Prompts are the extra help that gets a correct response to happen so you can reinforce it. The prompts you choose, and the order you deliver them in, shape how quickly a learner reaches independence and how many errors they make along the way.
Why it matters
A prompting plan that does not fit the learner causes predictable problems. Too little help early on produces long strings of errors, frustration, and sometimes problem behavior. Too much help for too long produces a learner who waits for you instead of responding to the natural cue. Both outcomes slow progress and cost the learner reinforcement.
Practice vignettes in this area usually describe a learner and a skill and ask which arrangement fits, or walk through a procedure and ask you to name it.
Key ideas
- A prompt is a supplementary antecedent. It is added to the natural discriminative stimulus (SD) to make the correct response more likely. The goal is always to transfer control from the prompt to the natural SD.
- Response prompts vs. stimulus prompts. Response prompts act on the learner's behavior: verbal instructions, gestures, models, and physical guidance. Stimulus prompts change the materials, such as making the correct card larger (within-stimulus) or adding an arrow (extra-stimulus). Within-stimulus prompts that exaggerate the feature that tells the stimuli apart tend to transfer control better than added cues or changes to an irrelevant feature such as size.
- Intrusiveness is a rough ranking. Physical guidance is usually considered the most intrusive, followed by modeling, gestures, and verbal prompts, but the exact order depends on the learner. Some learners respond better to a model than to words.
- Errorless teaching is a goal, not a single procedure. You arrange prompts so the learner responds correctly from the very first trials, then fade the help gradually. Most-to-least prompting, a 0-second prompt delay, and stimulus fading can all be used to keep errors low.
- Most-to-least (MTL) moves down the hierarchy across trials or sessions. You start with the level of help that reliably produces a correct response, often full physical guidance, and step down to a less intrusive prompt each time the learner meets a preset criterion.
- Least-to-most (LTM) moves up the hierarchy within each trial. You give the instruction and a brief chance to respond independently. If the learner does not respond correctly, you deliver the next, more intrusive prompt, and continue until the response occurs. This is also called the system of least prompts.
How to choose
Before you write the plan, ask:
- Is the skill new or partly learned? Brand-new skills, and learners who make many errors, often fit MTL or another errorless arrangement. Partly learned skills often fit LTM, because the learner gets a chance to show what they can already do.
- What happens after errors? If errors are followed by problem behavior or long strings of repeated mistakes, lean toward errorless teaching.
- Is there a history of prompt dependence? If the learner waits for help, LTM with a brief wait builds in independent opportunities.
- What does the learner accept? Physical guidance is intrusive. Watch for signs that the learner is withdrawing assent, and use the least intrusive prompt that still works.
- What do the data say? No procedure is best for every learner. Compare progress, errors, and prompted responses, and change course when the data call for it.
Definitions
- Prompt: an added antecedent that increases the likelihood of a correct response while the learner is still acquiring a skill.
- Errorless learning: teaching arranged so the learner makes few or no errors, by prompting immediately and fading assistance gradually.
- Most-to-least prompting: beginning with the most assistive prompt needed and fading to less intrusive prompts as the learner meets criterion at each level.
- Least-to-most prompting: offering an independent opportunity on each trial, then adding increasingly intrusive prompts only as needed.
- Prompt dependence: responding only when a prompt is given, even after the skill is in the learner's repertoire.
Examples from a session
Matching the procedure to the learner
- MTL for a new motor chain: Kofi, age 4, is learning to put on his coat at preschool and has none of the steps yet. His BCBA writes a plan that starts with full physical guidance, then moves to partial physical prompts, then to modeling, each after two consecutive sessions with all steps correct at the current level.
- LTM for a skill that is partly learned: Rosa, age 16, can already complete most steps of a laundry routine at home. On each step, her RBT waits about 4 seconds, then gives a verbal prompt, then a gesture, then a model if needed. Most days Rosa responds after the wait or the first prompt.
- Errorless arrangement for a discrimination: Eli, age 7, has a history of crying and leaving the table after errors. His team uses an immediate prompt on every trial and plans to fade it with a prompt delay.
- Non-example of MTL: Nia, age 10, has received full physical guidance on every hand-washing step for three months, with no criterion for stepping down. Starting with the most help is not MTL; without planned fading it is constant prompting.
Common mistakes / exam traps
- Assuming errorless means "no fading." Errorless teaching only works if you have a plan to fade the prompt; without one, you risk building prompt dependence.
- Skipping levels in MTL without data, or staying at one level long after criterion is met.
Exam trap
If the vignette says the learner first gets a chance to respond on their own and help increases only after no response or an error, it is least-to-most. If help starts high and decreases as the learner meets criterion, it is most-to-least. MTL fades across trials or sessions; LTM escalates within a single trial.
Exam trap
LTM vs. time delay: both include a wait, so look at what follows it. In time delay, one controlling prompt follows a delay that starts at 0 seconds and is held constant or lengthened to fade that prompt. In LTM, the wait is a fixed pause before each next, more intrusive level of a hierarchy.
Quick check
A BCBA chooses most-to-least prompting for a new skill. Which risk should the plan address most directly?