cumulative

G.11 Shaping dimensions of behavior

What the task asks: Use differential reinforcement of successive approximations to change the form of a response, or another dimension such as its rate, duration, latency or force, until it reaches a target.

Behavior-Change Procedures5 min read4 practice questionsOfficial outline (opens in a new tab)

The big idea

Shaping builds a behavior the learner does not yet show by reinforcing successive approximations: responses that come closer and closer to a goal, called the terminal behavior. At each step, responses that meet the current criterion are reinforced and responses that fall short are not. Once the learner meets the criterion reliably, the criterion moves a little closer to the goal.

Shaping matters because you cannot reinforce a behavior that never happens. A child with few speech sounds, or an employee who takes three minutes to start a task, needs a route from where they are to the goal. Shaping provides it, using what the learner already does as the starting point.

Breaking it down

Two parts working together

Shaping combines two operations. Differential reinforcement reinforces responses that meet a criterion and withholds reinforcement for those that do not (see G.2). A shifting criterion then raises the requirement step by step.

Differential reinforcement on its own produces response differentiation: a new, narrower response class made up of the reinforced variations. Shaping repeats that process with a moving target. Extinction of the old approximation helps too. Responses that stop being reinforced become more variable for a while, and that variability gives the learner a chance to emit the next, closer approximation.

Across and within topographies

Shaping across topographies changes the form of the response. Each approximation looks different from the last: “buh,” then “bub,” then “bubble.”

Step 1 reinforce "mm" Step 2 reinforce only "ma" Step 3 reinforce only "milk" "mm" "mm" "ma" "mm" "ma" "milk"
Nikhil's shaping steps from the example below. Once the criterion moves, the earlier approximation no longer earns milk, so only the response closest to the goal is reinforced at each step.

Shaping within a topography keeps the form the same and changes another dimension:

DimensionExample of shaping within a topography
RateWords read correctly per minute rise from 20 to 60.
DurationTime seated at a work table grows from 30 seconds to 10 minutes.
LatencyTime to start a task after an instruction drops from 3 minutes to 10 seconds.
Magnitude (force or intensity)Speaking volume rises from a whisper to a level a classmate can hear.

Designing a shaping program

  1. Define the terminal behavior and a mastery criterion, such as 10 minutes of seated work on three days in a row.
  2. Pick a starting response the learner already emits that resembles the goal in some way.
  3. Plan the steps, but treat the plan as a guess. The learner’s data decide when to move.
  4. Move to the next criterion only after the current one is met consistently. Steps that are too large put the whole response on extinction, and the learner may stop responding or show problem behavior. Steps that are too small, or staying too long at one level, can make that approximation hard to move past.
  5. If performance breaks down, return to the last step the learner met and use smaller increments.
  6. Reinforce immediately. A conditioned reinforcer such as a click or a quick “yes” marks the exact response that met the criterion.
  7. Keep reinforcing the terminal behavior once it appears, then plan for maintenance (G.16).

Shaping takes time and close attention, and progress is rarely smooth. Adding prompts, models or instructions (G.7, G.9, G.10) usually speeds it up, as long as they are faded later.

Percentile schedules

A percentile schedule makes the shaping criterion objective. The learner’s most recent responses are ranked, and a new response is reinforced only if it beats a set proportion of them. For example, a work bout earns a token only if it lasts longer than 7 of the last 10 bouts. The criterion rises automatically as performance improves and eases if it slips, which keeps reinforcement steady without relying on a staff member’s judgment.

Shaping, chaining and stimulus fading

All three move in small steps.

ProcedureWhat changes from step to stepWhat stays
ShapingThe response. Earlier approximations stop being reinforced and drop out.The antecedent, roughly
Chaining (G.12)Different responses are linked into a longer sequence.Earlier steps, which keep being performed
Stimulus fading (G.8)The antecedent, as an exaggerated cue or prompt is gradually removed.The response

Seeing it in practice

  • Speech. Nikhil, 3, says “mm” when he wants milk. His team reinforces “mm,” then only “ma,” then only “milk.” The form changes at each step, so this is shaping across topographies.
  • School. Itzel, 12, waits a long time before starting math worksheets. Her teacher praises starting within 2 minutes of the instruction, then within 90 seconds, then within 60. Only latency changes.
  • Workplace. Tomás packs 8 orders an hour and the goal is 20. His job coach uses a percentile schedule: an hour earns a bonus only if his count beats most of his recent hours.
  • Home, by accident. A parent ignores mild whining but gives in to screaming, and later gives in only when screaming includes kicking. The parent is shaping more intense problem behavior. Shaping works whether or not anyone plans it.

Common mix-ups

  • Shaping vs. stimulus fading. If the learner’s behavior changes across steps, it is shaping. If the cue changes and the behavior stays the same, it is fading.
  • Shaping vs. chaining. In shaping, the earlier approximation disappears. In chaining, every step remains part of the final performance.
  • Shaping vs. a changing-criterion design. Both raise a criterion in steps, but the design (D.7) is a way to show experimental control over a behavior already in the repertoire. Shaping is a procedure for producing a new form or level.
  • Reinforcing below criterion. Reinforcing responses that fall short of the current step undoes the differentiation you have built.
  • Accidental shaping vs. an extinction burst. A burst is a temporary rise when reinforcement stops. If reinforcement keeps following ever more intense versions, the behavior is being shaped.

Exam tip: Ask what changes from one step to the next. The response’s form or one of its dimensions points to shaping. The antecedent points to fading. A new link added to a sequence points to chaining.

Quick recap

  • Shaping is differential reinforcement of successive approximations to a terminal behavior.
  • Shaping across topographies changes form. Shaping within a topography changes rate, duration, latency or magnitude.
  • Move the criterion only after the current step is met reliably, and back up when performance breaks down.
  • Percentile schedules set an objective criterion that adjusts to the learner’s recent responses.
  • Shaping changes the response, fading changes the antecedent, and chaining links responses.

Further reading

  • Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapter 22.
  • Galbicka, G. (1994). Shaping in the 21st century: Moving percentile schedules into applied settings. Journal of Applied Behavior Analysis, 27(4), 739-760.

Key terms

Shaping
Building a new behavior, or a new level of an existing one, by reinforcing successive approximations to a terminal behavior. Responses that meet the current criterion are reinforced and others are not, and the criterion moves closer to the goal each time the learner meets it reliably.
Successive approximations
The series of intermediate responses in shaping, each closer to the terminal behavior than the one before. Each approximation is reinforced until it occurs reliably, and then reinforcement shifts to the next one.
Differential reinforcement
Reinforcing behavior that meets a defined criterion while withholding reinforcement for behavior that does not. In shaping, the criterion is a form of the response or a level of one of its dimensions. Procedures such as DRA, DRO, DRL and DRH apply the same logic to build alternatives or reduce problem behavior.
Response differentiation
The result of differential reinforcement within a response class: variations that meet the criterion become more frequent, variations that do not become less frequent, and a new, narrower response class emerges.
Percentile schedule
A rule for setting a shaping criterion from the learner's own recent performance. Each new response is compared with a set number of the most recent responses and is reinforced only if it ranks above a chosen point among them, so the criterion rises and falls with performance.

Check yourself

4 questions on G.11. Choose an answer to see whether you're right and why.

  1. Question 1A child says 'ba' when they want a ball. His therapist reinforces 'ba' for a few days, then reinforces only 'bal', and finally reinforces only 'ball'. Which procedure is this?
    Show the answer

    Answer: A. Each approximation has a different form, and earlier forms stop earning reinforcement, so this is shaping across topographies. Shaping within a topography would keep 'ball' the same and change a dimension such as volume. Nothing about the antecedent is being faded, and the sounds are not separate links that all stay in a chain.

  2. Question 2Over two weeks, a technician raised a learner's seated-work criterion from 1 minute to 2 and then 3 minutes, with steady success. The technician then jumped the criterion to 8 minutes. The learner now leaves the table after a minute or two and throws materials. What should the BCBA advise?
    Show the answer

    Answer: C. The jump was too large, so the learner's sitting rarely met the criterion and was in effect on extinction, which can bring problem behavior. Going back to the last criterion the learner met and moving in smaller steps restores reinforcement. Holding at 8 minutes prolongs extinction, and reinforcing any sitting throws away the progress already shaped.

  3. Question 3A job coach wants a shaping criterion for an employee's packing speed that does not rely on staff judgment and adjusts to recent performance. Each hour earns a bonus only if its count beats most of the employee's previous 10 hours. What is this arrangement?
    Show the answer

    Answer: B. Reinforcing a response only when it ranks high among the learner's recent responses defines a percentile schedule, so the criterion rises and falls with the employee's own performance. DRH also reinforces higher rates, but its criterion is a fixed value set by the analyst. A changing-criterion design is an experimental design, not a reinforcement rule.

  4. Question 4At the grocery store, a parent used to buy candy after one polite request. Over months the parent began holding out until the child cried, and lately gives in only when the child cries and kicks the cart. What best describes what has happened?
    Show the answer

    Answer: C. The parent reinforces only the more intense versions and keeps raising the requirement, which is differential reinforcement of successive approximations: shaping, even though nobody planned it. An extinction burst is a temporary increase when reinforcement stops altogether. Here reinforcement still follows the most intense version, so the behavior keeps escalating.

Adds a step to your study record. Saved in this browser only.

Type to search all 104 guides and the glossary.