Schedules of Reinforcement

Ferster and Skinner’s landmark 1957 book, Schedules of Reinforcement, mapped out how these schedules operate within operant conditioning, showing that each produces a distinct, predictable response pattern.

How and when a behavior is reinforced shapes its strength and persistence, often more than the reward itself.

Types of Reinforcement Schedules

All schedules work within operant conditioning (also known as instrumental conditioning). They differ in what triggers the reinforcer (responses made or time passed) and in whether the rule is predictable (Ferster & Skinner, 1957).

Schedules of reinforcement can be divided into two broad categories: continuous reinforcement, which reinforces a response every time, and partial reinforcement, which reinforces a response occasionally.

The type of reinforcement schedule used significantly impacts the response rate and resistance to the extinction of the behavior.

Research into schedules of reinforcement has yielded important implications for the field of behavioral science, including choice behavior, behavioral pharmacology, and behavioral economics.

Continuous Reinforcement

In continuous schedules, reinforcement is provided every single time after the desired behavior.

Because the behavior is reinforced every time, the association is easy to make and learning is quick. The flip side is rapid extinction once reinforcement stops.

Example: Candy Machines

Candy machines show continuous reinforcement in action. Every time we put money in (behavior), we receive candy in return (positive reinforcement).

However, if a candy machine were to fail to provide candy twice in a row, we would likely stop trying to put money in (Myers, 2011).

We have come to expect our behavior to be reinforced every time it is performed and quickly grow discouraged if it is not.

Continuous reinforcement is ideal for establishing a new response. It is poor for maintaining one.

Applied behavior analysts therefore teach on a continuous schedule, then thin the reinforcement gradually. Thinning too fast backfires. A sudden jump to a very lean schedule makes the response collapse.

Partial (Intermittent) Reinforcement Schedules

Unlike continuous schedules, partial schedules reinforce the desired behavior only occasionally rather than every time. Learning is slower at first, because the link between behavior and reinforcement is harder to spot.

The payoff is durability. Behavior learned this way is more resistant to extinction, because organisms keep going in hopes that they will eventually be rewarded.

For instance, slot machines at casinos operate on partial schedules. They provide money (positive reinforcement) after an unpredictable number of plays (behavior). Hence, slot players are likely to continuously play slots in the hopes that they will gain money in the next round (Skinner, 1953).

Partial reinforcement schedules are the most common in everyday life. They vary along two dimensions.

The first is whether the number of responses required is fixed or variable. The second is whether reinforcement depends on the number of responses (ratio) or on the passage of time (interval).

Each descriptor has a precise meaning:

• Fixed: the number of responses or the time between reinforcements is set and unchanging, so the schedule is predictable.
• Variable: the number of responses or the time between reinforcements changes randomly, so the schedule is unpredictable.
• Ratio: reinforcement occurs after a certain number of responses have been made.
• Interval: reinforcement follows the first response made after a period of time has passed.

Combinations of these four descriptors yield four kinds of partial reinforcement schedules: fixed-ratio, fixed-interval, variable-ratio and variable-interval.

The four partial schedules compare as follows:

Schedule	Rule	Response pattern	Resistance to extinction	Everyday example
Fixed-ratio	Reinforcement after a set number of responses	High, steady rate with a pause after each reward	Moderate; rises as the ratio grows	Piece-rate pay: $500 per 10 dresses
Variable-ratio	Reinforcement after an unpredictable number of responses	Very high, very steady rate; little or no pause	Highest of all	Slot machines, door-to-door sales
Fixed-interval	Reinforcement for the first response after a set time	Pause after reward, then accelerating “scallop”	Moderate; rises as the interval lengthens	Weekly quiz, monthly paycheck
Variable-interval	Reinforcement for the first response after an unpredictable time	Low but steady, stable rate	High (higher than fixed-interval)	Checking email, fishing

Fixed Interval Schedule

In operant conditioning, a fixed-interval schedule is when reinforcement is given to a response after a specific, predictable amount of time has passed.

Such a schedule results in a tendency for organisms to increase the frequency of responses closer to the anticipated time of reinforcement. However, immediately after being reinforced, the frequency of responses decreases.

The fluctuation in response rates means that a fixed-interval schedule will produce a scalloped pattern rather than steady rates of responding.

Plotted over time, the scallop is a flat pause after reinforcement that curves up into a burst of responding just before the deadline. Responses made before the interval ends achieve nothing.

Example: The Weekly Quiz

A teacher gives students a quiz every Monday.

Over the weekend, there is suddenly a flurry of studying for the quiz. On Monday, the students take the quiz and are reinforced for studying (positive reinforcement: receive a good grade; negative reinforcement: do not fail the quiz).

For the next few days, they relax after finishing the stressful experience, until the next quiz date draws near again.

Last-minute exam cramming shows the same scallop. So does the surge of legislative activity as a session’s deadline nears. A fixed weekly or monthly salary works similarly: effort tends to slacken just after payday and build toward the next one.

Variable Interval Schedule

In operant conditioning, a variable-interval schedule is when reinforcement is provided for the first response after a random, unpredictable amount of time has passed.

This schedule produces a low, steady response rate since organisms are unaware of the next time they will receive reinforcers.

The organism responds at a slow, even pace. That is often enough to collect the reinforcer soon after it becomes available, but there is no frantic acceleration like the fixed-interval scallop.

Examples: Pigeons, Phones and Fishing

A pigeon in Skinner’s box has to peck a key to receive a food pellet. It is given a food pellet after varying time intervals ranging from 2-5 minutes.

It is given a pellet after 3 minutes, then 5 minutes, then 2 minutes, etc. It will respond steadily since it does not know when its behavior will be reinforced.

People show the same pattern with their phones. Checking email or social media for messages that arrive at unpredictable times produces the same low, steady rate of checking throughout the day.

Fishing works the same way; a bite can come after any unpredictable wait. A teacher’s surprise pop quizzes keep students studying steadily instead of cramming.

Fixed Ratio Schedule

In operant conditioning, a fixed-ratio schedule reinforces behavior after a specified number of correct responses.

This kind of schedule results in high, steady rates of response. Organisms are persistent in responding because of the hope that the next response might be one needed to receive reinforcement.

Right after each reward, the organism briefly stops, then resumes at a high rate until the next reinforcer. This is the post-reinforcement pause, sometimes called “break-and-run.”

The pause lengthens as the ratio requirement grows, because a larger block of work looms before the next reward (Ferster & Skinner, 1957).

Example: Piece-Rate Pay

An example of a fixed-ratio schedule is a dressmaker paid $500 after every 10 dresses they make.

After shipping a batch of ten and collecting the $500, they are likely to take a short break before starting the next batch. This post-reinforcement pause is the schedule’s signature.

The same logic drives ordinary factory piece-rates and sales commission paid per fixed number of units. A teacher who awards a spelling star for every five correctly spelled words is using the identical rule. The more a person produces, the more reinforcement they earn, so output stays high.

Variable Ratio Schedule

A variable ratio schedule is a schedule of reinforcement where a behavior is reinforced after a random number of responses.

This type of schedule produces very consistent and high rates of responding, as the organism never knows exactly when reinforcement will occur. Due to its unpredictability, organisms persistently repeat the desired behavior, driven by the anticipation that the next response could lead to reinforcement.

Examples: Candy Rewards and Slot Machines

An example of a variable-ratio schedule is a child given candy for reading pages of a book.

They might get candy after 5 pages, then after 3, then 7, then 8. The reward comes at no fixed point.

This unpredictable reinforcement keeps them reading, even on pages that bring no reward.

Slot machines are the classic real-world case. Variable-ratio schedules show little or no post-reinforcement pause, because the next response always might pay off.

Skinner (1953) singled out gambling as the paradigm: variable-ratio contingencies can sustain enormous amounts of responding for very little net reinforcement.

Compound Schedules

Beyond the four simple schedules, researchers have combined the basic rules into several compound schedules. Three matter most.

Each extends Ferster and Skinner’s (1957) original framework.

Differential Reinforcement (DRL, DRH and DRO)

Differential-reinforcement schedules reward a response based on how fast or how often it occurs, not just on whether it occurs at all.

• DRL (differential reinforcement of low rates): reinforces a response only once a minimum time has passed since the last one, teaching an organism to respond slowly and deliberately.
• DRH (differential reinforcement of high rates): does the opposite and reinforces only bursts of rapid responding.
• DRO (differential reinforcement of other behavior): reinforces the absence of a target response, which reduces problem behavior without punishment.

A teacher might use DRO to reinforce a student for periods without talking out of turn. Together, DRL, DRH and DRO show that schedules can shape not just whether a behavior happens, but its pace.

Interlocking, Multiple and Chained Schedules

Other schedules combine two component schedules instead.

• Interlocking: the ratio and interval requirements are interdependent, so meeting one changes the other.
• Multiple: alternates between two component schedules and signals each switch with a cue, called a discriminative stimulus.
• Mixed: alternates between two component schedules with no signal of the switch.
• Chained: requires a sequence of component schedules to be completed before a single terminal reinforcer arrives.

These arrangements let researchers study how stimuli, sequences and past reinforcement history combine to control behavior. Comparing multiple with mixed schedules separates discriminated responding from purely consequence-driven behavior.

A chained schedule shows up in everyday routines too. Getting dressed, making coffee and leaving for work might all have to happen in sequence before a single paycheck follows.

Non-Contingent Reinforcement: Skinner’s “Superstition” Experiment

Not every schedule depends on behavior at all.

In a fixed-time or variable-time schedule, the reinforcer is delivered on a purely temporal basis, regardless of whether the organism responds. There is no genuine link between the response and the reward.

Skinner (1948) showed this vividly. He ran a well-known “superstition” experiment: hungry pigeons received food at a fixed interval, about every 15 seconds, no matter what they were doing.

The pigeons developed their own rituals. They turned counter-clockwise, tossed their heads, and swayed their bodies like a pendulum.

Whatever a bird happened to be doing when food arrived was accidentally, or adventitiously, reinforced. The finding shows that a reinforcement schedule can shape and sustain behavior even when the behavior itself has no real effect on when the reward arrives.

Together, these compound schedules show real flexibility in the model. Researchers can combine, chain, or strip away the response-reinforcer link entirely, and behavior still organizes itself around whatever schedule is in play.

Response Rates of Different Reinforcement Schedules

Ratio schedules, which are linked to the number of responses made, produce higher response rates than interval schedules.

As well, variable schedules produce more consistent behavior than fixed schedules; the unpredictability of reinforcement results in more consistent responses than predictable reinforcement (Myers, 2011).

On a ratio schedule, responding faster earns reinforcers faster. On an interval schedule, it does not. Extra responses before the interval ends are wasted.

A variable schedule removes the cue that a reward has just been delivered or is about to arrive. The post-reinforcement pause of fixed schedules vanishes. Output runs smooth and even.

Ferster and Skinner (1957) read these differences from the cumulative record: a running graph of total responses over time, recorded automatically by a Skinner box. A steeper slope means a faster rate.

Variable-ratio produces the fastest, steadiest responding of all, and fixed-interval the most uneven. The same reinforcer can drive very different behavior depending only on its schedule.

Ferster and Skinner (1957): Mapping Every Schedule

A later retrospective treated this scale of replication as unmatched in the operant literature (Morgan, 2010).

Extinction of Responses Reinforced at Different Schedules

Resistance to Extinction Across Schedules

Resistance to extinction is how long a behavior continues once reinforcement stops. A response high in resistance takes longer to disappear.

Different schedules build different levels of resistance. In general, schedules that reinforce unpredictably resist extinction more than predictable ones do.

The variable-ratio schedule resists extinction more than the fixed-ratio schedule. The variable-interval schedule resists extinction more than the fixed-interval schedule, as long as the average intervals are similar.

Within a fixed schedule, resistance grows as the requirement grows: a larger fixed ratio, or a longer fixed interval, both increase resistance to extinction.

Variable-ratio is the most resistant of all four. This helps explain the pull of gambling addiction. A slot machine (variable-ratio) is far harder to walk away from than a vending machine (continuous reinforcement) for exactly this reason.

The Partial-Reinforcement Extinction Effect

Gamblers often go through long losing streaks with no pay-off. They keep playing anyway.

They stay hopeful that reinforcement will come soon. This persistence is the partial-reinforcement extinction effect (PREE): behavior reinforced only intermittently persists longer after reinforcement stops than behavior that was reinforced every time.

One classic study shows why.

The same effect makes behavior kept up by lean, unpredictable rewards hard to eliminate. Nagging that occasionally works and a child’s tantrum that is sometimes given in to both fit this pattern.

Implications for Behavioral Psychology

In his article Schedules of Reinforcement at 50: A Retrospective Appreciation, Morgan (2010) reviews how schedules-of-reinforcement research has shaped behavioral science.

He highlights three areas where the framework remains productive: choice behavior, behavioral pharmacology and behavioral economics.

Choice Behavior

Behaviorists have long been interested in how organisms choose between alternatives and reinforcers. They study this choice using concurrent schedules.

Two or more schedules run at the same time on different response options. Operating two schedules at once, often both variable-interval, lets researchers see how organisms allocate their behavior between the options. This is choice behavior.

Herrnstein (1961) discovered the matching law using concurrent schedules: an organism’s response rate to each option closely matches the proportion of reinforcement it has produced.

Herrnstein (1970) later generalized this into a quantitative version of the law of effect.

Baum (1974) refined it further. His generalized matching law adds two systematic deviations: bias, a constant preference for one option, and undermatching, responding less extremely than strict matching predicts.

Take Joe, for instance. His father almost always gives him money when he asks, but his mother rarely does.

Because asking his father is reinforced more often, Joe is more likely to ask him rather than his mother for money.

Individuals choose behavior that provides the largest reward. Their choices also depend on the reinforcement’s rate, quality, delay and required effort.

In short, people prefer rewards that are larger, better, sooner and easier to get.

Behavioral Pharmacology

Schedules of reinforcement are used to evaluate preference and abuse potential for drugs. One key tool is the progressive-ratio schedule.

In a progressive-ratio schedule, the response requirement rises each time after reinforcement is attained. In pharmacology research, participants must make an increasing number of responses to earn an injection of a drug (reinforcement).

A single injection may eventually need thousands of responses. Participants are measured for the point where responding stops, called the “break point.”

Gathering data about drugs’ breakpoints allows researchers to categorize their abuse potential. Using the progressive-ratio schedule to assess drug preference is now standard practice in behavioral pharmacology.

Behavioral Economics

Operant experiments are an ideal way to study microeconomic behavior. Participants act as consumers, and reinforcers as commodities.

Researchers alter a commodity’s availability or price by changing the reinforcement schedule, then track how response allocation shifts.

They can change the ratio schedule, for example, so more or fewer responses earn the reinforcer. Two measures follow.

• Elasticity: how much demand changes as the effort “price” of a reinforcer rises or falls.
• Substitutability: whether one commodity can stand in for another when both are offered at the same response price.

The operant laboratory lets researchers manipulate independent variables directly and measure the dependent variables that follow.

Classrooms and Animal Training

Schedule principles guide everyday teaching and training. Both start with continuous reinforcement to build a response, then thin the rewards to a partial schedule so the behavior lasts.

A teacher shaping class participation might praise every attempt at first, then move to intermittent praise of successively better answers.

Assessment matters too. Weekly tests and end-of-term deadlines are fixed-interval schedules, so they predictably produce cramming scallops. Frequent, less predictable assessment approximates a variable schedule and encourages steadier study.

Humane animal training follows the same logic. Trainers establish a response with continuous reinforcement. They then thin to an intermittent, often variable, schedule so the behavior stays reliable and resistant to extinction (Ferster & Skinner, 1957).

A clicker serves as a conditioned reinforcer: a sound that has been paired with food, so it can bridge the delay before the food arrives.

Critical Evaluation

The schedules framework is one of psychology’s most precise and predictive bodies of law. But it inherits the limits of the behaviorist tradition it grew from.

Strengths

• Precise and predictive: produces a distinctive, replicable pattern on the cumulative record (a running graph of responses over time) for every schedule.
• Wide real-world reach: explains gambling, pay structures, drug-taking and consumer demand, and underpins applied behavior-change programs.

Schedules give operant psychology genuine predictive power. Each one yields a distinctive, replicable signature on the cumulative record, a running graph of total responses over time where a steeper slope means a faster rate.

The signature includes the fixed-ratio break-and-run, the fixed-interval scallop, and the steady variable-ratio rate (Ferster & Skinner, 1957). These patterns recur across species and settings.

The framework predicts the rate and persistence of behavior, not just that reward “works.” That precision matters. It explains behaviorism’s staying power.

The framework also reaches far beyond the laboratory, explaining gambling persistence, pay-structure effects, drug break points and consumer demand. It underpins token economies, programs where desired behaviors earn tokens later exchanged for rewards.

Ayllon and Michael (1959) showed that a psychiatric nurse could act as a “behavioral engineer,” systematically reinforcing adaptive behavior in patients. Their approach launched the token-economy tradition (Morgan, 2010).

Modern technology applies the same rules. Apps, feeds and games deliver notifications, likes and new content at unpredictable times, mirroring variable-ratio and variable-interval schedules to keep users returning.

Limitations

• Cognitive factors are neglected: humans can respond to a described rule rather than their actual reinforcement history.
• Reductionist and ethically fraught: omits meaning and motivation, and hands institutions a tool for covert behavioral control.
• Limited animal-to-human generalizability: the core findings come from rats and pigeons in tightly controlled chambers.

The “pure” schedule pattern often breaks down in people. Schedule effects come mainly from controlled animal chambers, but in humans, instructions, expectations and rule-governed behavior can override the contingencies.

A person who understands a schedule may respond to the rule itself rather than to their actual reinforcement history. That limits how well schedule laws generalize to humans.

Explaining behavior solely through reinforcement schedules also omits meaning, motivation and cognition, reducing purposive human activity to external contingencies rather than choice.

The same power raises ethical concerns. Employers, institutions and technology companies can arrange schedules that shape behavior without a person’s awareness or consent. Deliberate variable-reward design in gambling and technology is a direct ethical charge against applied schedule engineering.

Schedule effects were also established chiefly with rats and pigeons in tightly controlled chambers. Human behavior is shaped by language, self-awareness and culture. So these “pure” patterns may not transfer directly to people.

Rival Account: Latent Learning and Cognitive Maps

Tolman’s work on latent learning challenges the idea that reinforcement is necessary for learning. Tolman and Honzik (1930) gave rats no reward at all for ten days. The rats had learned the maze’s layout anyway.

Once a reward appeared, their performance improved almost instantly. That was far too fast for them to be learning the route from scratch.

Tolman concluded that reinforcement affects the performance of a learned response, not its acquisition. What an animal acquires is a cognitive map. This internal representation of the environment can be consulted flexibly once the animal has a reason to act (Tolman, 1948).

Schedule theorists explain a rat’s improving performance by its reinforcement history alone. The cognitive account needs an unobservable expectancy instead: that a given route leads to a given outcome.

Neither account cancels the other. Schedule effects are exceptionally reliable and precisely quantified, and latent learning is equally well replicated.

A reasonable synthesis is that both processes run in parallel. Schedules shape when and how vigorously a learned response is performed. Expectancies can determine what is learned even when no schedule is operating.

Contemporary Research

Schedule research has moved beyond the laboratory. A 2026 study examined how reinforcement operates in real social-media behavior.

The design was cross-sectional, so it cannot show whether heightened sensitivity to likes contributes to depression or follows from it. The findings complement, rather than settle, the critique of variable-reward design: they show an individual-differences signature of reinforcement in real platform data (Mirea et al., 2026).

Mini Quiz

Below are examples of schedules of reinforcement at work in the real world. Read the examples and then determine which kind of reinforcement schedule is being used.

1. Daniel is a door-to-door salesperson whose job is to sell insurance. He has successful days where 1 out of every 5 people he talks to will buy insurance. However, he also has unsuccessful days when only 1 out of every 15 people he talks to will buy insurance.
   Answer: Variable-ratio
2. Ashraf has to take medicine for pain relief. His medicine comes from a machine where he has to push a button to release a pill. To prevent Ashraf from taking too much medicine, the machine only releases a pill 4 hours after the last one.
   Answer: Fixed-interval
3. Nikki enjoys fishing on the weekends. It takes her on average 20 minutes to catch a fish but the time varies for each fish. It can take as little as 5 minutes or as long as 40 minutes for a fish to bite.
   Answer: Variable-interval
4. Lee recently adopted a young puppy. In order to housetrain it, he praises the puppy and gives it a treat every time it eliminates in the backyard instead of in the house.
   Answer: Continuous reinforcement schedule
5. Aisha is employed by a company to sell newspapers. For every 50 newspapers that she sells, she is paid a $10 commission.
   Answer: Fixed-ratio

References

Ayllon, T., & Michael, J. (1959). The psychiatric nurse as a behavioral engineer. Journal of the Experimental Analysis of Behavior, 2(4), 323–334.

Baum, W. M. (1974). On two types of deviation from the matching law: Bias and undermatching. Journal of the Experimental Analysis of Behavior, 22(1), 231–242.

Ferster, C. B., & Skinner, B. F. (1957). Schedules of reinforcement. Appleton-Century-Crofts.

Herrnstein, R. J. (1961). Relative and absolute strength of response as a function of frequency of reinforcement. Journal of the Experimental Analysis of Behavior, 4(3), 267–272.

Herrnstein, R. J. (1970). On the law of effect. Journal of the Experimental Analysis of Behavior, 13(2), 243–266.

Humphreys, L. G. (1939). The effect of random alternation of reinforcement on the acquisition and extinction of conditioned eyelid reactions. Journal of Experimental Psychology, 25(2), 141–158.

Mirea, D.-M., Chen, S., Ding, J., Palacio-Soto, K., Mildner, J. N., Kelley, S. W., Gillan, C. M., Nook, E. C., & Niv, Y. (2026). The reinforcement effect of social media likes in depression. JAMA Psychiatry, 83(9), 930–938. https://doi.org/10.1001/jamapsychiatry.2026.1430

Morgan, D. L. (2010). Schedules of reinforcement at 50: A retrospective appreciation. The Psychological Record, 60(1), 151–172.

Myers, David G. (2011). Psychology (10th ed.). Worth Publishers.

Skinner, B. F. (1948). “Superstition” in the pigeon. Journal of Experimental Psychology, 38(2), 168–172. https://doi.org/10.1037/h0055873

Skinner, B. F. (1953). Science and human behavior. Macmillan.

Tolman, E. C. (1948). Cognitive maps in rats and men. Psychological Review, 55(4), 189–208. https://doi.org/10.1037/h0061626

Tolman, E. C., & Honzik, C. H. (1930). Introduction and removal of reward, and maze performance in rats. University of California Publications in Psychology, 4, 257–275.

Further Information

• Schedules of reinforcement at 50: A retrospective appreciation
• Ayllon, T., & Michael, J. (1959). The psychiatric nurse as a behavioral engineer. Journal of the Experimental Analysis of behavior, 2(4), 323-334.