The big idea
Response maintenance is the extent to which a learner keeps performing a target behavior after some or all of the intervention that produced it has ended. Gabrielle, age 7, learned to wash her hands with a picture schedule and a token chart. She still washes them correctly three months after both were removed, which shows maintenance.
Maintenance is about time. Generalization (B.14) asks whether behavior shows up with new stimuli or in new forms. Maintenance asks whether it lasts. A skill that disappears the week the therapist stops visiting was not taught in a way that helps the person.
Breaking it down
What counts as maintenance
Two things must both be true:
- Some or all of the intervention has been withdrawn.
- The behavior continues to occur.
Behavior that continues while the full intervention is still running is just the intervention working. To check maintenance, analysts take maintenance probes: they measure the behavior at set points after the intervention ends (for example, two weeks, one month and three months later) without the programmed contingencies.
Maintenance and generalization
Cooper, Heron and Heward describe three forms of generalized behavior change: response maintenance, setting or situation generalization, and response generalization. Stokes and Baer (1977) defined generalization to include behavior change across time, so some authors call maintenance generalization across time. The exam treats maintenance as its own outcome, defined by the passage of time after intervention.
The forms can combine. A skill that is still occurring months later in a new classroom shows both maintenance and setting generalization.
Why some behavior lasts
When reinforcement stops entirely, how long behavior keeps going is its resistance to extinction. Behavior with a history of intermittent reinforcement usually persists longer in extinction than behavior reinforced every time. That is one reason practitioners thin a schedule before ending an intervention (see B.9).
In daily life, though, durable maintenance rarely depends on extinction alone. It depends on whether the behavior contacts reinforcement that already exists in the person’s environment. A behavior trap is a natural contingency of this kind: once a person enters it with a small, easily reinforced response, it keeps reinforcing and expanding the behavior. When Bashir learns to name dinosaurs, he may find that classmates who love dinosaurs talk with him at recess. Peer interaction, not the therapist, now keeps him talking.
Maintenance vs. maintaining variables
Practitioners often say problem behavior is “maintained by escape” or “maintained by attention.” That phrase names the reinforcer currently sustaining the behavior, which is its function. It has nothing to do with response maintenance after intervention. Both use the word maintain, and exam items test whether you can tell them apart.
Seeing it in practice
| Example | Intervention status | Behavior now | Maintenance? |
|---|---|---|---|
| Priya, 8, learned to raise her hand with a token system. Tokens ended in March. In May probes she raises her hand on 90% of opportunities. | Withdrawn | Continues | Yes |
| A staff member, Ximena, completes a session checklist every day while her supervisor gives daily feedback on it. | Still in place | Continues | Not yet: the intervention is still running |
| Marcus, 19, learned a bus route with a job coach. The coach faded out, and six months later he rides the route on his own. | Withdrawn | Continues | Yes |
| Kenji learned to request breaks at the clinic and now requests them at school. | Not the key detail | New setting | This is setting generalization |
| A teen’s flossing dropped to zero within a week of the dentist ending a sticker chart. | Withdrawn | Stopped | No |
Common mix-ups
- Maintenance is not acquisition. Data collected while the intervention is fully in place show acquisition or performance, not maintenance.
- Maintenance is not generalization. Time after intervention points to maintenance. A new person, place, material or response form points to generalization.
- Maintenance is not a maintaining variable. “Maintained by attention” describes function, not durability.
- Maintenance is not automatic. Stokes and Baer called the common approach of teaching and hoping for the best “train and hope.” Lasting behavior change has to be planned (see G.16).
- Continuous reinforcement is good for teaching, weaker for lasting. It produces fast acquisition but less persistence when reinforcement stops.
Exam tip: Find the moment the intervention ended. If the item describes behavior measured after that point in the same or a similar situation, think maintenance. If the key detail is a new person, place, material or response, think generalization.
Quick recap
- Response maintenance is behavior continuing after some or all of the intervention has been withdrawn.
- It is measured with probes taken after the intervention ends.
- Intermittent reinforcement and contact with natural contingencies, such as behavior traps, support maintenance.
- Maintenance is across time; generalization is across stimuli or responses.
- “Maintained by escape” describes function, not response maintenance.
Further reading
- Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapter 30.
- Stokes, T. F., & Baer, D. M. (1977). An implicit technology of generalization. Journal of Applied Behavior Analysis, 10(2), 349-367.
Key terms
- Response maintenance (Maintenance)
- The extent to which a learner continues to perform a target behavior after some or all of the intervention that produced it has been withdrawn.
- Resistance to extinction
- How long, or how many times, a behavior keeps occurring after extinction begins. It is greater after intermittent reinforcement (especially on variable schedules) than after continuous reinforcement, and it is also affected by the length of the reinforcement history, the current motivating operation and response effort.
- Intermittent reinforcement (Partial reinforcement)
- Any schedule in which some, but not all, occurrences of a response are reinforced. The basic intermittent schedules are fixed ratio, variable ratio, fixed interval and variable interval. Behavior maintained on intermittent schedules is more resistant to extinction than behavior maintained on continuous reinforcement.
- Stimulus generalization
- The occurrence of a trained response in the presence of stimuli that were not part of training but share features with the training stimuli. The more a new stimulus resembles the training stimulus, the more likely the response.
- Behavior trap
- A natural contingency of reinforcement that, once a person enters it with a small, easily reinforced response, keeps reinforcing and expanding the behavior without programmed help.
Check yourself
4 questions on B.15. Choose an answer to see whether you're right and why.
Show the answer
Answer: B. The intervention ended and the behavior kept occurring in the same situation two months later, which is response maintenance. Nothing about the setting or the response is new, so this is not generalization. Spontaneous recovery is a brief return of behavior during extinction, not a skill that lasts.
Show the answer
Answer: B. Maintenance is measured after some or all of the intervention has been withdrawn. While praise still follows every log, the data show the intervention working, not maintenance. Duration alone does not make it maintenance.
Show the answer
Answer: B. In a functional assessment, the maintaining variable is the consequence that currently keeps the behavior going. Response maintenance is a different idea: an outcome measured after intervention ends. The two share a word, and items sometimes test whether you can tell them apart.
Show the answer
Answer: D. Behavior with a history of intermittent reinforcement usually shows more resistance to extinction than behavior reinforced every time. This is one reason practitioners thin reinforcement before ending an intervention. Continuous reinforcement produces fast learning but weaker persistence when reinforcement stops, so the skill learned faster is not the one that lasts.
Adds a step to your study record. Saved in this browser only.