You cannot reinforce a behavior that never happens. If a learner has never said "more," never sat for ten minutes, or never spoken loudly enough to be heard at a counter, waiting for the full skill to appear could take forever. Shaping solves that problem: you start with something the learner already does and reinforce small steps toward the goal.
Why it matters
Shaping lets you teach behaviors that are not yet in a learner's repertoire, without needing the learner to imitate or follow a prompt first. It is used for first words, clearer speech, longer engagement, better handwriting, and many motor skills.
Practice items on this task ask you to recognize shaping, decide which response to reinforce under the current criterion, choose what to do when progress stalls, or tell shaping apart from chaining and prompt fading. Some describe a caregiver who shapes a problem behavior by accident.
Key ideas
- Start where the learner is. The first approximation is a response the learner already emits that resembles the goal in some way, such as "mm" on the way to "more."
- Reinforce the current step, not the old ones. Shaping is built on differential reinforcement. Responses that meet today's criterion get reinforced. Earlier approximations no longer do. When an old approximation stops paying off, the learner often repeats it more forcefully and then tries new variations. One of those is a slightly better version, and that is the one you reinforce.
- Raise the bar gradually. Once the learner meets the current step reliably, the criterion moves one step closer to the terminal behavior. Your supervisor's plan says what each step is and when to move on.
- Watch the step size. Steps that are too big can make responding fall apart. Staying too long at one step can make that approximation so strong that the next one is hard to get. Your supervisor decides whether to add a smaller step.
- Shaping can change form or a dimension. Some programs shape a new form of behavior (sounds becoming a word). Others keep the same form and shape a dimension of it, such as duration, rate, latency, or loudness.
- Accidents happen. The same process can strengthen problem behavior when adults respond only to louder or more intense versions of it.
Definitions
- Shaping: teaching a new behavior by differentially reinforcing successive approximations of it until the learner performs the terminal behavior.
- Successive approximations: a sequence of responses, each one closer to the terminal behavior than the one before.
- Terminal behavior: the final form or level of behavior the shaping program is working toward.
- Differential reinforcement: reinforcing responses that meet a criterion while withholding reinforcement for responses that do not.
- Shaping within a response topography: keeping the form of the behavior the same while shaping one of its dimensions, such as how long, how fast, or how loudly it occurs.
Examples from a session
Three learners, three kinds of shaping
- Home, a first word: Leo, 3, is learning to say "more" at snack. His RBT, Dana, first reinforces "mm." When Leo says "mm" reliably, Dana reinforces only "muh," then only "mor," and finally "more." Each time the criterion rises, the earlier sound no longer earns a cracker.
- School, duration: Diego, 9, stays with independent seatwork for about 2 minutes. His plan reinforces 3 minutes of work, then 4, then 5, moving up only after he meets each step across several sessions. The behavior looks the same; only its duration changes.
- Community program, loudness: Tomás, 24, speaks very softly when ordering at a café counter. His RBT reinforces requests at or above a target volume level and raises that level a little at a time. Requests at his old soft volume are not reinforced, but anything at or above the current level is.
Non-example: Dana says "more" and Leo echoes the full word; over sessions she says less and less of it until Leo asks on his own. That is prompt fading, not shaping. Shaping changes the response itself. Prompt fading changes the help that comes before the response.
Shaping by accident
A caregiver gives Eli, 5, the tablet when he whines. Over time she starts holding out, and gives in only when he screams. Weeks later, Eli screams loudly almost every time. Reinforcing only the more intense versions of a behavior shaped louder screaming. Report patterns like this to your supervisor rather than trying to fix them on your own.
Common mistakes / exam traps
- Mixing up shaping and chaining. Chaining links separate steps of a task analysis into one routine, like handwashing. Shaping changes one behavior gradually, like sounds turning into a word.
- Mixing up shaping and prompt fading. In prompt fading, the learner already gives the full response with help, and the help is reduced. In shaping, the response itself starts out incomplete and gets closer to the goal.
- Calling extinction of old approximations "punishment." When you stop reinforcing "mm," you are withholding reinforcement, not adding or removing anything to suppress behavior.
- Reinforcing everything to keep the learner happy. Reinforcing responses below the current criterion undoes the differential reinforcement that makes shaping work.
Exam trap
If the scenario describes the same behavior getting closer to a goal across sessions, think shaping. If it describes separate steps done in order, think chaining. If it describes less and less help, think prompt fading.
Exam trap
Extinction burst or a step that is too big? Right after the criterion rises, a learner who repeats the old approximation louder, then tries new variations, is showing extinction-induced variability; the step is working, so keep reinforcing only the new criterion. If responding stops or falls apart over several trials, the step is too big: return to the last step met and tell your supervisor.
Quick check
An RBT is shaping a learner's reach toward a cup. Last week she reinforced any arm movement toward the cup. This week, per the plan, she reinforces only reaches that come within a few inches of it. What should she do when the learner makes the smaller arm movement from last week?