Skip to content
Psych Dictionary

Operant conditioning

  • Concept · Learning theory
  • Strong evidence
  • Published October 18, 2026
  • 17 sources
  • 7 min read

In short

Operant conditioning is learning in which the consequences of a behavior change how often it occurs. Behavior followed by reinforcement becomes more frequent and behavior followed by punishment less so. B. F. Skinner developed its experimental analysis; Edward Thorndike’s law of effect was its precursor.

Key facts

Field
Learning theory
Originators
B. F. Skinner
First described
1898–1938, United States
Also known as
Instrumental conditioning; Type R conditioning

State of the evidence

Strong evidence

The effects of reinforcement, extinction and schedules on response rate were established in laboratory animals by Skinner and his colleagues and have been reproduced and refined since, including a review of fifty years of research on pausing under ratio schedules[1][2][3]. As a treatment method, contingency management for substance use disorders produced a mean effect of d = 0.42 across 47 treatment–control comparisons[4]. Several parts of the theory are contested: whether positive and negative reinforcement are distinct processes[5], how lasting the effects of punishment are[6][7], and how far conditioning can override a species’ instinctive behavior[8].

Evidence labels summarise the research literature; they are not recommendations. How labels are assigned

Operant conditioning is learning in which the consequences of a behavior change the probability that the behavior will occur again. A rat whose lever presses produce food presses more often; if the presses stop producing food, pressing eventually falls off[6]. B. F. Skinner, who named the process[9], defined an operant as behavior that is emitted by the organism rather than elicited by an identifiable stimulus, and chose a term that refers to an event following the behavior[1].

The contrast with classical conditioning is in the arrangement. In classical conditioning a reinforcing stimulus is paired with another stimulus; in operant conditioning it is contingent on a response[6]. A dog that salivates to a metronome has learned that one event predicts another. A rat that presses a lever for food has learned that its own behavior produces an outcome. The comparison of the two sets out the differences point by point.

Negative reinforcement is not punishment. Reinforcement of either kind makes a behavior more likely; punishment of either kind is defined by the opposite operations[6]. The section on the four contingencies treats the distinction directly, and a separate comparison sets negative reinforcement against punishment in detail.

Thorndike’s law of effect

Skinner traced the stamping in of rewarded behavior, which he equated with operant conditioning, to Edward Thorndike’s experiments with cats in puzzle boxes[6]. In Thorndike’s 1898 monograph, hungry cats shut in a box with food visible outside had to claw a loop or press a button to open the door. They hit on the successful movement by accident among many others, and over repeated trials their escape times fell gradually. Thorndike concluded that the successful impulse was “stamped in” by the pleasure that followed and the others “stamped out”, and took the gradual slope of the learning curve as evidence that the cats were not reasoning[10]. Pavlov later named Thorndike’s monograph as the starting point of systematic experimental work of this kind[11].

In 1911 Thorndike stated the principle as the law of effect: “Of several responses made to the same situation, those which are accompanied or closely followed by satisfaction to the animal will, other things being equal, be more firmly connected with the situation, so that, when it recurs, they will be more likely to recur”[12]. Responses followed by discomfort, in the same statement, have their connections with the situation weakened[12].

Skinner’s operant and the measure of response rate

Skinner’s 1938 book The Behavior of Organisms separated two kinds of behavior. Respondent behavior is correlated with specific eliciting stimuli, as a reflex is; operant behavior has no such correlated stimulus and is studied through the rate at which it occurs[1]. He called the use of rate the outstanding characteristic of his method[1]. To record it he built the cumulative recorder, which registered every response as an upward step in a horizontally moving line, so that the slope of the line showed the rate of responding[9].

Skinner defined reinforcers by their effect, not their physical properties. An event is a reinforcer if it strengthens the behavior it follows; this classification is empirical, he argued, and becomes circular only if the effect is then explained by saying that the event is reinforcing[6]. The same logic defines an aversive stimulus as one whose removal is reinforcing[6].

The four contingencies

A consequence can add a stimulus or remove one, and the stimulus can be appetitive or aversive. The combinations give four contingencies.

The four operant contingenciesA two-by-two grid. The columns say whether a consequence adds a stimulus or removes one; the rows say whether that stimulus is appetitive, such as food, or aversive, such as a loud noise. Adding an appetitive stimulus is positive reinforcement and the behavior increases. Removing an appetitive stimulus is negative punishment and the behavior decreases. Adding an aversive stimulus is positive punishment and the behavior decreases. Removing an aversive stimulus is negative reinforcement and the behavior increases. The two reinforcement cells lie on one diagonal and both increase behavior; positive and negative mean added and removed, not good and bad.Stimulus addedStimulus removedAppetitivestimuluse.g. foodPositivereinforcement▲ Behavior increasesFood is deliveredafter a lever pressNegativepunishment▼ Behavior decreasesFood is taken awayafter the responseAversivestimuluse.g. loud noisePositivepunishment▼ Behavior decreasesA loud noise startsafter the responseNegativereinforcement▲ Behavior increasesA loud noise stopsafter a lever pressHighlighted cells are reinforcement: the behavior becomes more frequent.“Positive” and “negative” mean added and removed, not good and bad.

Figure 1Reinforcement increases the behavior it follows and punishment decreases it; “positive” and “negative” say whether a stimulus is added or removed. After Skinner (1953).

Text description

A two-by-two grid. The columns say whether a consequence adds a stimulus or removes one; the rows say whether that stimulus is appetitive, such as food, or aversive, such as a loud noise. Adding an appetitive stimulus is positive reinforcement, and the behavior increases. Removing an appetitive stimulus is negative punishment, and the behavior decreases. Adding an aversive stimulus is positive punishment, and the behavior decreases. Removing an aversive stimulus is negative reinforcement, and the behavior increases. The two reinforcement cells lie on one diagonal. A note states that positive and negative mean added and removed, not good and bad.

Skinner called a reinforcer positive when the reinforcing event adds something, such as food or water, and negative when it removes something, such as a loud noise, a very bright light, extreme cold or electric shock; in both cases the probability of the response increases[6]. Punishment is the opposite pair of operations: the removal of a positive reinforcer or the presentation of a negative one[6]. The diagram labels these two forms negative and positive punishment, on the same added-or-removed logic. Behavior that ends an aversive stimulus is escape, and Skinner’s examples are ordinary: putting fingers in the ears or closing a window against a loud noise[6].

The positive–negative distinction is itself contested within behavior analysis. Michael argued in 1975 that it is ambiguous, because adding a stimulus requires its previous absence and removing one requires its previous presence, and that no behavioral or physiological evidence shows two distinct processes[5]. Alan Baron and Mark Galizio reconsidered the question thirty years later and found no new evidence that allows an event to be classified reliably as one kind of reinforcer rather than the other[5].

Shaping by successive approximation

Shaping produces a behavior that the organism would rarely or never emit by reinforcing closer and closer approximations to it. Skinner described the procedure with a pigeon: food is given first for a slight turn toward a spot on the wall, then only for movements toward it, then only when the head moves slightly forward, and finally only when the beak touches the spot; a hungry bird can usually be brought to peck the spot in two or three minutes[6]. He compared operant conditioning to a sculptor shaping a lump of clay, in which no point in the process produces anything very different from what preceded it[6].

The method was described in The Behavior of Organisms, but according to Gail Peterson’s historical account Skinner did not shape a response by hand until 1943, on the top floor of a flour mill in Minneapolis; the experience led him to coin the term shaping[13].

Reinforcement does not require that the behavior actually produce the reinforcer. In Skinner’s “superstition” experiment, food was delivered to hungry pigeons at regular intervals regardless of what they did, and six of eight birds developed stereotyped responses, such as turning counter-clockwise or thrusting the head into a corner of the cage. Skinner attributed them to the accidental coincidence of a response with food[14].

Schedules of reinforcement

Reinforcement can follow every response or only some of them. Charles Ferster and Skinner reported schedule performances at length in Schedules of Reinforcement (1957), the book a later review treats as the starting point of research on pausing under ratio schedules[2]. Schedules are defined by whether reinforcement depends on a number of responses (ratio) or on the passage of time (interval), and whether the requirement is fixed or varies around an average.

Four basic schedules of reinforcementA two-by-two grid. Columns are fixed and variable schedules; rows are ratio schedules, which count responses, and interval schedules, which count time. Fixed ratio: a reinforcer follows a set number of responses, and the typical pattern is a pause after each reinforcer followed by rapid responding, called break and run. Variable ratio: the number of responses required varies around an average, and responding is fast and steady with few long pauses. Fixed interval: the first response after a set time is reinforced, and responding is sparse early in the interval and accelerates toward its end, called the scallop. Variable interval: the first response after a varying time is reinforced, and responding occurs at a relatively constant rate that is lower than on ratio schedules. A note states that these are typical patterns rather than invariant ones: pausing also occurs on variable-ratio schedules, and ratio schedules sustain higher response rates than interval schedules.FixedVariableRatioCounts responsesFixed ratio(FR)Reinforcer after aset number ofresponsesTYPICAL PATTERNA pause after eachreinforcer, thenrapid responding(“break and run”)Variable ratio(VR)Reinforcer after anumber of responsesthat varies aroundan averageTYPICAL PATTERNFast, steadyresponding with fewlong pausesIntervalCounts timeFixed interval(FI)First responseafter a set time isreinforcedTYPICAL PATTERNLittle respondingearly in theinterval,accelerating towardits end (the“scallop”)Variableinterval (VI)First responseafter a time thatvaries around anaverage isreinforcedTYPICAL PATTERNA relativelyconstant rate,lower than on ratioschedulesTypical patterns, not invariant ones: pausingalso occurs on variable-ratio schedules, andratio schedules sustain higher rates thaninterval schedules.

Figure 2Intermittent schedules deliver a reinforcer after a number of responses (ratio) or after a period of time (interval), on a fixed or variable basis. Patterns as typically described since Ferster and Skinner (1957).

Text description

A two-by-two grid. Columns are fixed and variable schedules; rows are ratio schedules, which count responses, and interval schedules, which count time. On a fixed-ratio schedule a reinforcer follows a set number of responses, and the typical pattern is a pause after each reinforcer followed by rapid responding, called break and run. On a variable-ratio schedule the number of responses required varies around an average, and responding is fast and steady with few long pauses. On a fixed-interval schedule the first response after a set time is reinforced, and responding is sparse early in the interval and accelerates toward its end, a pattern called the scallop. On a variable-interval schedule the first response after a varying time is reinforced, and responding occurs at a relatively constant rate that is lower than on ratio schedules. A note states that these are typical rather than invariant patterns: pausing also occurs on variable-ratio schedules.

The characteristic patterns are well documented, with qualifications. Ratio schedules sustain higher response rates than interval schedules, and ratio schedules fail to maintain responding when reinforcement is too sparse, a breakdown called ratio strain[3]. Fixed-ratio performance is described in textbooks as “break and run”, a pause after reinforcement followed by rapid responding, and variable-ratio performance as fast and steady[2]. A review of fifty years of research found that the pause depends on the upcoming ratio as well as the one just completed, and that pausing also occurs on variable-ratio schedules[2]. Fixed-interval schedules produce responding that accelerates through each interval, shown as a scalloped cumulative record, although some records show a break-and-run pattern instead[15][16]. Variable-interval schedules maintain a relatively constant rate[15].

Punishment and its by-products

Skinner regarded punishment as the commonest technique of control in modern life and argued that it has unfortunate by-products: the aversive stimuli it requires generate emotions, including tendencies to escape or retaliate, and disabling anxieties[6]. He also argued that its suppressive effect is temporary. He cited Thorndike’s later finding that saying “wrong” to human participants did not weaken the preceding behavior, although “right” strengthened it, and laboratory work in which punishment at the start of extinction reduced the rate of responding only until punishment stopped, after which the responses reappeared[6].

That position is historical theory, and the question was not closed by it. A 2002 review of basic and applied research on punishment concluded that further understanding of punishment processes is still needed before a systematic and effective technology of behavior change can be built on them[7].

Limits set by instinct

Skinner’s own experiments used rats and, from the 1940s, pigeons[9]. Keller and Marian Breland, who conditioned over 6,000 animals of 38 species for zoos, museums, fairs and television, reported repeated failures of conditioning outside the experimental chamber[8]. A raccoon reinforced for dropping coins into a container rubbed them together instead, and pigs trained to carry wooden coins to a “piggy bank” increasingly dropped and rooted them, even when this delayed food and the animals were hungrier. The Brelands called the pattern instinctive drift, “learned behavior drifts toward instinctive behavior”, and rejected three assumptions they had brought from the laboratory: that the animal arrives as a blank slate, that species differences are insignificant, and that all responses are about equally conditionable to all stimuli[8].

Applications in treatment

One clinical application with meta-analytic evidence is contingency management, which arranges reinforcing consequences to promote abstinence in the treatment of substance use disorders[4]. A meta-analysis of 47 treatment–control comparisons published between 1970 and 2002 found a mean effect of d = 0.42, larger for opiate and cocaine use than for tobacco; effects declined after treatment ended[4]. Negative reinforcement has its own applied literature: Brian Iwata’s 1987 article describes behavior acquired or maintained through negative reinforcement, the treatment of such behavior, and negative reinforcement used as therapy[17].

Sources17

Numbered in order of first citation.

  1. 1.Skinner, B. F. (1938). The behavior of organisms: An experimental analysis. Appleton-Century.
  2. 2.Schlinger, H. D., Derenne, A., & Baron, A. (2008). What 50 years of research tell us about pausing under ratio schedules of reinforcement. The Behavior Analyst, 31(1), 39–60. https://doi.org/10.1007/BF03392160
  3. 3.Baum, W. M. (1993). Performances on ratio and interval schedules of reinforcement: Data and theory. Journal of the Experimental Analysis of Behavior, 59(2), 245–264. https://doi.org/10.1901/jeab.1993.59-245
  4. 4.Prendergast, M., Podus, D., Finney, J., Greenwell, L., & Roll, J. (2006). Contingency management for treatment of substance use disorders: a meta‐analysis. Addiction, 101(11), 1546–1560. https://doi.org/10.1111/j.1360-0443.2006.01581.x
  5. 5.Baron, A., & Galizio, M. (2005). Positive and negative reinforcement: Should the distinction be preserved?. The Behavior Analyst, 28(2), 85–98. https://doi.org/10.1007/BF03392107
  6. 6.Skinner, B. F. (1953). Science and human behavior. Macmillan.
  7. 7.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. https://doi.org/10.1901/jaba.2002.35-431
  8. 8.Breland, K., & Breland, M. (1961). The misbehavior of organisms. American Psychologist, 16(11), 681–684. https://doi.org/10.1037/h0040090
  9. 9.B. F. Skinner Foundation (n.d.). Biographical information. B. F. Skinner Foundation. www.bfskinner.org/archives/biographical-information/
  10. 10.Thorndike, E. L. (1898). Animal intelligence: An experimental study of the associative processes in animals. The Psychological Review: Monograph Supplements, 2(4), i–109. https://doi.org/10.1037/h0092987
  11. 11.Pavlov, I. P. (1927). Conditioned reflexes. Oxford University Press.
  12. 12.Thorndike, E. L. (1911). Animal intelligence: Experimental studies. Macmillan.
  13. 13.Peterson, G. B. (2004). A day of great illumination: B. F. Skinner's discovery of shaping. Journal of the Experimental Analysis of Behavior, 82(3), 317–328. https://doi.org/10.1901/jeab.2004.82-317
  14. 14.Skinner, B. F. (1948). 'Superstition' in the pigeon. Journal of Experimental Psychology, 38(2), 168–172. https://doi.org/10.1037/h0055873
  15. 15.Catania, A. C. (1962). Independence of concurrent responding maintained by interval schedules of reinforcement. Journal of the Experimental Analysis of Behavior, 5(2), 175–184. https://doi.org/10.1901/jeab.1962.5-175
  16. 16.Dews, P. B. (1978). Studies on responding under fixed-interval schedules of reinforcement: II. The scalloped pattern of the cumulative record. Journal of the Experimental Analysis of Behavior, 29(1), 67–75. https://doi.org/10.1901/jeab.1978.29-67
  17. 17.Iwata, B. A. (1987). Negative reinforcement in applied behavior analysis: An emerging technology. Journal of Applied Behavior Analysis, 20(4), 361–378. https://doi.org/10.1901/jaba.1987.20-361

Further reading

  • Skinner, B. F. (1953). Science and human behavior. Macmillan.
  • Ferster, C. B., & Skinner, B. F. (1957). Schedules of reinforcement. Appleton-Century-Crofts.

Cite this entry

  • APA 7

    Psych Dictionary. (2026, October 18). Operant conditioning. In Psych Dictionary. https://www.psychdictionary.com/concepts/operant-conditioning

  • MLA 9

    “Operant Conditioning.” Psych Dictionary, 18 Oct. 2026, www.psychdictionary.com/concepts/operant-conditioning.

Written by Editorial team. First published October 18, 2026. Editorial standards

Educational reference material. Nothing on this site is a diagnosis, a treatment recommendation or a substitute for care from a qualified clinician.