Behavioral
Extinction Burst
The temporary surge of behavior that occurs when a previously reinforced response stops being rewarded
An extinction burst is the temporary increase in frequency, intensity, or variability of a previously reinforced behavior when reinforcement is withdrawn. The phenomenon was systematically described by B. F. Skinner in The Behavior of Organisms (1938) and forms a fundamental principle of operant conditioning. When a pigeon previously rewarded for pecking a key suddenly stops receiving food, it pecks faster, harder, and more variably before the behavior eventually decreases — the burst phase. The same pattern appears in toddlers escalating tantrums when parents stop responding, gamblers chasing losses at extinguishing slot machines, and abandoned vending machines being shaken. Lerman, Iwata, and Wallace (1999) reviewed 113 cases in applied behavior analysis and found extinction bursts in 39% of clinical extinction procedures, with intensity declining over a few sessions when extinction was applied consistently.
- First describedB. F. Skinner (1938)
- Clinical incidence39% of cases (Lerman, Iwata, Wallace 1999, n=113)
- Three burst featuresIncreased frequency, force, variability
- Typical durationFirst few sessions if extinction is consistent
- Critical conditionTotal reinforcement withdrawal — partial schedules sustain behavior
- Real-world examplesTantrums, vending-machine shaking, app-notification checking
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why extinction burst matters
- Parenting. Knowing tantrums temporarily worsen when ignored helps parents persist through the burst rather than capitulate and re-establish the behavior.
- Behavior therapy. Applied behavior analysis programs for autism, OCD, and habit reversal must plan for and manage potential bursts.
- Habit change. Quitting smoking or social media produces predictable craving spikes that fade if not reinforced.
- Animal training. Trainers expect and ride out variability bursts when shaping new behaviors.
- Workplace management. Removing previously available perks or attention triggers brief escalation before adjustment; consistent leadership prevents reinforcing the burst.
- Product design. User-engagement systems should anticipate burst behavior when notifications change to avoid alienating users.
- Public policy. Removing perverse incentives often produces short-term protests that fade when rules are enforced consistently.
Common misconceptions
- Bursts always happen. About 60% of clinical extinction procedures show no detectable burst, especially when alternatives are reinforced.
- The burst means extinction is failing. The opposite — the burst signals that reinforcement removal is being detected; persistence resolves it.
- Capitulation just delays the inevitable. It actually strengthens the behavior under partial reinforcement, making future extinction harder.
- Extinction equals punishment. Extinction removes reinforcement; punishment adds an aversive consequence. They have different effects and side effects.
- Bursts last weeks. Most resolve within a few sessions of consistent extinction; long-lasting bursts usually indicate accidental reinforcement.
- It only applies to children and animals. Adults show identical burst patterns in habits, addiction recovery, and digital behavior.
Frequently asked questions
What is an extinction burst?
A short-lived increase in the frequency, intensity, or variability of a behavior immediately after the reinforcement that maintained it is removed. If pressing a button produced food and then stops, the organism presses faster and harder before pressing slows down. The burst is one of three classic extinction effects — the others are increased response variability and emotional behavior (frustration, aggression).
Why does it occur?
Two functional explanations dominate. (1) Variability: when familiar responses fail, the organism samples nearby behaviors more aggressively in search of what works — an evolved adaptive search. (2) Frustration-aggression: nonreinforcement is aversive, producing arousal that energizes the established response. Amsel's frustration theory (1958) provided early formal support for the second account; both likely operate.
How common is it in clinical work?
Lerman, Iwata, and Wallace (1999) reviewed 113 extinction procedures in applied behavior analysis for severe behavior problems and found bursts in 39% of cases, with about 25% showing aggression. Bursts were more likely when extinction was implemented alone rather than combined with differential reinforcement of alternative behavior. Modern practice typically pairs extinction with reinforcement of alternatives to mitigate bursts.
Why does intermittent reinforcement prevent extinction?
Behaviors maintained on partial schedules — say, a slot machine paying out only sometimes — are far more resistant to extinction than continuously reinforced behaviors. The organism cannot distinguish a no-reward streak from random variation, so persistence pays. This partial reinforcement extinction effect (PREE) is one reason gambling, intermittent texting, and unpredictable bosses produce such durable, hard-to-extinguish behaviors.
What does it look like in everyday life?
A toddler whose tantrum is no longer rewarded with attention escalates volume and duration before the tantrum frequency drops. A user who stops getting notifications opens the app more often before checking less. A jammed vending machine prompts shaking, kicking, button-mashing — variability within the same act. Pets ignored at meal time bark louder before quieting. Each illustrates the same operant pattern.
How do clinicians manage it?
Three standard tactics. (1) Differential reinforcement of alternative behavior (DRA) — reward an acceptable substitute. (2) Antecedent control — prevent the trigger that elicits the unwanted behavior. (3) Consistency — partial reinforcement of the unwanted behavior during extinction worsens bursts and may permanently strengthen it. Crisis-management plans and parent training programs incorporate all three to ride out bursts safely.
Is extinction burst inevitable?
No. Lerman et al. found roughly 60% of extinction procedures showed no detectable burst. Inevitability depends on reinforcement history, schedule, and whether alternatives are reinforced. Combining extinction with concurrent reinforcement of alternative behavior dramatically reduces burst likelihood. The popular advice "things will get worse before they get better" overstates a partial finding.