Operant conditioning

What it says

Classical conditioning is about signals. Operant conditioning is about consequences: what follows a behavior changes how often the behavior comes back.

There are four boxes, and the words are easy to get wrong. Reinforcement makes a behavior more likely, punishment makes it less. Positive means something is added, negative means something is taken away. So negative reinforcement is not punishment: it is relief, and relief is one of the strongest teachers there is.

Two more ideas carry most of the practical weight. Shaping builds a behavior nobody has yet by reinforcing closer and closer tries. Schedules decide how stubborn it becomes. A reward that comes after an unpredictable number of tries produces the most persistent behavior of all. It is the pattern behind both a slot machine and a person who keeps checking a phone.

The four boxes

Read each one as what happened right after the behavior.

1. Positive reinforcement

   Something good is added. A dog sits and gets a treat, so it sits again.

2. Negative reinforcement

   Something unpleasant is taken away. The seatbelt chime stops when you buckle, so you buckle sooner next time.

3. Positive punishment

   Something unpleasant is added. A hot pan stings, so you use the handle.

4. Negative punishment

   Something good is taken away. The phone goes in a drawer, so the behavior that cost it drops off.

Where it came from

Thorndike put cats in puzzle boxes in the 1890s and timed how long they took to get out. They got faster, and he concluded that acts followed by satisfaction are more likely to be repeated. That is the law of effect, published in Animal Intelligence in 1911.

Skinner built the apparatus that made the field measurable and set out the account in The Behavior of Organisms in 1938. With Charles Ferster he published Schedules of Reinforcement in 1957, which is still the reference for how timing changes behavior.

What holds up, and what does not

The schedule findings replicate and they are used every day in animal training and in programs built around clear rewards. The four boxes are a genuinely useful tool for reading your own habits.

Two limits are well established. Punishment suppresses a behavior without teaching anything to do instead, so it tends to need repeating. And a reward can crowd out interest in a task somebody already enjoys. A large review in 1999 looked at tangible rewards offered for a task people already found interesting. Once the reward stopped, they spent less time on it (Deci, Koestner and Ryan).

What it does not mean

Reinforcement is defined by what happens to the behavior, not by what the giver meant. Attention given to stop a tantrum can reinforce the tantrum. That is why the honest test is always the rate of the behavior afterward, not anyone's intention.

Check yourself

Three questions on what is above. Each answer says why.

1. A chime sounds until a driver buckles the seatbelt, and buckling becomes quicker over the following week. Which box is that?

   1. Positive punishment
   2. Negative reinforcement
   3. Negative punishment
   Show the answer

   Answer: B. Something unpleasant was removed and the behavior went up, so it is negative reinforcement. Reinforcement is read off the rate of the behavior.

2. Which schedule produces the most persistent behavior?

   1. A reward every single time
   2. A reward after an unpredictable number of tries
   3. A reward once a day, at the same time
   Show the answer

   Answer: B. An unpredictable number of tries before the reward is the hardest pattern to give up, because no single failure tells you anything.

3. Pick the statement the 1999 review supports.

   1. Rewards always increase interest in a task
   2. A tangible reward for an already interesting task can reduce time spent on it later
   3. Punishment teaches a replacement behavior
   Show the answer

   Answer: B. Deci, Koestner and Ryan found that tangible rewards offered for an interesting task cut the time people spent on it once the reward stopped.

Sources

1. Thorndike, E. L. (1911). Animal Intelligence: Experimental Studies. Macmillan.
2. Skinner, B. F. (1938). The Behavior of Organisms: An Experimental Analysis. Appleton Century Crofts.
3. Ferster, C. B., and Skinner, B. F. (1957). Schedules of Reinforcement. Appleton Century Crofts.
4. Deci, E. L., Koestner, R., and Ryan, R. M. (1999). A meta analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6).

Practice it on Learn

More theories

• Classical conditioning

  Classical conditioning explained: the four terms, what Pavlov actually did, acquisition and extinction, and why pairing two events is not enough on its own.

• Learned helplessness

  Learned helplessness: the 1967 experiment, the attribution version, and the 2016 reformulation that turned the original claim around.

• The Little Albert experiment

  The Little Albert experiment: what Watson and Rayner did in 1920, what the study really showed, the ethics, and what happened to the child.