Punishment · The hub
Punishment
The other half of operant conditioning: making behavior less likely. This page defines punishment precisely, separates its two types from extinction and from negative reinforcement, and gives an honest reading of whether it works.
Definition
Punishment is the process in which a consequence that follows a behavior makes that behavior less likely in the future. The consequence is called a punisher. As with reinforcement, the definition is functional: a consequence that does not reduce the behavior is not punishment, however unpleasant it was meant to be.[1]
In everyday speech "punishment" means a penalty someone intended. In behavior analysis it means a consequence that worked.
In brief
- Punishment is any consequence that makes the behavior it follows less likely; a consequence that does not reduce the behavior is not punishment.
- Positive punishment adds a stimulus, negative punishment removes one; both decrease behavior, unlike negative reinforcement, which increases it.
- It works only when immediate, consistent, intense from the outset, and paired with a reinforced alternative — conditions rarely met outside a laboratory.
The two types of punishment
Positive punishment
A stimulus is added after the behavior and the behavior decreases. You touch the hot pan and it burns; you speed and get a ticket.
Definition, examples, side effects ›Negative punishment
A stimulus is removed after the behavior and the behavior decreases. A teenager misses curfew and loses the car keys; a player fouls and sits out.
Time-out, response cost, and doing them right ›Three things punishment is not
- It is not negative reinforcement. Negative reinforcement increases a behavior by removing something aversive. Punishment decreases behavior. The word "negative" is shared; the direction is opposite. Why students mix these up ›
- It is not extinction. In extinction nothing is added or removed as a consequence; the reinforcer that used to follow the behavior simply stops arriving. Ignoring a tantrum is extinction (if attention was the reinforcer); taking away the tablet for the tantrum is negative punishment. Extinction ›
- It is not defined by intent. A parent who "punishes" a child by yelling, and finds the behavior increasing, has reinforced it — attention was the reinforcer. The behavior, not the parent, decides what the consequence was.
Does punishment work?
Skinner thought not. An early experiment by his student W. K. Estes found that punishing rats' lever pressing suppressed it only temporarily; when the punishment stopped, the pressing returned, and the total number of responses to extinction was about the same as for unpunished rats.[2] Skinner concluded that punishment merely suppresses, and argued against it for the rest of his career.
Later research complicated the picture. Azrin and Holz's 1966 review, still the reference work, found that punishment can produce large and lasting decreases when it is immediate, delivered every time, intense enough from the outset (rather than escalated gradually), and combined with reinforcement of an alternative behavior.[3] A review of the applied literature three decades later reached the same conclusion for clinical settings.[4] Punishment, in other words, works — under conditions that are rarely met outside a laboratory, and at a price.
The side effects
- It teaches nothing new. Punishment says what not to do. The gap is filled by whatever else the person can do, which may be worse. Reinforcement of an alternative fills the gap deliberately.
- Escape and avoidance. The punished organism learns to avoid the punisher as much as the behavior — the child learns not to get caught, the employee stops reporting mistakes. Avoidance learning ›
- Emotional and aggressive responding. Aversive stimulation produces fear and aggression, sometimes toward bystanders.[3]
- It models aversive control. People treated with punishment use punishment; the punisher's own behavior is negatively reinforced by the brief pause in the problem, which is why punishing tends to escalate.
- Corporal punishment specifically. A 2016 meta-analysis of 75 studies covering more than 160,000 children found spanking associated with worse, not better, outcomes — more aggression, more antisocial behavior, more mental-health problems — with no evidence of benefit.[5]
What to do instead
| Alternative | How it works | Where it is explained |
|---|---|---|
| Differential reinforcement of an alternative | Reinforce a behavior that does the same job for the person; the problem behavior loses its function | Glossary: DRA |
| Extinction | Identify the reinforcer maintaining the behavior and stop it arriving; expect a burst first | Extinction |
| Antecedent change | Remove the cue, add a prompt, change the setting so the behavior is never triggered | The ABC model |
| Time-out done correctly | Brief, calm removal from reinforcement — which only works if the situation left was reinforcing | Negative punishment |
| Response cost | A predictable, proportionate loss of a token or privilege inside a system that also pays for good behavior | Negative punishment |
Modern applied behavior analysis treats reinforcement-based procedures as the default and reserves punishment for narrow, supervised cases where reinforcement alone has failed and the behavior is dangerous, with consent and the least restrictive procedure as ethical requirements.[6]
Key takeaways
- In behavior analysis, punishment means a consequence that worked: the behavior became less likely. The behavior, not the punisher's intent, decides what the consequence was.
- Punishment is not negative reinforcement, which increases behavior by removing something aversive, and it is not extinction, in which the reinforcer simply stops arriving. Ignoring a tantrum is extinction if attention was the reinforcer; taking away the tablet for it is negative punishment.
- Punishment can produce large, lasting decreases when it is immediate, delivered every time, intense enough from the outset, and combined with reinforcement of an alternative. Those conditions are rarely met outside a laboratory.
- Its side effects: it teaches nothing new, it produces escape and avoidance of the punisher, it elicits fear and aggression, and it models aversive control. Spanking is associated with worse outcomes and no evidence of benefit.
- Reinforcement-based procedures are the default. Differential reinforcement of an alternative, extinction, antecedent change, and time-out or response cost done correctly come first; punishment is reserved for narrow, supervised cases where the behavior is dangerous.
Check yourself
A parent yells at a child for whining, and the whining increases over the following weeks. What kind of consequence was the yelling?
A reinforcer. Punishment is not defined by intent: the behavior, not the parent, decides what the consequence was. Attention was added and the behavior increased, so the yelling was positive reinforcement.
A child throws tantrums for attention. One parent stops responding to tantrums entirely; the other takes away the tablet each time. Which one is punishment?
Taking the tablet is negative punishment: a stimulus is removed as a consequence. Ignoring the tantrum is extinction, because nothing is added or removed; the reinforcer that used to follow the behavior simply stops arriving. Expect a burst before extinction works.
A manager waits until the monthly review to reprimand an employee for skipping safety steps, and mentions it only sometimes. Why is this unlikely to reduce the behavior?
Punishment produces lasting decreases only when it is immediate and delivered every time. A delayed, occasional reprimand meets neither condition, and it teaches nothing about what to do instead.
A teenager loses her phone for a day each time she misses curfew, and missed curfews decrease. A friend calls this negative reinforcement, because something was removed. What is wrong with that?
The direction. Negative reinforcement increases a behavior by removing something aversive; here a behavior decreased when something wanted was removed, which is negative punishment. The two share the word "negative" and nothing else.
Explain it to a friend. Explain why behavior analysts reach for reinforcement before punishment, using an example from your own life in which a penalty failed to change what someone did.
Go deeper
Frequently asked questions
What is punishment in psychology?
Any consequence that makes the behavior it follows less likely in the future. Positive punishment adds a stimulus (a burn, a fine); negative punishment removes one (losing privileges). It is defined by its effect on behavior, not by anyone's intention.
Is negative reinforcement a type of punishment?
No. Negative reinforcement increases behavior by removing something unpleasant; punishment decreases behavior. They share the word "negative" and nothing else.
Does punishment work?
It can reduce behavior when it is immediate, consistent, sufficiently intense from the start, and paired with reinforcement of an alternative. Those conditions are rarely met in daily life, and punishment carries side effects — avoidance, aggression, no new learning — which is why behavior analysts use reinforcement first.
What is the difference between punishment and extinction?
Punishment adds or removes a stimulus as a consequence. Extinction changes nothing except that the reinforcer stops coming. Both reduce behavior; extinction usually produces a temporary burst first.
References
- Skinner, B. F. (1953). Science and Human Behavior. Macmillan.
- Estes, W. K. (1944). An experimental study of punishment. Psychological Monographs, 57(3), 1–40.
- Azrin, N. H., & Holz, W. C. (1966). Punishment. In W. K. Honig (Ed.), Operant Behavior: Areas of Research and Application (pp. 380–447). Appleton-Century-Crofts.
- Lerman, D. C., & Vorndran, C. M. (2002). On the status of knowledge for using punishment: Implications for treating behavior disorders. Journal of Applied Behavior Analysis, 35(4), 431–464.
- Gershoff, E. T., & Grogan-Kaylor, A. (2016). Spanking and child outcomes: Old controversies and new meta-analyses. Journal of Family Psychology, 30(4), 453–469.
- Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied Behavior Analysis (3rd ed.). Pearson.