effectiveness of reward-based methods, one to the opposite, and three find no difference (10.1371/journal.pone.0247321). It is a protocol whose own results were still pending at the time of publication.

10. What the research does not (yet) show

• Not that reward-based training gets results faster or more reliably for every training goal. The literature on effectiveness is mixed; the literature on welfare is not.
• Not that a clicker is superior to a marker word. This has been tested in dogs and not confirmed.
• Not that praise works as a stand-alone reward. In direct choice experiments, it clearly loses out to food and petting.
• Not that variable reinforcement has a quantifiable advantage over continuous reward in dogs. The principle comes from learning research in other species; comparative studies in dogs are lacking.
• Not that intrinsic and extrinsic motivation can be cleanly separated in dogs. The distinction is a thinking tool, not a measurable quantity.
• Not how often pain is behind a loss of motivation in dogs in general. The available figures come from behavioral medicine referral cases.

11. In practice at unterHUNDs

Recall. Instead of putting dogs under pressure with repeated calling, we build the recall deliberately through positive reinforcement. High-value rewards are especially important in the building phase. Only once the behavior is solid does the reward become more variable.

Loose-leash walking. Instead of leash jerks, we reward every moment of loose walking. This way, loose-leash walking grows out of positive expectation rather than out of avoiding discomfort.

With fear. We work especially gently with fearful dogs, and always below their stress threshold. Here, rewards serve to build positive associations – motivation comes from safety, not from pressure.

12. Conclusion

Motivation is not an add-on to training but its foundation. When your dog understands that the desired behavior pays off for him and the situation makes sense to him, he shows it more often, more consistently and more reliably.

What matters is less the perfect technique than the willingness to look closely: what does this dog really value – rather than what he ought to value? What is too hard here and now? And if motivation suddenly collapses: could something physical be behind it?

A motivated dog doesn't just learn more efficiently. He also works in a more relaxed, more confident way, with a genuine willingness to cooperate.

Key takeaways

• Motivation is a state that fluctuates with location, the dog's condition on the day and task difficulty, not a fixed character trait.
• In direct choice experiments, most pet dogs preferred food to petting and petting to vocal praise (Feuerbacher & Wynne, 2012, 2014, 2015).
• A clicker offered no measurable advantage over a marker word or food alone (Chiandetti et al., 2016; 51 dogs in three groups).
• In about a third of dogs referred for behavioral medicine, a painful condition was involved (Mills et al., 2020).
• In surveys, more frequent use of punishment went along with more reported problem behavior, not with better obedience (Hiby et al., 2004).
• Among 92 filmed family dogs, those from schools using aversive methods showed more stress behavior and greater cortisol increases (Vieira de Castro et al., 2020).
• Prediction error as a mechanism comes mainly from recordings in monkeys, not from studies in dogs (Schultz, 2015).

References

Chiandetti, C., Avella, S., Fongaro, E., & Cerri, F. (2016). Can clicker training facilitate conditioning in dogs? Applied Animal Behaviour Science, 184, 109–116. https://doi.org/10.1016/j.applanim.2016.08.006

Feuerbacher, E. N., & Wynne, C. D. L. (2012). Relative efficacy of human social interaction and food as reinforcers for domestic dogs and hand-reared wolves. Journal of the Experimental Analysis of Behavior, 98(1), 105–129. https://doi.org/10.1901/jeab.2012.98-105

Feuerbacher, E. N., & Wynne, C. D. L. (2014). Most domestic dogs (Canis lupus familiaris) prefer food to petting: population, context, and schedule effects in concurrent choice. Journal of the Experimental Analysis of Behavior, 101(3), 385–405. https://doi.org/10.1002/jeab.81

Feuerbacher, E. N., & Wynne, C. D. L. (2015). Shut up and pet me! Domestic dogs (Canis lupus familiaris) prefer petting to vocal praise in concurrent and single-alternative choice procedures. Behavioural Processes, 110, 47–59. https://doi.org/10.1016/j.beproc.2014.08.019

Hiby, E. F., Rooney, N. J., & Bradshaw, J. W. S. (2004). Dog training methods: their use, effectiveness and interaction with behaviour and welfare. Animal Welfare, 13(1), 63–69. https://doi.org/10.1017/S0962728600026683

Mills, D. S., Demontigny-Bédard, I., Gruen, M., Klinck, M. P., McPeake, K. J., Barcelos, A. M., Hewison, L., Van Haevermaet, H., Denenberg, S., Hauser, H., Koch, C., Ballantyne, K., Wilson, C., Mathkari, C. V., Pounder, J., Garcia, E., Darder, P., Fatjó, J., & Levine, E. (2020). Pain and problem behavior in cats and dogs. Animals, 10(2), 318. https://doi.org/10.3390/ani10020318

Pawlow, I. P. (1927). Conditioned Reflexes: An Investigation of the Physiological Activity of the Cerebral Cortex. Oxford University Press, London. No DOI has been assigned to this work.

Schultz, W. (2015). Neuronal reward and decision signals: from theories to data. Physiological Reviews, 95(3), 853–951. https://doi.org/10.1152/physrev.00023.2014

Skinner, B. F. (1953). Science and Human Behavior. Macmillan, New York. No DOI has been assigned to this work.

Vieira de Castro, A. C., Fuchs, D., Morello, G. M., Pastur, S., de Sousa, L., & Olsson, I. A. S. (2020). Does training method matter? Evidence for the negative impact of aversive-based methods on companion dog welfare. PLOS ONE, 15(12), e0225023. https://doi.org/10.1371/journal.pone.0225023

Vieira de Castro, A. C., Araújo, Â., Fonseca, A., & Olsson, I. A. S. (2021). Improving dog training methods: efficacy and efficiency of reward and mixed training methods. PLOS ONE, 16(2), e0247321. https://doi.org/10.1371/journal.pone.0247321

Frequently asked questions about motivation in dog training