How dogs learn: rewards, timing and management
Behavior that pays is repeated. This chapter covers the marker, the timing of a reward, management, and why this manual uses reward-based training without pain, intimidation or physical corrections.
- Start here: the essentials
- 1. Behavior that pays is repeated
- 2. Timing, and the marker that buys time
- 3. What the veterinary bodies and the studies say
- 4. Management: the other half of training
- 5. Age, breed and a new dog
- 6. See it: Rosie's first ten days
- 7. Practice: charge the marker and start the log
- 8. Your dog
- 9. Check your work
- 10. Knowledge check and scenario
Start here Essentials
At minimum, give your dog a marker that means "food is coming," and write down what you will manage instead of train. The research, the consequence table's examples and the age detail are Go deeper.
- Choose a marker. A short word such as "yes," or a clicker. Use the same one every time (timing and the marker).
- Charge it. Say the marker, then hand over a small treat, ten times in a row. Do three sessions of about a minute each, spread across the day (practice).
- Play the name game once. Say your dog's name; the moment the dog looks at you, mark and treat (practice).
- Write three management items. Three things you will prevent with gates, leashes, doors or tidying, in the household sheet (management).
- Log the first session. One row in the sessions sheet: date, skill, place, minutes, repetitions.
Minimum time: about 15 minutes across one day. Done when: your dog turns toward you at the marker sound before the treat appears, and one row is in the log.
1. Behavior that pays is repeated Understand
Every training method in this manual rests on one observation: a behavior that is followed by something the dog wants becomes more likely to happen again. That is reinforcement: any consequence that makes the behavior before it more likely in the future. VCA Animal Hospitals' page on reinforcement and rewards puts the practical version plainly: the reward has to follow the behavior immediately, or it risks rewarding whatever the dog did in between. Physical principle
Notice what the definition does not say. It does not say "food," and it does not say "what you meant to reward." The dog decides what counts as a reward, and the dog's next behavior is the only evidence of whether something worked. A piece of kibble may be a reward in the kitchen and worth nothing in a park full of squirrels. A pat on the head may feel like praise to you and like an imposition to a dog that ducks away from hands. A reward, in this manual, is anything the dog will work to get: food, a toy, a game, a sniff of a lamp post, the door opening. Whether it reinforces is settled by what happens next time.
The same rule runs in the background whether or not anyone is training. VCA notes that attention, even scolding, can reinforce an unwanted behavior. A dog that jumps on a visitor and gets pushed, talked to and looked at has been paid in attention. A dog that barks at the treat cupboard until someone opens it has been paid in food. The American Veterinary Society of Animal Behavior (AVSAB) makes the same point in its position statement on dominance: most of the behaviors owners complain about, such as jumping up, stealing food and not coming when called, happen because they were rewarded by accident and nothing else was taught. That is good news. It means most everyday problems are a matter of changing what pays.
Get the behavior, then name it
VCA's order is: get the behavior first, then add the cue. A cue is the word or hand signal that tells the dog which behavior will pay right now. Saying "sit" ten times to a dog that does not yet know what sitting earns teaches the dog that "sit" is background noise. So this manual gets the behavior by one of three routes and names it only once the dog offers it readily:
- Luring: guiding the dog into position with food in your hand, for example moving a treat up and back over the dog's head so the rear end folds into a sit. VCA's training basics page says to switch to an empty hand within 5 to 10 repetitions, so the hand movement, not the food, becomes the signal. First behaviors teaches this in full.
- Capturing: waiting for the dog to do the behavior on its own (lie down on a mat, look at you) and marking and rewarding it when it happens.
- Shaping: rewarding small steps toward a behavior, one approximation at a time, such as a glance at the mat, then a step toward it, then a paw on it.
Every time while learning; sometimes once it is reliable
The American Animal Hospital Association's 2015 behavior management guidelines say that a new behavior is learned best when it is rewarded every time, and maintained best when rewards come at random intervals once the behavior is reliable. VCA says the same. The first is called continuous reinforcement; the second, a variable schedule (or intermittent reinforcement). Convention
Why the switch? Every-time rewards make the connection clear while the dog is still working out what earns the treat. Once the behavior is solid, unpredictable rewards keep the dog trying, and you no longer need a pouch of food to get a sit at the door. Switching too early is one of the commonest ways a cue falls apart, so this manual switches only after the log shows the behavior is reliable in that place. "Reliable" is measured, not felt: the cue inventory asks for five out of five before a stage counts. Proof it and keep going covers fading food in detail.
The four consequences
Behavior science sorts consequences into four kinds by two questions: is something added or taken away, and does the behavior become more or less likely afterwards? "Positive" and "negative" here mean added and removed, not good and bad. The definitions below follow the Merck Veterinary Manual's professional chapter on treating behavior problems and Karen Pryor Clicker Training's glossary.
| Consequence | What happens | Effect on the behavior | Everyday example | In this manual |
|---|---|---|---|---|
| Positive reinforcement | Something the dog wants is added | More likely | The dog sits; a treat arrives | The main tool |
| Negative punishment | Something the dog wants is taken away | Less likely | The dog jumps; the person turns and steps away, so the attention stops | Used mildly, alongside teaching what to do instead (Everyday problems) |
| Negative reinforcement | Something the dog wants to avoid is taken away | More likely | Leash pressure stops when the dog stops pulling | Not taught; named so you recognize it |
| Positive punishment | Something the dog wants to avoid is added | Less likely | A leash jerk, a shout, a squirt of water after a bark | Not taught; named so you recognize it |
This table describes how learning works. It is not a menu. This manual teaches positive reinforcement, uses the mildest form of negative punishment (a reward stops when the dog does something unwanted) only alongside teaching what to do instead, and does not teach how to apply punishment or pressure. The table also assumes a healthy, comfortable dog. A dog that is sick, in pain or frightened does not learn the way it predicts, and a sudden change in behavior is a veterinary question before it is a training one: the ASPCA's behavioral help page says only a licensed veterinarian can diagnose the medical conditions behind it. Ask the authority
This manual's choice in one line: reinforce what you want, and prevent what you do not want with management, so the unwanted behavior stops paying. Merck's professional chapter explains why punishment is a poor teacher even where it suppresses a behavior: punishment does not teach the dog what to do, and it can create a conditioned fear of the person or the place. The methods section gives the position statements and studies behind that choice.
Worked examples of the four consequences. The same everyday moment can be read four ways, depending on what the dog gets and what it does next time.
- The dog jumps on you when you come home, and you say hello and pet it. Attention was added and jumping continues: positive reinforcement of jumping, by accident. The fix is to change what pays (Everyday problems): greet only when four paws are on the floor, and teach a sit at the door so the dog has a way to earn the greeting.
- The dog mouths your hand during play, and you stand up and leave for twenty seconds. Play was taken away. If mouthing drops over the next week, that was negative punishment. It works best when the dog also has a toy to bite and a way to earn play back. On its own it teaches only what not to do.
- A trainer tightens the leash until the dog sits, then lets it go slack. Pressure was removed when the dog sat: negative reinforcement. This manual does not teach this, because the dog learns the sit to make discomfort stop, and AVSAB's 2021 statement lists leash jerks and pressure-based methods among those to avoid.
- A dog barks at the window and someone sprays it with water. Something unpleasant was added: positive punishment, if the barking drops. AVSAB names squirt bottles in its list of tools to avoid. The barking also has a cause (a passing dog, a delivery), and the spray does nothing about it. Everyday problems treats window barking with management first.
Extinction is the fifth term you will meet: when a behavior that used to pay stops paying, it fades. Merck's owner chapter on behavior modification warns that a behavior on its way out often gets worse first. The dog that used to be fed for barking at the table barks louder and longer before giving up. If the household gives in at the loudest moment, it has just rewarded louder barking. That is why the household agrees the rules before a plan starts (Set up to train).
2. Timing, and the marker that buys time
Dogs connect a consequence with what they were doing at that moment. VCA's training basics page says the marker lasts a fraction of a second and the reward comes right after. If you ask for a sit, the dog sits, and you then fish in your pocket for three seconds while the dog stands up to see what you are doing, the treat arrives on a standing dog. You meant to reward the sit; you rewarded standing up. Physical principle
Human hands are slow. Nobody can deliver a treat in a fraction of a second while the dog's rear is still touching the floor, especially when the dog is across the room. The solution is a marker, also called a bridge: a short, distinct signal that means "that was it, and food is coming." The marker does the timing; the treat follows when your hands catch up.
Karen Pryor Clicker Training's introduction to clicker training explains that the marker does not have to be a clicker. A clicker makes a sound that is the same every time and stands out from speech, which helps. A word works too, and your voice goes everywhere you do. Pick a word that you do not say all day: "yes" works for many people; "good" is said too often to stand out; some people use a tongue click. Whatever you choose, everyone in the household uses the same one (the household sheet in Set up to train records it). Preference
Two rules for the marker
- Mark the instant, then treat. Karen Pryor's fifteen tips for getting started say to click during the behavior and deliver the treat after. Mark while the rear touches the floor, not after the dog has already stood up again. Preference
- Every mark is followed by a treat. The same tips say one click per behavior, and that a click is followed by a treat even if you clicked by mistake. The marker's meaning comes from what follows it. A marker that is sometimes followed by nothing slowly loses its meaning. Preference
Charging the marker
A new marker means nothing to the dog. Charging the marker (also called loading it) gives it meaning by pairing it with food before you ask the dog to do anything: say the marker, then give a treat, many times. This is classical conditioning, the kind of learning in which a neutral signal predicts something good and starts to produce the same anticipation. KPCT's glossary defines it, and the contrast with operant conditioning, in which the dog's own behavior changes because of what follows it. The marker is built classically and then used operantly.
You know the marker is charged when the dog reacts to the sound itself: an ear flick, a head turn toward you, a glance at your treat hand, before the treat moves. That is the Done when for this chapter, because it is visible and it does not depend on any cue yet. Test it
Treat size matters here, because you will give many. VCA's training basics page says treats the size of a pencil eraser. Set up to train sets a calorie budget for them and shows how to take part of the dog's daily food for training, so treats do not add weight.
Common timing problems, and how to see them. Ask someone to film a two-minute session on a phone and watch it at half speed. Three patterns show up often:
- The hand moves before the marker. If your hand reaches for the pouch first, the dog learns to watch your hand, and the marker becomes redundant. Keep hands still, mark, then move.
- Late marks. The dog sits, you mark as it rises. Over a session, the dog begins to bounce up out of the sit, because that is what was marked. Mark earlier, at the moment the rear lands.
- Talking through the session. Praise and chatter in between make the marker one sound among many. Be quiet between marks; keep praise for after the treat, or leave it out.
The rate of reward. KPCT's glossary calls a high rate of reinforcement critical for a new behavior. One practitioner's guideline, discussed in Set up to train, is roughly 10 to 15 rewards a minute while a behavior is new. That is the reason sessions are short and treats are small.
3. What the veterinary bodies and the studies say
This manual teaches with rewards only. That is not a style preference. It is the position of the veterinary organizations that write guidance on dog behavior, and it rests on a body of studies. This section gives each document and study with what it found and the limits its authors state, so you can tell the position from the evidence and recognize the alternatives when you meet them.
The position statements
- AVSAB, Humane Dog Training (2021). AVSAB's position statement on humane dog training recommends that only reward-based methods be used for all dog training, including training for behavior problems. It says aversive methods (those that work by force, pain or discomfort) should not be used, and it states that there is no evidence aversive methods are more effective in any context. It lists the tools and techniques to avoid: choke chains, prong collars, shock collars, squirt bottles, shaker cans, compressed air, shouting, staring, "alpha rolls," "dominance downs," leash jerks and flooding. It also notes that recall is the most common reason owners give for using shock collars, which is why Come when called spends time on a reward-based recall and a long line.
- AVSAB board statement (2025). AVSAB's position statements page carries a 2025 board statement. It reaffirms the 2021 statement as AVSAB's current view and says an outdated, since-rescinded AVSAB document suggesting aversive tools may sometimes be appropriate is still circulating. That older document is AVSAB's 2007 statement on punishment, which treated punishment as a last resort rather than ruling it out; it is no longer AVSAB policy and this manual cites it only as history.
- AAHA, Canine and Feline Behavior Management Guidelines (2015). AAHA's guidelines, written for veterinary practices, oppose aversive techniques including prong and choke collars, alpha rolls, dominance downs, shock collars, withholding food and hitting. They say to switch trainers who use prong or shock collars, leash jerks or "dominance," and that aggression or any clinical diagnosis does not go to a trainer for primary treatment. AAHA has published no newer behavior guideline.
- CCPDT, LIMA policy (2021). The Certification Council for Professional Dog Trainers publishes its Least Intrusive, Minimally Aversive (LIMA) policy. It says LIMA does not justify punishment, or electronic, choke or prong collars, in place of positive-reinforcement strategies that would work. It orders interventions in a six-step hierarchy, with health first and positive punishment last. That is what one standards body publishes for its certified trainers; it is a different position from AVSAB's, and the Go deeper block below compares them. Convention
What the statements share: the first question about any behavior is health, positive reinforcement is the first teaching tool, and pain, fear and startle are named as risks. Where they differ is whether positive punishment is ever on the list. AVSAB's answer is no; CCPDT's hierarchy keeps it as a last step. This manual follows AVSAB and AAHA, because they are the veterinary bodies and because the studies below found no case where the aversive method was needed.
The studies
Four studies come up most often. Each one is summarized with its design, because the design decides how far the finding reaches.
| Study | Design | Finding | Limits the authors state |
|---|---|---|---|
| Vieira de Castro et al., 2020 (PLOS ONE) | 92 pet dogs from seven training schools in Porto, grouped by how much the school used aversive methods; sessions filmed, saliva cortisol measured, and a cognitive-bias test outside training | Dogs from the aversive group showed more stress behaviors in class, a larger rise in cortisol after training and more "pessimistic" choices in the bias test | Dogs were not randomly assigned, so the authors say the design does not show cause. The study did not compare how well the dogs learned. |
| China, Mills and Cooper, 2020 (Frontiers in Veterinary Science) | 63 dogs with off-leash problems in three groups of 21: trainers using shock collars; the same trainers without them; trainers focused on rewards. Up to 150 minutes of training over 5 days; videos coded blind | The reward group responded to a single "come" about 82.5 percent of the time against 71.0 and 72.4 percent, with shorter response times. The authors conclude a shock collar is unnecessary for effective recall training. | The reward group had different trainers and was trained about four months later, in spring; results may reflect trainer style |
| Ziv, 2017 (Journal of Veterinary Behavior; Purdue summary page) | A narrative review of 17 studies | Aversive methods can put dogs' physical and mental health at risk; no evidence that positive punishment is more effective than positive reinforcement, and some evidence the opposite is true | The reviewed studies had small samples, missing effect sizes and possible bias in coding behavior |
| Hiby, Rooney and Bradshaw, 2004 (Animal Welfare) | A questionnaire to 364 owners about how they trained seven basic tasks | Owner-rated obedience rose with the number of tasks trained with rewards, not with punishment; the number of problem behaviors rose with the number of tasks trained with punishment | Owner reports, and a correlation: it cannot separate cause from effect (owners of difficult dogs may punish more) |
Read together, the honest summary is this. The studies found no evidence that aversive tools are needed, including for recall, and they found signs that welfare is at risk when they are used. None of them proves cause on its own: the designs are observational or use different trainers, and the authors say so. What this manual takes from them is the conclusion the position statements draw: when a reward-based method was compared, it did at least as well, with fewer stress signs. Convention
No method guarantees a result, and this manual does not promise one. Some problems need a professional working with the dog in person; Read your dog and Proof it and keep going say which and whom.
The dominance model, and why this manual does not use it
You will hear dogs described as "dominant," "stubborn" or "testing you," and owners told to be the "pack leader." That language comes from a model in which dogs misbehave to raise their rank and owners must assert theirs. AVSAB's dominance statement rejects the model as a general guide to behavior: dominance, in behavior science, describes a relationship between two particular animals over a particular resource, not a personality trait, and it does not explain why a pet jumps, pulls or ignores a cue. The statement adds that confrontational techniques aimed at "dominance" frequently provoked aggression in the study it cites, and that what is often called an alpha roll among wolves is a posture the lower-ranking animal offers on its own.
This manual does not use the word "stubborn" for this reason. A dog that does not respond is telling you something specific: the cue is not learned in this place, the reward is not worth it here, something else is more interesting, or the dog is uncomfortable or unwell. Each of those has a fix, and none of them is about rank.
LIMA and AVSAB, compared. CCPDT's hierarchy, in its own order, is: (1) health, nutrition and physical factors, addressed by a veterinarian; (2) antecedents, meaning changing the setting and the triggers; (3) positive reinforcement; (4) differential reinforcement of an alternative behavior, meaning rewarding a replacement and removing the reward for the problem; (5) negative punishment, negative reinforcement or extinction, in no set order; (6) positive punishment. Steps one to four are this manual's whole method. Step one is why every chapter sends sudden changes to the veterinarian first; step two is management; steps three and four are First behaviors, Loose-leash walking, Come when called, Settle, stay and leave it and Everyday problems. The difference with AVSAB is steps five and six. LIMA permits them in the last resort under a qualified professional; AVSAB's 2021 statement and its 2025 board statement say aversive methods should not be used. A trainer who says they "follow LIMA" may still use tools AVSAB and AAHA advise against, which is why the questions for a trainer in Proof it and keep going ask about equipment by name.
Two manuals this manual does not follow. Two 2004 youth dog-project manuals from university extension services, Wisconsin 4-H's "The Basics of Training Your Dog" and Purdue's 4-H-357-W, teach with a training collar and leash jerks. They were standard material twenty years ago and are still online. This manual names them so that a reader who finds them knows why their method differs from this one: the veterinary position statements written since then advise against those techniques.
Reading a study for yourself. Three questions sort most training claims. Were the dogs assigned to methods at random, or did owners choose (and might the owners differ)? Were the same trainers used for each method? Was the outcome measured by someone who did not know which group the dog was in? China, Mills and Cooper used blind coding but different trainers; Vieira de Castro used schools the owners had chosen; Hiby relied on owner reports. Each design has a known weakness, and each paper names it.
4. Management: the other half of training
Management means arranging the dog's surroundings so the unwanted behavior cannot happen, or cannot pay, while you teach something else. A baby gate across the kitchen door, a leash on when the doorbell rings, shoes in a closet, the trash can in a cupboard, blinds closed on the window where the dog barks at passers-by: none of these teach anything, and all of them stop the dog practicing and being rewarded for the behavior you are trying to replace.
Management matters for three reasons:
- Rehearsal builds habits. Each time a dog steals a sandwich from the counter, counter-surfing pays again. Training a "leave it" while the counter is still loaded with food is training against a jackpot. Clear the counter first; train second.
- It protects people and the dog while skills are new. A recall taught last week does not keep a dog out of the road. A long line and a closed gate do (Come when called).
- It is the first step in every problem plan. CCPDT's hierarchy puts changing the setting (antecedents) before any consequence, and the problem plan in Everyday problems opens with a management change.
Management is not failure and not a shortcut. Some management is permanent and sensible: many households keep the trash in a cupboard for the life of the dog. Other management is a bridge until a skill is reliable. Either way, write it down. The household sheet has a management list, and everyone who lives with the dog reads it.
Three management items for most homes: a gate or closed door that keeps the dog out of one room it should not be in unsupervised; something about the front door (leash on before it opens, or a gate in the hall); and one tidying rule (shoes, children's toys, food on counters). A puppy needs more (House training and Crate, pen and alone time); a dog with a concern needs management aimed at it (Read your dog).
5. Age, breed and a new dog
The rules of learning are the same at every age. What changes is how long the dog can concentrate, what it can physically do, and what else is going on in its body and its life.
| Stage (setup's bands) | What changes for training | What does not change |
|---|---|---|
| Puppy (under 6 months) | Short attention; sessions of a few minutes, a few times a day (VCA); the socialization window is open and closing (Socialization and confidence); bladder control is still developing (House training) | Reinforcement, timing and the marker work from the first day home |
| Adolescent (6 to 18 months) | A measured dip in responsiveness to the main carer around 8 months in one study (below); size and strength rise faster than self-control | Cues learned earlier are still there; they need practice in more places |
| Adult (18 months to about 7 years) | A history you may not know; habits that were rewarded for years take longer to replace | Adults learn new cues with the same method |
| Senior (about 7 or older) | Hearing, vision, joints and pain can change what the dog can do; hand signals help a dog whose hearing fades (First behaviors) | Older dogs learn. A senior that stops responding to a cue it knew is a veterinary question first. |
If you chose "I don't know yet" for age in setup, a veterinarian can estimate it from the teeth and body at the first visit. Until then, plan for short sessions and watch for tiredness. Ask the authority
Adolescence
Many owners report that a dog who knew its cues at five months seems to forget them at eight. One study measured this. Asher and colleagues, in "Teenage dogs?" (Biology Letters, 2020), asked the carers and trainers of 285 guide-dog candidates (golden retrievers, Labradors, German shepherds and crosses) to score each dog's trainability at 5, 8 and 12 months. Carers scored trainability lower at 8 months than at 5, and it was higher again by 12. Trainers who knew the dogs less well did not see the dip: their scores rose from 5 to 8 months. In a separate test, dogs were less likely to sit on their carer's cue at 8 months than at 5, but more likely to sit for a stranger.
The authors state their limits: these were guide-dog candidates in a standard program, much of the data came from questionnaires, and the finding needs confirming in other dogs. So this manual takes a narrow lesson from it. A dip in responsiveness around eight months has been measured in one population, it was specific to the main carer, and it passed. It is a reason to keep practicing and lower the difficulty for a while, not a reason to change methods.
Adolescence in more detail. The same study found that separation-related behavior peaked at 8 months and was linked to lower trainability for the carer at that age only. It also found that female dogs with more attention-seeking and attachment behavior at 5 months reached puberty earlier than their breed average. The authors read this as a parallel with human adolescence, where insecure attachment is linked to earlier puberty. That is an interesting hypothesis with one sample behind it; this manual does not build advice on it. In practice, for an adolescent dog: go back a step in difficulty (a closer distance, a quieter place) when a cue falls apart; keep sessions short; keep up the long line for recall (Come when called); and record where the cue works in the cue inventory, so a bad day outside does not erase what the log shows indoors.
Breed
The setup form does not ask for breed. Morrill and colleagues' study in Science (2022; open-access copy) surveyed the owners of 18,385 dogs, 49 percent of them purebred, and sequenced the DNA of 2,155. Breed explained about 9 percent of the variation in behavior between individual dogs, and the authors conclude that breed offers little predictive value for an individual. Some traits were more heritable than others, and breed averages do differ. But knowing the breed tells you little about the dog in your kitchen; watching that dog tells you much more. Test it
What the breed finding does and does not say. The design was owner surveys plus genetics, and the 9 percent figure is about variation across individual dogs. It does not say breed history is meaningless: breed groups were bred for different jobs, and averages differ on some traits. It says the spread within any breed is wide, so a label (from a shelter card or a DNA kit) is a weak guide to how one dog will behave. For training, the practical rule is the same either way: find out what this dog works for and where it struggles, and write it down.
A new dog
A dog in a new home is coping with new people, smells, sounds and routines all at once. San Diego Humane Society's page on helping a dog adjust describes the first two weeks as a decompression period. It suggests keeping a routine, keeping exercise low-intensity, holding off on visitors, dog parks and greetings with other dogs on walks, and waiting until the dog has settled in for two weeks before in-person classes (one week for puppies). It mentions the "3-3-3 rule" (three days, three weeks, three months) as a general guide and says the time can be shorter or longer for an individual. Preference
This manual takes the idea, not the numbers: expect less at first. A new dog that does not eat in a session, startles at the marker, or will not settle is not failing to learn. Keep sessions to a minute, use food the dog takes readily, and let the log show when the dog starts to relax. The adjustment period is a time for lower expectations, not a schedule. If the dog stays frightened or stops eating, that is a question for a veterinarian (Read your dog covers the signs).
6. See it: Rosie's first ten days See it
Known. Priya Raman adopted Rosie, a terrier mix of about 3 years (the shelter's estimate), 22 lb, on 2026-10-02. Priya lives alone in a third-floor city apartment with her cat, Pickle, and has about 10 minutes a day for training. Rosie's history is unknown; the shelter card says "nervous with men" and nothing about bites. Rosie barks at dogs through the window and lunges at them on leash.
What Priya did. On day 3 (2026-10-05) she chose "yes" as her marker and charged it: ten marks and treats from Rosie's kibble, three times that day, each under a minute. By the third session Rosie's ears went up and her head turned toward Priya at "yes." She then played the name game in the living room: 10 tries, 9 turns of the head. Her first sessions row reads 2026-10-05, name game, living room, 2 minutes, 10 repetitions, 9 successes, no distraction, kibble. Her three management items went in the household sheet: a gate between Rosie and Pickle whenever Priya cannot watch both; the leash on before the apartment door opens; blinds closed on the street window.
Assumed. Rosie's age, about 3 years, is the shelter's estimate, so Priya plans as for an adult. The veterinarian can estimate age at Rosie's first check, and that settles it. Priya also assumes kibble is worth working for indoors; the log tests that, and Set up to train tests higher-value food for harder places.
Open. Whether Rosie is house trained: two weeks of the potty log settles it (by day 10 she has had one accident, after a 7-hour absence). Whether the window barking and lunging are fear or frustration: provisional for now, recorded in the observations sheet from Read your dog, and settled by a credentialed behavior consultant. Whether the cat and dog can share a room unsupervised: unresolved, so the gate stays. Rosie's bite history is unknown, and this manual treats "not sure" as "yes" for safety: management, then a veterinarian, then a behavior professional, before any work on the lunging.
What this shows. In ten days with an unknown dog, Priya did not try to fix the lunging. She built the marker, captured one easy behavior, and managed the problems so they would not be rehearsed. Under these assumptions, that is the whole of this chapter for her.
- Karen Pryor Clicker Training: fifteen tips for getting started: short practical rules on marker timing and session length, from a practitioner. This manual uses its timing rules.
- VCA: using reinforcement and rewards: a veterinary hospital group's owner page on timing, reward schedules and fading food, the basis for this chapter's first section.
- AVSAB: Humane Dog Training position statement (PDF): the full statement with its references, if you want to read the evidence list yourself.
7. Practice: charge the marker and start the log Practice
Minimum: three sessions of about a minute charging the marker, one round of the name game, three management items written, and one row in the log. Time: about 15 minutes across one day, plus 5 minutes of writing. Materials: about 30 small treats (pieces the size of a pencil eraser, or part of the dog's own kibble), a marker word or a clicker, a pouch or a bowl on a shelf, and the training log.
- Set up. Pick a quiet room the dog knows. Put the treats where you can reach them but the dog cannot (a pouch, or a bowl on a shelf). Have the dog nearby; it does not need to do anything.
- Charge the marker. Say your marker once, in a normal voice, then hand over one treat within a second or two. Wait a few seconds; repeat. Ten times, then stop. Keep your treat hand still until after the marker, so the sound, not your hand, predicts the food.
- Repeat twice more today. Two more sessions of ten, a few hours apart. In the third session, say the marker while the dog is looking away. If the head turns toward you or your treat hand before you move, the marker is charged. If not, do another session tomorrow and check again. Test it
- Play the name game. With the dog a few feet away and not looking at you, say its name once. The moment it looks at you, mark and treat. If it does not look within a couple of seconds, make a small sound to help, and mark the look. Ten tries is plenty. Count how many times it turned on the name alone.
- Write three management items. Walk through the home with the question "what does the dog do here that I do not want it practicing?" Choose three and write the fix for each in the household sheet's management list.
- Log it. Add a row to the sessions sheet: today's date, skill "Name game," place, minutes, attempts (10) and successful responses (how many of the 10 were clean). The sheet's percent is successful responses divided by attempts. In notes, write one thing to change next time.
- Capture one behavior. Once the marker is charged, sit with the treats and wait. Each time the dog sits (or lies down) on its own, mark and treat. Do not say "sit." After a few sessions, notice whether the dog sits more often near you. That is reinforcement you can see, and it is the start of First behaviors.
- Film a session. Ask someone to film two minutes on a phone. Watch it at half speed for the three timing problems in the timing section: hands moving first, late marks, chatter.
- Try the name game in a second room. Then a third. Write each in the log as its own row. If the success count drops, the name is not yet learned there, which is normal; Proof it and keep going calls this generalization.
8. Your dog Your dog
9. Check your work Check your work
Criteria
- Does your dog orient to the marker (a head turn, an ear flick, a look at your hand) before your hand moves?
- Did every mark get a treat, including any you made by mistake?
- Can you say in one sentence, in your own words, what reinforcement is? (A good answer mentions that the dog's next behavior, not your intention, shows whether something reinforced.)
- Are your three management items specific enough that someone else in the house could follow them?
- Does the log have one row with a date, a skill, minutes and a count?
Worked example
Priya's work, checked (fictional example). Rosie turned her head at "yes" before any hand movement in the third session: the marker passes. Every mark got a piece of kibble. Priya's sentence: "Whatever Rosie gets right after she does something makes her more likely to do it again, so I have to watch what I'm paying for, including barking." It passes, and the last clause is the point of the chapter. Her management items name the place and the fix (gate between Rosie and Pickle when unsupervised; leash on before the door opens; blinds closed on the street window). The log row is complete. What her work does not settle: whether the lunging is fear or frustration, and whether Rosie can be trusted with the cat. Both are open, and both stay managed.
Common mistakes
- Reaching for the treat before the marker. The dog watches your hand instead of listening. Check by keeping your hand behind your back until after the sound.
- Using a word you say all day. "Good" and "OK" are spoken constantly and lose meaning. If the dog does not orient after three sessions, try a rarer word or a clicker.
- Repeating the name. "Rosie, Rosie, ROSIE" teaches that the first two do not count. Say it once; help with a small sound if needed; mark the look.
- Sessions that run long. A dog that wanders off, sniffs the floor or stops taking food has had enough or is uncomfortable. End earlier next time, and see Read your dog for the signals.
- Writing management as wishes. "Stop the barking" is a goal; "blinds closed on the street window" is management.
What is still unresolved
This chapter cannot tell you whether your dog's existing problems come from learning, fear, pain or something else; Read your dog teaches the signals that start to answer that, and the veterinarian settles anything medical. It cannot tell you what your dog finds most rewarding in harder places; Set up to train tests that. And the marker, charged at home, may need recharging somewhere new.