Prompting and Prompt Fading: The RBT Field Guide

Prompting is the most-used procedure in applied practice and one of the most-tested topics on the RBT exam. The reason it generates so many wrong answers is that candidates learn the prompt hierarchy and stop there — while the exam is usually asking about fading, not prompting.

Every prompt you deliver creates a small problem you now have to solve. That framing is the whole topic.

Prompt types, from most to least intrusive

Two distinctions the exam cares about:

Response prompts versus stimulus prompts. Response prompts (physical, model, gestural, verbal) act on the learner's behavior. Stimulus prompts (positional, visual highlighting) act on the materials themselves. A common question describes moving the correct card closer and asks for the classification — that is a stimulus prompt, specifically position.

Intrusiveness is not the same as effectiveness. A full physical prompt is the most intrusive but not automatically the "strongest" choice; intrusiveness describes how much you control the response, not how well it teaches.

The two hierarchies

Most-to-least (MTL)

Start with the level of assistance that guarantees success, then systematically reduce it.

• Minimizes errors during acquisition
• Best for brand-new skills, safety-relevant skills, and learners with a history of frustration or error patterns
• Risk: prompt dependency if fading is not deliberate

Least-to-most (LTM)

Start with the smallest prompt, escalating only if the learner does not respond.

• Promotes independence and gives the learner the chance to respond unaided
• Best for skills partially in the repertoire
• Risk: more errors, and errors can be practiced

Exam questions frequently present a learner and ask which hierarchy fits. The deciding factors are almost always: Is the skill new? Is there a safety concern? Does the learner have an error history? Yes to any of those pushes toward most-to-least.

Time delay

A third, quieter procedure that shows up regularly.

• Constant time delay: a fixed pause (say 3 seconds) between the instruction and the prompt, every trial
• Progressive time delay: the pause increases gradually — 0 seconds, then 1, then 2, then 3

Time delay is a fading procedure disguised as a prompting procedure. It works by giving the natural cue a chance to control responding before assistance arrives. A scenario describing "waits three seconds before providing the model" is time delay, not simply modeling.

The part that matters most: stimulus control transfer

The purpose of a prompt is to be removed. Until the learner responds to the natural discriminative stimulus — the actual instruction, the actual materials, the actual social situation — the skill is not acquired.

A learner who says "ball" only after you say "say ball" has not learned to label the ball. They have learned to echo you. Control sits with your prompt, not the object.

Transferring control means the response occurs when the natural cue is present and the prompt is absent. Practically:

• Fade on a plan written into the program, not on how the session feels
• Use time delay to insert opportunities for independent responding
• Reinforce unprompted responses more richly than prompted ones — differential reinforcement of independence is the engine of fading
• Track prompt level as data, not just correct/incorrect. "Correct with full physical" and "correct independent" are not the same outcome, and if your data sheet cannot tell them apart, fading cannot be evaluated

Prompt dependency: how it happens

Prompt dependency is not a learner trait. It is a procedural outcome, and it is produced by predictable staff behaviors:

• Prompting before the learner has had a chance to respond
• Delivering the same reinforcer for prompted and independent responses
• Fading on intuition rather than on the written criterion
• Repeating the instruction while the learner is still processing the first one

The fix is usually structural rather than motivational: insert a wait, differentially reinforce independence, and follow the fading criteria as written.

An exam-shaped example

A learner has mastered handwashing with a partial physical prompt at step four of the task analysis. During the last five sessions the learner has completed step four independently on every trial. What should the technician do?

The tempting answer is to keep prompting because it is working. The correct answer is to follow the program's fading criteria and move to the next, less intrusive prompt level — the mastery criterion has been met, and continuing to prompt maintains dependency.

Note the phrasing, though: follow the program. If the program does not specify a fading criterion, the answer becomes contacting your supervisor, not inventing one. Prompt fading schedules are programmatic decisions.

A short checklist for real sessions

• Do I know the current prompt level for this target, and the criterion for reducing it?
• Am I waiting long enough before prompting?
• Am I reinforcing independent responses more than prompted ones?
• Is my data recording prompt level, not just accuracy?
• Have I reported to my supervisor if the learner is stuck at a level for several sessions?

Answer those five honestly and prompt dependency largely takes care of itself — and the exam questions on this topic stop being tricky.