Edward Thorndike (1898) was one of the first researchers to study what was later to be called "operant conditioning". Thorndike put hungry cats in puzzle boxes and put pieces of tasty fish outside them to motivate them to escape. To escape, the cats had to push a lever and pull on a string. When first put in the box, the cats would behave in a fairly random way, pacing up and down, scratching at the sides etc., but eventually, by chance, they would push the lever and pull on the string in the correct sequence. They would then be rewarded by being able to eat the fish.

Thorndike found that the cats would become faster and faster at escaping, while never showing any insight into how the puzzle box was solved. There was no sudden drop in escape time, which would be expected if the cat had solved the puzzle mentally or realised how its actions led to escape; this provides support for the behaviourist view that only behaviours are important and not cognitions: the escape behaviour was simply strengthened by the reward of a piece of fish. From this Thorndike developed his 'law of effect', that behaviours that have a pleasant outcome are stamped in and those that have unpleasant consequences are stamped out.

Click here for a video of Thorndike's experiment

Burrhus F. Skinner (1938) Elaborated on Thorndike’s “law of effect”. He criticised Thorndike’s use of “Stamps out” and “Stamps in” as being “mentalist”. He said the terms “Strengthen” and “Weaken” should be used instead: Responses that have positive outcomes are strengthened and responses that have negative outcomes are weakened.

Skinner was very critical of "mentalism". Mentalism is when inferences are made about what an organism is thinking. Skinner, like other behaviourists such as John B. Watson believed that since mental events, such as thoughts cannot be observed directly they should not be studied. Instead, according to Skinner, psychologists should study only antecedents of behaviour, the behaviours themselves and the consequences of the behaviour. This is the ABC of operant conditioning.

The Skinner Box

Skinner invented the “Skinner Box” as an alternative to Thorndike’s puzzle box. One reason for this was because it allowed tighter control of extraneous variables by automating the experimental process (e.g., rewards, such as food pellets or water could be dispensed and punishments (electric shocks) could be given. It also recorded the animals’ responses eliminating any experimenter bias. In addition, it allowed Rats or Pigeons to do things instead of just escaping; for example, rats could pull a lever or pigeons could peck a disc.

Reinforcement and Punishment

In Operant Conditioning, reinforcements increase the likelihood of behaviour being repeated. Punishments decrease the likelihood of it being repeated. Whether an outcome is a reinforcement, punishment or neither can only be judged after observing whether the frequency of the behaviour increases, decreases or does not change.

Positive and Negative Reinforcements and Punishments

Reinforcements and punishments can be either positive or negative (this is often a source of confusion). Positive and negative do not refer to whether the outcome is good or bad, but to whether the outcome involves presenting something (positive) or taking something away (negative).

Positive reinforcements and punishments produce their effects by being presented

Negative reinforcements and punishments produce their effects by being removed


Here are some examples:

Note: The outcomes in the table are only correct if they have the expected effect on the behaviour.

Click here for classical conditioning