Cue levels in speech therapy: a prompt hierarchy for data
A cue level records how much help a student needed to produce a response. Writing it down with each trial lets you show progress before the percent correct moves: a student who needed a model in September and only a reminder in December has made real progress, even at the same accuracy.
The four common levels
| Level | What it means | Articulation example | Language example |
|---|---|---|---|
| Independent | No help beyond the instruction and the materials. | Names the picture of the sun with a correct /s/. | Answers "Why did the boy cry?" about a story with no prompting. |
| Minimal | One light cue: a reminder, a gesture or an expectant pause. | "Remember your snake sound," or pointing to your mouth. | Pointing back to the picture: "Look at his face." |
| Moderate | More direct or repeated cues that narrow down the response. | A picture of tongue placement plus a verbal reminder. | A choice of two: "Did he cry because he fell, or because he was hungry?" |
| Maximal | The response is shown or done with the student. | A full model to imitate: "Say sun." | The answer is modeled for the student to repeat. |
Districts and graduate programs define these levels differently. Some go by the number of cues (one cue is minimal, two or three is moderate); others go by the kind of cue. The exact definitions matter less than using the same ones every session, and writing them on your data sheet so anyone who takes data for you scores the same way.
Kinds of cues
A cue type describes what you did; a cue level describes how much help it was. Most goals are written against the level, but knowing which types work for a student is worth a note.
- Verbal: a reminder, question or instruction. "Slow down." "What sound does it start with?"
- Visual: pictures, written words, a mirror, a visual schedule.
- Gestural: pointing, or a hand signal for a sound or a strategy.
- Tactile: touch cues for placement or rate.
- Phonemic: the first sound of the word. "It starts with /b/…"
- Semantic: a hint about meaning. "You sit on it."
- Model: the full correct response for the student to imitate.
Least-to-most or most-to-least?
Both are standard prompting approaches, and they answer different questions.
- Least-to-most: start with no help and add the next level only after an error or no response. It shows what the student can do alone, so it suits probes and skills that are partly learned.
- Most-to-least: start with plenty of support and fade it as the student succeeds. It keeps errors down while a skill is new.
Either way, record the level at which the response was actually produced.
Recording cue levels
- Mark each trial with a result and a level. On paper, a box per trial with
+or−and a letter is enough:+I,+Min,−Mod. Put the key at the bottom of the sheet. - Decide which levels meet the goal. If the goal says "given minimal cues," correct responses at the independent and minimal levels meet it; responses that needed moderate or maximal support don't.
- Report the spread, not just the total. "Of 40 trials, 22 were independent, 12 needed minimal cues and 6 needed moderate cues." That sentence shows support fading even when accuracy is flat.
Goals written around cue levels
A goal like "Sam will produce /r/ in words with 80% accuracy given minimal cues across three consecutive sessions" needs three things in the data: percent correct, the cue level of every trial, and results session by session. (The percent accuracy guide covers how to calculate and combine the percentages.)
Goals written around independence, such as "…will request help independently in 80% of opportunities," measure something different: the share of opportunities where the student needed no cue at all. Track those as percent independent, not percent correct.