Negative reinforcement (R–) means taking something away that increases the frequency of the behavior. An example would be pushing on the dog’s bottom until they sit, at which point you let go. Assuming that the dog sits more often, the behavior of sitting is reinforced by removing the pressure on the dog’s bottom. And negative punishment (P–) means taking something away that makes the behavior decrease in frequency. For example, your dog jumps on you and you turn away from them or even leave the room for thirty seconds every time. You are taking your attention away and the dog is less likely to jump up in future (but remember what I said earlier about extinction bursts!).
Examples of operant conditioning: reward-based training uses R+ and P–
| ANTECEDENT | BEHAVIOR | CONSEQUENCE | RESULT |
| You say “sit” | The dog sits | R+ Something good happens e.g., chicken, cheese, or treats; a quick game of tug; attention such as petting | The behavior happens more often |
| You come home | The dog jumps on you | P– Something good is taken away e.g., the flow of chicken, cheese, or treats stops; the game of tug ends; the person stops giving attention or leaves the room | The behavior happens less often |
| You greet the dog | The dog jumps on you | P+ Something bad happens e.g., a tug on the leash, pressure to push the dog’s bottom down, a zap from the shock collar | The behavior happens less often |
| You say “sit” while tugging the leash, pushing the dog’s rear end, or applying the shock collar | The dog sits | R– Something bad is taken away or stops e.g., tugging on the leash stops, pressure is no longer applied to the dog’s rear end, the shock from the collar stops | The behavior happens more often |
The table has examples of reinforcement and punishment. Note the consequence has to have an effect on behavior. For example, if you pet the dog intending it to be positive reinforcement but it has no effect on the dog’s behavior, then the petting wasn’t actually reinforcing to the dog.
Consequences are not the only way to change behaviors; we can also change the antecedents, something dog trainers call antecedent arrangements. For example, suppose the dog has a habit of drinking from the toilet bowl. The antecedent is that the lid is up allowing access to the toilet water. A very sensible antecedent arrangement would be to ensure the lid is never left up, so it is not possible for the dog to drink the toilet water. Of course, you should also ensure the dog has access to a suitable water supply!
DOG TRAINING: THE LINK BETWEEN TRAINING METHODS AND BEHAVIOR
REWARD-BASED METHODS ARE those that use positive reinforcement (R+) and/or negative punishment (P–), or humane management strategies (such as putting a lid on the garbage can to keep dogs from raiding the trash, or using a no-pull harness for dogs that pull on-leash).6 Exercise and enrichment are also often part of the solution to resolving behavior problems (see chapters 9 and 10).
Eighty-eight percent of dog owners do at least some training at home, according to a report in the Journal of Veterinary Behavior, but it seems that most do not use reward-based methods exclusively.7 Unfortunately, when people use outdated methods to train dogs, perhaps because they don’t realize science recommends reward-based methods, they are using methods that rely on fear and pain. It’s just a tap, a correction, or information, they say. But prong collars, choke collars, leash corrections, electronic collars, and alpha rolls (rolling the dog on their back and holding them there until they stop moving) work because they are painful or fearful for the dog. These are aversive methods.
A survey published in the Journal of Veterinary Behavior asked owners about their dog training methods and attendance at dog obedience classes.8 The owners were then asked to look at a list of thirty-six possible dog behavior problems, including attention-seeking issues (e.g., jumping up, pawing, or mouthing the owner), fear issues (e.g., avoiding or hiding from familiar or unfamiliar people), and aggression, and indicate which one(s) their dog exhibited. Seventy-eight percent of dogs jumped up at people, 75 percent pawed at people or demanded attention, and 74 percent were excitable with visitors. These are all friendly, pro-social behaviors (at least in the eyes of the dog!). The three most common behavioral issues people described as problematic were aggression towards family members, house soiling when the owner was at home, and chewing or destroying things when the owner was out. Owners who used only positive reinforcement in training were less likely to report behavior problems related to fear, aggression, and attention seeking. Interestingly, the highest levels of fear, aggression, and attention seeking were found in dogs whose owners used both positive reinforcement and positive punishment (so-called “balanced” dog training methods).
A study in Vienna published in Applied Animal Behaviour Science looked at whether the size of the dog made any difference.9 Dog owners in that city are required to register their dogs, and researchers sent a questionnaire to a random sample of owners, which means the results of the survey are representative of the population there. The study rated a dog as small (up to 20 kg or 44 pounds in weight) or large. Eighty percent of owners used punishment to train their dog, most commonly leash jerking, scolding, and holding the dog’s muzzle. Ninety percent of owners used rewards either often or very often. For both small and large dogs, the more often their owners used punishment, the more aggressive and more excitable the dog. The relationship was strongest for small dogs. In contrast, the more often people used rewards, the more obedient they rated their dog, and also less aggressive and less excitable. Another finding of note is that owners of small dogs are less consistent with their training, put less emphasis on training, and engage in fewer activities with their dog than those who have large dogs. And consistency matters when it comes to obedience: the less consistent the owner, the less obedient their dog.
In another study published in Applied Animal Behaviour Science fifty-three dog owners were asked about how they had trained their dog and video was taken of them asking their dog to sit, lie down, and stay.10 Researchers then gave the owner a bag of treats and a ball to use as rewards if they wished, and gave them five minutes to teach their dog a novel task—touching one of two spoons on command. All of the participants had used a mix of rewards and punishment to train their dog in the past. If the owners had tended to use punishment more often than rewards, the dogs were less playful with the owner and less interactive with the researcher. The dogs whose owners had previously used rewards more were quicker to learn the new task. Dogs also performed better at learning the new task if their owners were patient and used more rewards. The most likely reason for the improved results is motivation.
Another study, published in the Journal of Veterinary Behavior, observed dogs at two different dog training schools, one that used positive reinforcement and another that used negative reinforcement.11 Dogs in the negative reinforcement group showed more signs of stress, such as a lowered body posture (keeping their body closer to the ground), whereas dogs in the positive reinforcement group looked at their owners a lot more. This matters because you need the dog’s attention to ask them to do something. So positive reinforcement was not just better for the dog’s welfare but also for the human–canine bond.