Skip to content
Behavior Analysis Concepts

Shaping in ABA: Definition, Examples, When to Change Steps and Practice Items

Shaping in ABA reinforces successive approximations of a target behavior. Definition, examples, a worked graph, how RBTs record it and 6 practice items.

By Editorial team

Updated 20 min read

Single-line drawing of an open hand at the lower left and a smaller hand at the upper right, joined by a rising row of stepping stones on a pale blue background.
On this page

Shaping in ABA is a procedure that teaches a new behavior by reinforcing successive approximations, responses ever closer to a target behavior, while earlier versions stop being reinforced (Association for Science in Autism Treatment). A child reinforced for “wa”, then only “wa-da”, then only “water” is being shaped. RBTs implement it under RBT Test Content Outline (3rd ed.) task C.10.

This page is for RBTs and students, with a short section for parents at the end. It has examples of shaping in ABA, a worked example with a graph that shows when to advance, hold or back up a step, how to document shaping, the exam task codes, and six practice items. More concept explainers are in Behavior Analysis Concepts.

What is shaping in ABA?

In plain words, shaping in ABA means building a behavior the learner does not show yet by reinforcing small changes toward it, one step at a time. The Association for Science in Autism Treatment (ASAT) gives the full definition: “A process in which an instructor teaches a new behavior by systematically and differentially reinforcing ‘successive approximations’ of a target behavior while no longer reinforcing previous approximations of the behavior.” Its example is a child who is reinforced for “mmm”, then for “ma” (when “mmm” no longer is), on the way to “mommy”.

In our reading of common practice, a shaping procedure runs in this order, with the supervising BCBA writing steps 1 to 4 into the plan and the RBT carrying out steps 5 to 7 and reporting what happens:

  1. Define the terminal behavior, the final form or level you are shaping toward, so two observers would agree (“says ‘water’ clearly enough that an unfamiliar adult understands”, “works for 5 minutes without stopping”).
  2. Pick a starting behavior the learner already shows that resembles the terminal behavior. A response that never occurs can’t be reinforced.
  3. Write the successive approximations, responses each closer to the terminal behavior than the one before, and the criterion a response must meet at each step.
  4. Set the rules for advancing and backing up.
  5. Use differential reinforcement: reinforce, right away, every response that meets the current step, and never responses below it, including last week’s step.
  6. Move the criterion when the data meet the advance rule, and back up when they meet the back-up rule.
  7. Record every session, so the next person runs the right step.

Examples of shaping in ABA

Examples of shaping in ABA include building a first word sound by sound, lengthening how long a learner works, and shortening how long they take to start. In each one the instruction stays the same, and only responses that meet the current step are reinforced. All six learners are fictional.

Learner and goal Shaping steps
Ana, 3, asking for a drink Target: says “water”. Reinforced in turn: “wa”, then “wa-da”, then “water”. Changes: form of the word
Ben, 9, independent work Target: works for 10 minutes without stopping. Reinforced in turn: 2, then 4, 6, 8 and 10 minutes. Changes: duration
Cara, 14, answering in class Target: answers loudly enough to be heard from the back row, about 25 feet. Reinforced in turn: answers heard at 6 feet, then 12, 18 and 25 feet. Changes: magnitude (volume)
Dev, 6, starting a task Target: starts within 10 seconds of the instruction. Reinforced in turn: starts within 40, then 30, 20 and 10 seconds. Changes: latency
Eli, 5, kicking a ball Target: swings his leg and kicks the ball forward. Reinforced in turn: touches the ball with his foot, then pushes it, then swings and kicks. Changes: form of the movement
Fay, 11, reading aloud Target: reads a passage at 60 words per minute. Reinforced in turn: 35, then 45, 55 and 60 words per minute. Changes: rate

Shaping in ABA therapy is also used to make a replacement behavior easier to understand: a learner who reaches up to be lifted is first reinforced for any sound with the reach, then only for “uh”, then only for “up”.

When shaping happens by accident

The same process can make problem behavior worse. A parent who ignores a quiet “phone” and gives in when the request turns into yelling is reinforcing only the louder version. As the parent gets used to the noise, it takes louder yelling before they give in. Nobody planned the steps, but the criterion kept rising. If you see a pattern like this in a session, describe it in your notes and tell your supervisor.

Shaping vs chaining and fading: what is changing?

In shaping, the response itself changes (its form, duration, latency, magnitude or rate) while the instruction stays the same. In chaining, different responses are linked into one routine, and the learner either takes over more of the links (forward and backward chaining) or needs less help on every link (total-task chaining). In prompt and stimulus fading, the help shrinks while the response stays the same. So when a scenario confuses you, ask: what is changing?

Procedure What is changing, example and RBT TCO task
Shaping Changes: the response (its form or one dimension); the criterion for reinforcement moves. Fictional example: “wa” → “wa-da” → “water” to ask for a drink. RBT TCO C.10
Chaining Changes: how many links of a routine the learner does alone, or how much help each link needs (total task). Fictional example: handwashing (turn on water, wet hands, soap, rub, rinse, dry). RBT TCO C.5
Prompt fading Changes: the help given before or during a response. Fictional example: touching the red card (guided by the hand, then a touch at the elbow, then a gesture). RBT TCO C.7
Stimulus fading Changes: a feature of the materials or instruction, such as a highlight that grows paler. Fictional example: the correct picture starts larger than the others and shrinks to the same size. RBT TCO C.7

Shaping also differs from plain differential reinforcement. Both reinforce responses that meet a criterion, but in shaping the criterion keeps moving toward the terminal behavior. The Praxis Learn lessons on task analysis and chaining (C.5) and prompting and prompt fading (C.7) cover the other two procedures.

What are the types of shaping in ABA?

Behavior analysts usually describe two types of shaping in ABA. In shaping across topographies, the form of the response changes from step to step (“wa” to “water”, a touch of the foot to a kick). In shaping within a response topography, the form stays the same and one dimension changes. Both fit the definition Athens, Vollmer and St. Peter Pipkin (2007) use, citing Cooper, Heron and Heward (1987): “the gradual modification of some property of responding.” Dimensions that can be shaped include:

  • Duration: how long the behavior lasts (working for 2, then 4 minutes).
  • Latency: how soon it starts after the instruction (within 40, then 30 seconds).
  • Magnitude: how forceful or loud it is (answers heard from farther away).
  • Rate: how often it occurs in a period of time (words read per minute).

The BCBA outline words this task as “Shape dimensions of behavior” (G.11). The BCBA lesson on shaping across dimensions on Praxis Learn has more examples.

Percentile schedules: shaping without fixed steps

Fixed steps are not the only way to set a criterion. In a percentile schedule, a response is reinforced only if it beats a set rank among the learner’s own most recent responses, so the criterion moves with the learner’s performance. Galbicka (1994) proposed percentile schedules to standardize shaping “across subjects, trainers, and times” while keeping “the sensitivity to the individual’s repertoire that is the hallmark of shaping.” Athens, Vollmer and St. Peter Pipkin (2007) used one to lengthen task engagement for 4 students and found it worked best when the criterion was computed from a relatively large number of recent observations (20). In their version, with 20 observations ranked, a response had to beat the 10th-ranked one to be reinforced.

How big should each step be? Advance, hold or back up

Step size in shaping is a clinical judgment: steps small enough that the learner keeps meeting them, large enough that the behavior keeps moving. Athens, Vollmer and St. Peter Pipkin (2007), citing earlier authors, wrote that shaping techniques used in applied research “more often resemble an ‘art form’ than an established procedure”. Citing Galbicka (1994), they give examples of the decisions a trainer faces, including “How quickly should the criterion for reinforcement be increased?”, “How large an increase in this criterion should be made?” and “What should happen when the learner has a setback?” In our reading, a written plan should answer those questions before the session, so they don’t rest on the RBT’s judgment in the moment.

The table below shows one set of rules a supervisor might write. It is our example of practice judgment, not a BACB rule. Your supervising BCBA sets the numbers for each learner; task E.2 of the RBT outline is to “seek and prioritize clinical direction from a supervisor in a timely manner.”

What the data show (example rule) Decision and what the RBT does
At least 4 of 5 opportunities meet the step in 2 consecutive sessions Advance: start the next step when the plan says to, and record the change
Anything in between, including the first sessions after a step change Hold: keep reinforcing only responses that meet the current step
1 or 0 of 5 opportunities meet the step in 2 consecutive sessions Back up: return to the last step the learner met, and tell the supervisor the same day
Responding stops, or emotional responding grows, within a session Back up or stop: follow the plan’s safety and back-up steps, record what you saw, contact the supervisor
No change in either direction for several sessions Hold and report: report the stall; the supervisor decides on a smaller step, a new reinforcer or a new plan

In our reading, two errors show up in the data. Steps that are too big make responding fall apart: fewer attempts, more errors, sometimes crying or leaving. Steps that are too small, or staying too long at one step, can make that approximation so strong that the next one is hard to get.

What should you do if the learner gets frustrated or stops responding during shaping?

If the learner gets frustrated or stops responding during shaping, stop raising the criterion, follow the plan’s back-up rule, record what happened in observable terms, and tell your supervisor. If the plan has no back-up rule, keep the current step and contact the supervisor before changing anything. Don’t drop the program, invent a new step or start reinforcing an old approximation on your own. In our reading, pushing materials away, crying or leaving can also mean the learner is withdrawing assent; follow what the plan says to do then.

Worked shaping example: reading the graph session by session

Fictional example. Maya, 8, works on a math worksheet for about 30 seconds before she stops, looks around and leaves her seat. The terminal behavior is working for 5 minutes without stopping. Each session has 5 independent work periods. Her plan uses the rules in the table above: advance after 2 consecutive sessions with at least 4 of 5 periods meeting the step, and back up to the last step met after 2 consecutive sessions with 1 or fewer.

Line graph of fictional shaping data across 22 sessions: the percentage of work periods in which the learner met the step rises to 80 to 100% at 30 seconds, 1 minute and 2 minutes, falls to 20% and 0% when the step jumps to 4 minutes, recovers after a return to 2 minutes, then reaches 80 to 100% at 3, 4 and 5 minutes.
Figure 1. Percentage of 5 work periods per session in which Maya met the current step. Dashed lines mark step changes; the label above each phase is the step. Fictional data.
Sessions and step Result and decision
Sessions 1–3, step 30 seconds 3, 4, 5 of 5 periods met the step → advance after session 3
Sessions 4–6, step 1 minute 3, 4, 4 of 5 periods met the step → advance after session 6
Sessions 7–10, step 2 minutes 2, 3, 4, 5 of 5 periods met the step → hold through session 9; advance after session 10
Sessions 11–12, step 4 minutes 1, 0 of 5 periods met the step → back up: two sessions at 1 or fewer; supervisor told the same day
Sessions 13–14, step 2 minutes 4, 5 of 5 periods met the step → the supervisor adds a 3-minute step; advance after session 14
Sessions 15–17, step 3 minutes 3, 4, 4 of 5 periods met the step → advance after session 17
Sessions 18–20, step 4 minutes 3, 4, 5 of 5 periods met the step → advance after session 20
Sessions 21–22, step 5 minutes 4, 5 of 5 periods met the step → terminal behavior met; the supervisor sets the mastery and maintenance plan

Three things to read from the graph:

  • The dip after each step change is expected. Sessions 4, 7, 15 and 18 are lower because the old approximation no longer pays off, and how long Maya works varies before it settles. One low session right after a change is a reason to hold, not to back up.
  • The 4-minute step was too big. It added 2 minutes at once, twice the largest jump she had met (1 to 2 minutes). Her session 12 times (50 seconds to 2 minutes 30 seconds) showed she could work past 2 minutes in some periods but was nowhere near 4, so the supervisor added a 3-minute step rather than retrying 4.
  • The back-up rule protected the program. Without it, Maya would have kept working at a step where almost nothing was reinforced, which is how a learner stops trying.

Why do earlier approximations go on extinction?

In shaping, the previous approximation goes on extinction: once the criterion moves, the old version no longer produces reinforcement. Athens, Vollmer and St. Peter Pipkin (2007), summarizing Galbicka (1994), describe the differential reinforcement in shaping as producing “a mixture of both extinction of and reinforcement of responding until the terminal criterion is reached.” The variation extinction produces is where the next, closer approximation comes from. Task D.6 of the RBT outline is to “describe secondary effects of extinction (e.g., extinction burst, response variation, resurgence, emotional responding)”. All four can appear at a step change:

  • Extinction burst: the old version gets louder or more frequent at first. A learner whose “wa” no longer gets water says “wa! wa!” more loudly.
  • Response variation: new forms appear. Some, like “wa-da”, are closer to the target, and those are the ones you now reinforce.
  • Resurgence: a behavior that was reinforced in the past and had stopped comes back when the current response stops being reinforced, such as reaching or crying for the drink.
  • Emotional responding: crying, pushing materials away, leaving the table.

Under a rule like the example above, a burst in the first few trials of a new step doesn’t mean the step is too big; data across sessions decide that. The secondary effects of extinction lesson on Praxis Learn covers D.6.

How do RBTs document shaping?

RBTs document shaping by recording, every session, the step being run, how many opportunities met it, the actual values, any prompts, and every step change with the rule that triggered it and who approved it or was told. Read the prior session notes and the shaping graph before you start, so you run the current step and not last week’s.

Record these in each shaping session:

  • Program and step, with the criterion as the plan words it (“Step 4: 4 minutes of continuous work”).
  • Opportunities and how many met the step (“1 of 5”), plus the actual values for duration, latency or volume programs. The values show the supervisor where a smaller step belongs.
  • Prompts, recorded separately from independent responses.
  • The reinforcer and when it was delivered.
  • Step changes: what changed, which plan rule triggered it, and who approved it or was told.
  • Anything that might affect progress, such as illness or a schedule change (RBT outline task E.3), and any extinction effects described as actions.

Fictional data sheet, Maya, session 12, step 4 (4 minutes):

Work period Time worked before stopping, result and notes
Work period 1 2 min 10 s; didn’t meet 4 minutes; no reinforcer; looked up, left seat
Work period 2 1 min 45 s; didn’t meet 4 minutes; no reinforcer
Work period 3 2 min 30 s; didn’t meet 4 minutes; no reinforcer; pushed worksheet away, cried about 20 s
Work period 4 50 s; didn’t meet 4 minutes; no reinforcer; pushed worksheet away
Work period 5 1 min 20 s; didn’t meet 4 minutes; no reinforcer

Fictional example, weak note: “Maya had a hard day and couldn’t do her work. She got frustrated. Will try again next time.”

Fictional example, objective note: “Independent work shaping, step 4 (4 min continuous work). 0 of 5 work periods met the step; times worked 0:50 to 2:30 (median 1:45). Pushed the worksheet away in periods 3 and 4 and cried for about 20 s in period 3; crying stopped without intervention and did not recur, and Maya returned to the worksheet each time it was re-presented, so per the plan the session continued at step 4. Second consecutive session at 1 or fewer (session 11: 1 of 5), which meets the plan’s back-up rule; next session returns to step 3 (2 min). Supervisor (J.R., BCBA) informed by message at 3:40 p.m.; she will review adding a 3-minute step.”

What changed:

  • The step and its criterion are named, so the next RBT runs the same step.
  • Counts and the range of times replace “couldn’t do her work”.
  • “Frustrated” became what Maya did and for how long.
  • The note names the rule behind the back-up and who was told, and when.

Task E.4 asks RBTs to “communicate objectively what occurred during the session in accordance with applicable legal, regulatory, and workplace requirements.” For the rest of a session note, see how to write your first RBT session notes.

Shaping on the RBT and BCBA exams: C-11, C.10 and G.11

RBT shaping questions are keyed to task C.10 of the RBT Test Content Outline (3rd ed.), “Implement shaping procedures.” Older study guides cite C-11, the same task’s code in the retired RBT Task List (2nd ed.).

Document Code, wording and use
RBT Task List (2nd ed.) C-11: “Implement shaping procedures.” Retired; the 3rd-ed. outline covers RBT exams from 2026
RBT Test Content Outline (3rd ed.) C.10: “Implement shaping procedures.” Used for RBT exams from 2026; Domain C, Behavior Acquisition, 19 of 75 scored questions
BCBA Test Content Outline (6th ed.) G.11: “Shape dimensions of behavior.” Used for BCBA exams from 2025; Domain G, Behavior-Change Procedures, 25 of 175 scored questions
RBT Initial Competency Assessment (effective January 1, 2026) Task 9: “Shaping: Implement shaping procedures.” Shown to an assessor with a client or in a role-play

One trap: in the 3rd-ed. RBT outline, C.11 is “Implement token economies.” An old flashcard that says “C-11: shaping” uses the retired numbering. The tasks that sit next to shaping on the exam are C.5 (chaining), C.7 (prompts and fading), D.4 (extinction), D.6 (secondary effects of extinction) and, for BCBA candidates, G.12 (chaining).

For a study plan around the whole outline, see the RBT exam study guide. The RBT C.10 shaping lesson on Praxis Learn has key terms, exam traps and a quick check.

Shaping practice questions

Every scenario is fictional. Decide before you open the answer.

1. Reading the step data

Jonah’s plan shapes the volume of his answers in group. It says: advance after 2 consecutive sessions in which at least 4 of 5 answers meet the current volume step; back up after 2 consecutive sessions with 1 or fewer. His last three sessions at the current step were 3 of 5, 4 of 5 and 5 of 5. What should the RBT do at the next session?

  • A. Hold for one more session to make sure the step is stable
  • B. Move to the next volume step and record the change
  • C. Back up, because one of the three sessions was 3 of 5
  • D. Ask Jonah to repeat each answer more loudly until it meets the step
Answer and rationale

B. The last two sessions (4 of 5, then 5 of 5) meet the plan’s advance rule, so the criterion moves. A is tempting, but holding longer than the plan says can make the current volume so strong that the next step is harder to reach. C misreads the rule, and D adds a prompt the plan doesn’t include. Task C.10.

2. What is changing?

When asked “Touch cup,” Ravi touches the card with the printed word CUP. At first a picture of a cup is printed above the word; over sessions the picture is made lighter until only the word is left. Ravi touches the right card every time. Which procedure is this?

  • A. Shaping, because the picture changes a little at a time
  • B. Shaping within a topography, because Ravi’s touching gets more accurate
  • C. Stimulus fading, because a feature of the card changes while the response stays the same
  • D. Chaining, because the picture is removed in steps
Answer and rationale

C. Ravi gives the same correct response from the start; what changes is the card, so this is stimulus fading. A is the classic trap: in shaping the response itself changes, not the materials. B fails because his accuracy doesn’t change, and D because touching a card is one response, not a sequence of different responses. Task C.7, contrasted with C.10.

3. Across or within?

Sami is learning to draw a circle, with no outline or dots on the page. His RBT reinforces any curved mark, then only curves that go at least halfway around, then only closed circles. Which describes this program?

  • A. Shaping across topographies
  • B. Shaping within a topography, on the magnitude dimension
  • C. Chaining, because a circle is drawn in several strokes
  • D. Stimulus fading, because the drawing gets closer to a circle
Answer and rationale

A. A curve, a half circle and a closed circle are different shapes on the page, so the form of the response is what is shaped. Within-topography shaping, as in B, would keep the shape and change something like how hard Sami presses. C fails because each attempt is one response that gets closer to a circle, not separate steps linked in order. D needs help in the materials, such as a dotted outline, that is gradually removed; there is none here. Task G.11 (RBT C.10).

4. Which dimension?

After his dad says “Shoes on,” Omar sits for about 45 seconds before he picks up a shoe. His plan reinforces picking one up within 35 seconds, then 25, then 15; how long he takes to put the shoes on is not timed. Which dimension is being shaped?

  • A. Duration
  • B. Rate
  • C. Magnitude
  • D. Latency
Answer and rationale

D. The criterion is how soon Omar starts after the instruction, which is latency. A is the trap: duration would be how long putting on the shoes takes, and the plan doesn’t time that. Rate would count how often a response occurs in a period of time, and magnitude would be its force or size. Task G.11.

5. Right after the step changes

Kofi’s step has just moved from “mo” to “mor”. In the first trials he says “mo” louder and faster, then “moh”, then “mor-r”, then “mor”. What is happening, and what should the RBT do?

  • A. Punishment is suppressing “mo”; stop the program and tell the supervisor
  • B. An extinction burst and response variation; keep reinforcing only responses that meet the new step and record what happened
  • C. The step is too big; go back to “mo” after the first louder trial
  • D. Resurgence; reinforce “mo” again until he calms down
Answer and rationale

B. The louder, faster “mo” is an extinction burst, and the new forms are response variation, which is what produced “mor”. C is tempting, but a burst in the first trials is not evidence the step is too big; data across sessions decide that. A mislabels extinction as punishment, and D would reinforce the old approximation and undo the step. Task D.6.

6. The session note

Which note entry best documents a shaping step change?

  • A. “Step 3 (2 min continuous work): 5 of 5 periods met; second consecutive session at 4 or more. Per plan advance rule, step 4 (4 min) starts next session; BCBA notified.”
  • B. “Step 3 met. Moving to step 4 next session.”
  • C. “Maya worked for 2 minutes five times and was proud of herself.”
  • D. “Step 3: 100%. Advanced to step 4 per BCBA.”
Answer and rationale

A. It names the step and its criterion, the count, the rule that triggered the change and who was told. D is close but leaves out the criterion and the number of opportunities, so a reader can’t check the decision. B gives no data, and C describes a feeling instead of what Maya did. Task E.4.

For more exam-style items across Domain C, try the free RBT Behavior Acquisition practice exam on Praxis Learn.

Shaping at home: a short guide for parents

At home, shaping behavior means praising or rewarding a slightly better version of a skill than your child showed before, and no longer rewarding the old version. Your child’s BCBA can tell you which step the team is on, so everyone asks for the same thing. Two everyday examples (fictional):

  • First words at snack: “wa” gets water this week; once “wa” is easy, the team asks for “wa-da”.
  • Staying at the dinner table: 1 minute, then 2, 3 and 4, with praise or a favorite activity when the timer ends.

Two home skills are often called shaping but are mostly chaining. Using a spoon (hold, scoop, lift to the mouth) and tying shoes (cross, loop, pull tight) are sequences of different steps, so they are usually taught with a task analysis. Shaping fits one part of the sequence, such as scooping with less spilling.

A few habits help:

  • Use the same step across adults. If one adult still rewards the old version, progress stalls.
  • Expect a short burst. When the old version stops working, your child may try it harder for a while. Stay with the plan, and see extinction bursts for families.
  • Tell the BCBA what you see. If your child seems stuck or upset at a step, the team can make the step smaller.

This section is general information, not advice for a specific child.

Sources

All examples on this page are fictional. Sources last checked on October 28, 2026.

Reviewed by the Praxis Notes team on October 28, 2026.

About the author

Praxis Notes Team

Editorial team

The Praxis Notes team consists of experienced ABA professionals, software developers, and content creators dedicated to improving documentation efficiency for RBTs and BCBAs. Our team combines deep clinical expertise with innovative technology to create tools and resources that save time while maintaining the highest standards of compliance and quality.

More from Praxis Notes Team