A form of learning in which the consequences of a behaviour change how often the behaviour occurs in the future. Behaviours followed by something the animal finds rewarding tend to be repeated. Behaviours followed by something unpleasant tend to drop away.

Operant conditioning is one of the major frameworks within learning theory and underlies most of what we currently understand about how training works across species. The framework was developed primarily by B.F. Skinner in the 1930s and has been refined and extended substantially since.

The framework identifies four basic consequences that can follow a behaviour, often described as the four quadrants: positive reinforcement (something pleasant added, behaviour increases), negative reinforcement (something unpleasant removed, behaviour increases), positive punishment (something unpleasant added, behaviour decreases), and negative punishment (something pleasant removed, behaviour decreases).

Operant conditioning is the mechanism that underlies most practical training, from teaching a horse to move off the leg, to teaching a parrot to step up onto a hand, to teaching a dog to sit on cue, to teaching a zoo animal to present for a voluntary blood draw. It is distinct from classical conditioning, which is about learned associations between stimuli rather than between behaviour and consequence. The two run continuously and simultaneously in every training interaction, and modern training practice attends to both.

The behavioural science literature consistently finds that training built primarily on reinforcement (positive or negative, applied with good timing) produces better welfare and behavioural outcomes than training built primarily on punishment. Most modern applied training across species reflects this consensus.