Prompt Hierarchy in ABA: The Levels, Most-to-Least vs Least-to-Most
Key Takeaways
- 01A prompt hierarchy ranks help from most to least so you can say exactly how much support a learner needed and then fade it on purpose.
- 02In a session, you wait roughly three to five seconds at each level before moving up. Then you record which prompt level the trial needed, such as partial physical.
- 03Time delay isn't a rung on the ladder. It keeps the prompt the same and changes when it arrives. Prompt dependency is usually a teaching artifact, so set a written fading criterion before teaching starts.
- 04G.7 covers selecting and evaluating prompting procedures and G.8 covers designing the fading. Both sit in Behavior-Change Procedures, the largest domain at 25 questions.
Helping someone is easy. Helping them in a way you can later stop is the actual skill, and a prompt hierarchy is the tool for it. The point of ordering prompts is not tidiness. It is that an ordered list lets you say exactly how much support a learner needed today, compare it to last week, and prove the support is coming down.
The ladder, from most help to least
These are response prompts, which act on the person:
- Full physical. Hand-over-hand through the whole movement.
- Partial physical. A touch at the elbow or wrist to start or steer.
- Model. You do it, then the learner copies.
- Gestural. A point or a tap toward the right item.
- Verbal. Anything from the full instruction to a first-sound cue.
- Independent. No prompt. This is the target.
Stimulus prompts are a separate family and they act on the materials rather than the learner: moving the correct card nearer, making it larger, coloring it, exaggerating the difference between it and the distractor. They fade by shrinking that alteration until the array is plain, which is why the fading procedure has its own name, stimulus fading.
One honest caveat on the ordering above. Intrusiveness is a rule of thumb, not a law, and it varies by learner. Verbal prompts sit low on the ladder yet are famously the hardest to remove, because a learner can wait out a silence for a very long time if a spoken cue has always eventually arrived. Plenty of clinicians deliberately rank verbal prompts as more intrusive than a gesture for that reason.
Which direction you run it changes what you are optimizing
| Most-to-least | Least-to-most | |
|---|---|---|
| Starts with | The strongest prompt that guarantees success | No prompt, then escalating help |
| Errors | Close to none, by design | Expected, and informative |
| Best for | New skills, safety-critical steps, learners with a history of failure | Skills partly in the repertoire, when you want to know what they can already do |
| The risk | Prompts linger longer than they need to | Repeated errors get practiced |
Most-to-least is the backbone of errorless teaching. The claim is not that errors are morally bad, it is that a learner who has never performed the wrong version has nothing to unlearn, which matters most when the wrong version is dangerous or when the learner arrives discouraged.
A worked least-to-most trial
Target: the learner puts a coat on a hook after coming in. Wait roughly three to five seconds at each level before moving up.
- Give the natural instruction and wait. Nothing happens.
- Gesture toward the hook. The learner looks but does not move.
- Model it: hang your own coat, then hand theirs back.
- Partial physical: touch the elbow to start the reach. The learner completes the hang.
- Reinforce the completed response, and record that the trial needed a partial physical prompt.
That last line is the whole reason for the hierarchy. "Needed partial physical, down from full physical last Tuesday" is progress you can see. "Did it with help" is not data.
Time delay is not a rung on the ladder
It is frequently taught as though it were one, and it is not. Time delay keeps the prompt the same and changes when it arrives. Constant time delay inserts a fixed pause, often three seconds, between the instruction and the prompt on every trial. Progressive time delay stretches that pause across trials, from zero seconds to one, to two, to four.
What that buys you is the transfer of control: the learner starts responding to the instruction because it now reliably beats the prompt to the punch. It pairs naturally with either direction and is one of the cleanest ways to get out of a stuck hierarchy.
Prompt dependency, and why it is usually the adult's doing
A learner who waits for help rather than responding has learned something perfectly sensible: waiting works. The pattern is a teaching artifact, not a trait, and it comes from prompting at the same level for weeks because that level produces a correct response and correct responses feel like progress.
The tells are worth knowing. The learner performs at a stable level and stops improving. They begin correctly and then stall, watching you. They perform for one staff member and not another, which usually means the two are prompting at different levels. The fix is a written fading criterion set before teaching starts, so the decision to drop a level is made by the data rather than by whoever is in the room.
Where this sits on the exam
Two tasks in the 6th Edition Test Content Outline cover this and they are easy to conflate. G.7 is selecting and evaluating prompting procedures, naming errorless, most-to-least and least-to-most directly. G.8 is designing the fading, including prompt delay and stimulus fading. Both sit in Behavior-Change Procedures, the largest domain on the exam at 25 questions.