Positive reinforcement, explained

Positive reinforcement is a consequence added straight after a behavior that makes the behavior more likely to happen again. Positive means added, not pleasant, and reinforcement means the behavior increases; both words describe what happens, not what anyone intended. If a dog gets a biscuit for sitting and sits no more often afterward, the biscuit reinforced nothing, however kindly it was meant.

The opposite case makes the idea clearer. Negative reinforcement also strengthens a behavior, but by taking something unpleasant away, such as easing leash pressure the moment a dog steps back beside you, so the unpleasant thing has to be there first. Positive reinforcement needs nothing unpleasant at all. A handout on the method hosted by AVSAB, the American Veterinary Society of Animal Behavior, describes the reward as payment for a behavior the trainer wants repeated.

Every Dogtown program, from Tiny Tails Puppy School to both board and train stays, is fear- and force-free and built on positive reinforcement, led by Gabriella Pope. The first days of a stay include finding out what a dog will work for, because value is personal and it moves. A dog that takes cheese eagerly in a quiet room may ignore it beside the turf yard, and a worried dog may refuse food altogether, which tells the trainer the task is too hard, not that the dog is stubborn.

Value is then matched to difficulty. The AVSAB handout advises the highest-value reward, given generously and quickly, for hard or emotionally difficult work; we keep ordinary pay for easy work, so the best rewards keep their meaning. Delivery follows one order: the marker lands as the behavior happens, the reward follows within a second or two, and, as the handout recommends, a cue is added once the dog offers the behavior around nine times in ten.

Researchers have measured preference directly by letting dogs choose. In a 2015 study, Erica Feuerbacher and Clive Wynne compared petting with vocal praise in shelter dogs and owned dogs, with strangers and with owners. Every group preferred petting, the dogs showed no sign of tiring of it across eight sessions, and a person offering only praise drew about as little approach as a person offering nothing. The authors concluded that praise probably has to be conditioned before it works as a reward.

The method shapes the relationship too. AVSAB's 2021 position statement cites studies in which dogs trained with rewards looked toward their owners more often during training, and in which dogs from reward-based schools played more in their owner's presence than a stranger's, and followed and greeted their owners more. It recommends reward-based training for all dogs and finds no evidence that aversive methods do better in any context.