Shaping in ABA: Successive Approximations, With a Worked Example
Key Takeaways
- 01Shaping builds a behavior that has never happened by reinforcing closer and closer approximations of it.
- 02In a session, you reinforce the nearest response, then stop reinforcing yesterday's approximation and raise the criterion. If the old approximation still pays, the learner has no reason to improve.
- 03Shaping changes a dimension of one behavior, such as duration or topography. Chaining links separate behaviors into a sequence. Write the criterion and the move-up rule before the session.
- 04Task G.11 is shaping dimensions of behavior. Exam items often offer shaping and chaining for the same case, so ask whether one behavior is getting closer to a standard or several are being joined.
Most teaching procedures assume the behavior happens at least sometimes, and then work on when or how often. Shaping is what you reach for when the behavior has never happened at all. You cannot reinforce a response the learner has never made, so you reinforce the nearest thing they do make, and then move the goalposts.
What "successive approximations" actually commits you to
Two things are happening at once, and both have to happen or it is not shaping.
You reinforce a response closer to the target than the last one. That is the forward half, and it is the part everyone remembers.
You stop reinforcing the approximation you were accepting yesterday. That is the differential half, and it is the part that gets skipped. If the old approximation still pays, there is no reason on earth for the learner to produce a better one. Shaping without withholding is just reinforcement, and it stalls exactly where you stopped raising the bar.
You are changing a dimension, not adding a step
This is the distinction that separates shaping from the procedure next to it. Shaping changes some measurable property of one behavior. Chaining links separate behaviors into a sequence.
The dimensions available to you are the ones you can measure: frequency, rate, duration, latency, magnitude or force, and topography, meaning the physical form the response takes. Deciding which dimension you are shaping before you start is what keeps a program honest, because "better" is not a dimension and cannot be graphed.
| Procedure | What it changes | Use it when |
|---|---|---|
| Shaping | A dimension of one behavior | The target response has never occurred |
| Chaining | How separate behaviors link up | The steps exist but the sequence does not |
| Prompting | Nothing; it evokes the response | You need the behavior to happen now so you can reinforce it |
Prompting and shaping are not rivals and are frequently run together. A prompt gets you a response to work with; shaping refines it once you have one.
A worked example: tolerating a haircut
Target: the learner sits in the chair while clippers run for two minutes. Right now they leave the room when the clippers come out. The dimension being shaped is duration.
- Reinforce staying in the room while the clippers sit switched off on the counter.
- Reinforce walking past the chair. Staying in the room no longer earns anything on its own.
- Reinforce sitting in the chair for three seconds, clippers off.
- Reinforce sitting while the clippers run across the room.
- Reinforce sitting while the clippers run within arm's reach, then touching the neck, then one pass, then thirty seconds.
Every rung drops the one below it. That is uncomfortable to do and it is the procedure working: brief protest at a raised criterion is expected, and it is not a sign the plan is wrong.
The criterion is where shaping goes wrong
Both failure modes come from the same decision, made badly.
Raise it too fast and the learner cannot reach the new requirement, nothing gets reinforced, and the behavior you had built weakens. The response is to drop back one rung, reinforce a few times, and split the gap you just tried to jump.
Raise it too slowly and you cement the approximation. A learner reinforced two hundred times for a partial response has a very well-established partial response, and the fix costs far more than the caution saved.
The practical guard is to write the criterion and the move-up rule before the session, not during it. "Three consecutive trials at the current criterion, then advance" is a rule. "When it looks solid" is a feeling, and it drifts by whoever is running the session.
Where this sits on the exam
Task G.11 of the 6th Edition Test Content Outline is "shape dimensions of behavior," and the phrasing is the hint: items tend to give you a scenario and ask which dimension is being shaped, or to offer shaping and chaining as options for the same case. Sort those by asking whether the target is one behavior getting closer to a standard, or several behaviors getting joined up.