TL;DR: Prompts help learners produce correct responses, but only systematic prompt fading builds independence. Choosing the right hierarchy for the right learner — and then deliberately removing it — is one of the highest-leverage clinical skills in ABA.
Prompting is foundational to ABA. Every DTT trial, every NET session, and every skill acquisition program depends on it. Yet prompting is also one of those concepts that trainees learn in a textbook and then immediately underuse, overuse, or apply inconsistently in session.
The goal of prompting is not to help a learner succeed in the moment. It’s to transfer stimulus control from the practitioner to the natural environment. That distinction shapes every decision in a prompt hierarchy.
The Four Core Hierarchy Formats
Each format has a different risk profile and a different best use case.
Most-to-Least (MTL)
You start with the most intrusive prompt — often full physical guidance — and gradually fade toward less intrusive levels as the learner demonstrates mastery. MTL is errorless by design: the learner is set up for correct responses from the first trial.
Best for: New skills where errors are likely to be reinforced, where frustration tolerance is low, or where correct form is safety-critical from the start.
Risk: Prompt dependency. If fading is not systematic and data-driven, learners can require physical or full vocal prompts indefinitely.
Least-to-Most (LTM)
You start with the least intrusive prompt — often no prompt at all, just the SD — and add prompts only after the learner fails to respond within the specified wait interval. LTM promotes independence by design.
Best for: Learners with an established prompt-fading history, or skills where you suspect some existing ability. Also a good default for maintenance and generalization probes.
Risk: Errors occur. LTM requires waiting for a failed response, which can be frustrating for learners with low frustration tolerance and increases practice of incorrect responses.
Graduated Guidance
A hands-on approach where the practitioner follows the learner’s movements — providing full support only where resistance or inaccuracy occurs and releasing contact as the learner moves independently through the step.
Best for: Multi-step motor chains (dressing, handwashing, hygiene routines) where the level of support needed varies across steps.
Risk: Requires high clinical precision. Under-guiding and over-guiding are both common fidelity errors, especially mid-chain.
Time Delay
You insert a pause between the SD and the prompt, creating a window for the learner to respond without prompting. Delay can be constant (same interval every trial) or progressive (interval increases across sessions as independence grows).
Best for: Learners who have partial acquisition of a skill but are waiting for a prompt rather than attempting independently. Pairs well with LTM once initial skills are acquired.
Risk: Ineffective if the delay is too short (no real independence opportunity) or too long (learner disengages or becomes frustrated before the prompt arrives).
Choosing the Right Hierarchy
The decision is not about which method you prefer or which feels easiest to implement — it’s about what the learner’s history tells you.
| If the learner… | Consider… |
|---|---|
| Has high error rates and low frustration tolerance | MTL to build an early success history before fading |
| Has some ability but consistently waits for prompts | Time delay to probe independent responding |
| Is learning a multi-step motor chain | Graduated guidance with step-specific support levels |
| Is generalizing a skill they’ve already mastered | LTM to promote fluency across novel contexts |
One underappreciated point: the hierarchy you use for acquisition may not be the right one for maintenance. A learner who reaches mastery at vocal prompt level has not generalized — they’ve built dependence at a new level. Document the prompt level at acquisition, plan the fade explicitly, and treat independence as the criterion, not just accuracy.
Prompt Fading Is the Actual Skill
Most prompting errors happen not in delivery but in fading. Common patterns that erode gains:
- Prompt creep: Adding prompts back after they’ve been faded, usually triggered by errors, without a systematic decision rule. Every team member re-prompts slightly earlier. Within a few sessions, the program has regressed.
- Level-skipping: Jumping from full physical to vocal without working through intermediate steps, resulting in errors and regression at the skipped levels.
- Intuition-based fading: Moving across levels because the session “felt good” rather than because mastery criteria were met. If the criterion isn’t written down, it won’t be applied consistently across staff.
A fading plan should specify:
- The starting prompt level
- The mastery criterion for each level drop (e.g., 3 consecutive correct unprompted responses)
- The error response procedure (e.g., return to previous level and reset the criterion)
- Who makes the fading decision and on what data schedule
The more explicit this is in the program, the more consistent the implementation across everyone on the team.
Why This Is Hard to Learn Without Reps
Prompt hierarchies are easy to describe in training and surprisingly hard to execute in session. The clinical decisions happen in real time: how long to wait before adding a prompt, whether a partial response warrants prompting up or down, how to manage a learner who’s testing the delay rather than attempting independently.
These decisions don’t come from knowing the theory. They come from having practiced the decision-making under conditions that resemble real sessions — with a learner who doesn’t always cooperate, in a time frame that doesn’t allow for extended reflection. That’s a skill built through reps, not reading.
How Kipr Helps
Kipr is a simulation platform for ABA practitioners. Instead of working out prompting decisions exclusively in live sessions with real clients, BCBAs and RBTs can practice delivering prompt hierarchies, managing fading decisions in-session, and adjusting in real time — with AI-simulated learner personas that reflect realistic response patterns.
This creates space to build the kind of procedural fluency and in-session judgment that comes from reps — before those decisions affect actual clients. For BCBAs supervising a team, Kipr can also give your technicians targeted simulation practice on prompting consistency and fading before they run these procedures in live sessions.
Join the Kipr waitlist for early access.