Translated from the Portuguese original. Read the original (Português)
Escape and avoidance are the two faces of negative reinforcement. In escape, what bothers you is already happening and the action ends it — you leave the noisy room. In avoidance, the action keeps the discomfort from arriving — you never enter the room at all. Both are strengthened by the relief they produce. Avoidance, however, has a trap: when it works, the person never finds out what would have happened without it. That detail is exactly what keeps fears, anxiety, and much of procrastination going.
In Behavior Analysis (Cooper, Heron & Heward), escape is the contingency in which the response ends or removes an aversive stimulus that is already present; avoidance is the contingency in which the response prevents, delays, or reduces the presentation of an aversive stimulus that is not yet present. Both are cases of negative reinforcement: the future frequency of the response increases. Neither should be confused with punishment, which has the opposite effect — it decreases behavior.
The difference is about timing
Escape and avoidance really are similar: in both, behavior produces relief, and in both the process at work is negative reinforcement. What separates them is when the aversive stimulus enters the picture.
- Escape: the aversive stimulus is already there. The response interrupts it. The relief is immediate and observable — the bad thing is over.
- Avoidance: the aversive stimulus has not arrived yet. The response makes sure it never arrives (or arrives later, or with less intensity). The “relief” here is the absence of something.
The same source of discomfort can produce both things in the same person, at different times. Someone who has felt sick from the smell of a cleaning product might leave the room (escape) and, the following week, switch products so the smell never shows up at all (avoidance).
Escape or avoidance? Train your eye
The question that settles almost every case is simple: was the discomfort already happening when the person acted? Tap the cards to check:
Two kinds of avoidance
Not every avoidance response has a prior warning. Experimental research describes two quite different arrangements:
Signaled (or discriminated) avoidance
There is a warning that precedes the aversive event, and responding in the presence of that warning cancels the event. Over time, the warning itself starts to feel uncomfortable: through respondent conditioning, it gets paired with the aversive stimulus and acquires a conditioned aversive function.
This is the basis of the two-factor theory, proposed by Mowrer in 1947: a respondent process creates the discomfort in the presence of the signal, and an operant process strengthens the response that ends that signal. It's the most intuitive way to think about avoidance — we see someone's name on the phone screen and just don't answer.
Unsignaled avoidance (free-operant, or “Sidman avoidance”)
Here there is no external warning. The aversive event would simply happen again after a certain amount of time, and each response delays the next occurrence by a fixed interval. In 1953, Murray Sidman showed that organisms learn to respond at a stable rate based on that delay alone — with no signal to warn them.
In everyday life, this is the backdrop of many maintenance routines: checking your balance from time to time, servicing the car, answering messages before they pile up. There's no specific warning — what keeps the response going is the repeated postponement of something unpleasant.
But if there's no signal to “switch off,” what reinforces the response? Herrnstein and Hineline (1966) showed that the reduction in the overall frequency of the aversive stimulus is already enough to maintain the behavior — a clear, punctual relief after each response isn't necessary. Instead of looking moment by moment, the explanation looks at the aggregate: over time, responding makes the aversive event happen less often.
The trap of avoidance: it protects itself from disconfirmation
Escape behaviors are relatively easy to observe and test — you just check whether the discomfort ended. Avoidance is different: when it works, nothing happens. And that is exactly what makes it so stubborn.
Someone who avoids an elevator out of fear of a panic attack leaves the situation with the impression that “it worked.” But the missing piece of information is still missing: what if I had gotten in? Because the avoidance response is emitted beforehand, the person never makes contact with what would have happened without it — and the frightening expectation is never put to the test. Avoidance, therefore, does not correct itself over time: it persists precisely by preventing its own correction.
This mechanism explains a lot:
- Fears and phobias that go on for years without fading, even when the person knows, rationally, that the risk is small;
- Safety behaviors — carrying medication “just in case,” only going out with company, checking things over and over — which give immediate relief and preserve the fear;
- Procrastination, where putting off the task brings relief now and raises the cost later;
- Withdrawal and isolation in depressive suffering, when avoiding commitments reduces short-term discomfort while cutting off access to activities that provide reinforcement.
This is why exposure worksIf the problem is that the expectation is never tested, the solution is to test it. In exposure with response prevention, the person approaches what they fear without emitting the avoidance response — in a planned, gradual way, and with consent. Craske and colleagues (2014) point to expectancy violation and the removal of safety signals among the ingredients that make exposure more effective: the new learning depends on the person discovering, in practice, that the feared outcome does not come true.
Escape and avoidance in the clinic and at school
In applied settings, one of the most consistent findings is the weight of demand escape. Beavers, Iwata and Lerman (2013) reviewed roughly three decades of research — 435 studies, with data from 981 functional analyses — and task escape shows up among the most frequently identified functions in challenging behavior.
The reading is straightforward: if screaming, dropping to the floor, or biting makes a difficult activity go away, those behaviors are negatively reinforced and tend to repeat. There's no “bad intent” in this — there's an effective contingency at work.
What evidence-based practice suggests, in these cases:
- Teach an alternative response that works as well or better. Carr and Durand (1985) showed that teaching a child to ask for help or a break — functional communication training — reduces problem behavior, because the new response accesses the same consequence in a more efficient and socially appropriate way.
- Work on the antecedent, not just the consequence: adjust how hard the task is, break it into smaller steps, revisit missing prerequisites, offer choices, mix in easy requests, and schedule breaks that don't depend on the behavior. The demand is often aversive because it's poorly calibrated.
- Use differential reinforcement to strengthen engagement and asking for a break.
- Handle escape extinction with caution. Simply “not letting the person escape” can cause a temporary worsening and strain the relationship, and should rarely be the only strategy. It appears as a component of a plan that, above all, teaches alternatives and makes the task more accessible.
In every case, step zero is a functional assessment: finding out why the behavior happens before deciding what to do. Two identical behaviors can have opposite functions.
Avoidance isn't drama, laziness, or a lack of willpowerIt is learned behavior, maintained by real, immediate relief. Calling someone who avoids a situation “dramatic” ignores the contingency at work — and, worse, doesn't help at all. It's also worth remembering: sometimes avoiding is the right response (no one needs to put up with an environment that causes harm, or a genuine sensory overload). The target for intervention is the avoidance that keeps narrowing a person's life — and the way forward is to broaden their repertoire with consent, never to force exposure.
Test your understanding
Quick quiz
Three questions to reinforce the concept. You see the explanation right away.
Further reading
- Cooper, J. O.; Heron, T. E.; Heward, W. L. Applied Behavior Analysis. 3rd ed., 2020 — chapter on negative reinforcement, escape, and avoidance.
- Mowrer, O. H. “On the dual nature of learning — a re-interpretation of ‘conditioning’ and ‘problem-solving’”. Harvard Educational Review, 17, p. 102-148, 1947.
- Sidman, M. “Avoidance conditioning with brief shock and no exteroceptive warning signal”. Science, 118(3058), p. 157-158, 1953.
- Herrnstein, R. J.; Hineline, P. N. “Negative reinforcement as shock-frequency reduction”. Journal of the Experimental Analysis of Behavior, 9(4), p. 421-430, 1966.
- Carr, E. G.; Durand, V. M. “Reducing behavior problems through functional communication training”. Journal of Applied Behavior Analysis, 18(2), p. 111-126, 1985.
- Beavers, G. A.; Iwata, B. A.; Lerman, D. C. “Thirty years of research on the functional analysis of problem behavior”. Journal of Applied Behavior Analysis, 46(1), p. 1-21, 2013.
- Craske, M. G.; Treanor, M.; Conway, C. C.; Zbozinek, T.; Vervliet, B. “Maximizing exposure therapy: an inhibitory learning approach”. Behaviour Research and Therapy, 58, p. 10-23, 2014.
- Catania, A. C. Learning. 4th ed. (Portuguese ed.: Aprendizagem: Comportamento, Linguagem e Cognição.)
This entry is educational and does not replace assessment, diagnosis or treatment by a qualified health professional. Every person is unique — for a specific case, seek individual guidance.