Successive approximations
The series of intermediate responses in a shaping program, each one closer to the target behavior than the step before it.
Successive approximations are the steps of a shaping program. The first approximation is a response the learner already makes that resembles the target in some way. Each following approximation is a little closer to the terminal behavior, and the last approximation is the target itself. The steps can differ in form (the sounds of a word), or in a dimension such as duration, distance, force or latency.
In shaping, only the current approximation and responses closer to the goal are reinforced. Earlier approximations, which used to be reinforced, are now placed on extinction. This mix of reinforcing the new step and no longer reinforcing the old one is what moves behavior forward.
The supervisor decides the steps and the criterion for moving between them, such as meeting a step on a set number of opportunities. The RBT's job is to know the current step, reinforce responses that meet it immediately, withhold reinforcement for responses below it, and record data on the step reached each session. If the learner shows frustration or stops responding, the plan usually calls for returning to the last step the learner met. Step size matters: steps that are too large lead to frequent nonreinforcement, and steps that are too small slow progress.
Successive approximations describe a change in the response. Changing a prompt or a stimulus gradually is fading, not a series of successive approximations of a behavior.
Example and non-example
| Example | To build in-seat behavior, the steps in a plan are sitting for 5, 15, 30 and 60 seconds. The client now meets the 15-second step, so the RBT reinforces sitting for 15 seconds or more and no longer reinforces 5 seconds. |
|---|---|
| Non-example | An RBT moves from a full physical prompt to a partial physical prompt to a gesture while teaching a client to hang up a coat. The response stays the same; the help is being faded, so these are prompt-fading steps, not successive approximations. |
Don't confuse it with
| Prompt fading | Prompt fading gradually removes help for a response the learner is already making, while successive approximations are changes in the response itself. |
|---|---|
| Task analysis | A task analysis breaks a routine into a sequence of different steps that all occur each time, while successive approximations replace one another as the learner progresses. |
Test yourself
-
Question 1A plan accepts "ba" as a correct mand for "ball" for now. A child reaches toward a ball and says "ba." What should the RBT do?
Show answer
D: Give the ball right away and record a mand. The RBT reinforces whatever response the plan currently accepts, immediately, using the item requested. The supervisor decides when to raise the requirement.
- A: The plan accepts "ba" at this stage; requiring more changes the criterion.
- B: This punishes an acceptable response and removes the reinforcer.
- C: A delayed reinforcer weakens the link between the mand and the ball.
Task C.4 · 40-hour curriculum section E
-
Question 2How does chaining differ from shaping?
Show answer
D: Chaining links several responses; shaping refines one. Chaining teaches a sequence of separate responses (the steps of a task analysis), while shaping reinforces successive approximations of a single response until the target form is reached.
- A: Shaping is built on reinforcement of successive approximations.
- B: Both are used with learners of all ages.
- C: This reverses the two definitions.
Task C.5 · 40-hour curriculum section E
-
Question 3What is shaping?
Show answer
A: Reinforcing successive approximations of a terminal behavior. Shaping builds a new behavior by reinforcing responses that are closer and closer to the target behavior while no longer reinforcing earlier, less similar forms.
- B: Reinforcing steps of a task analysis in order from the first step is forward chaining.
- C: Gradually removing an added cue is stimulus fading.
- D: Reinforcement on a time schedule regardless of behavior is noncontingent reinforcement.
Task C.10 · 40-hour curriculum section E
Answer every question to see your score.
On the RBT exam
This term belongs to task C.10 Shaping in content area C, Behavior Acquisition, which is 19 of the 75 scored questions.