Error correction is one of the most consequential decisions a behavior therapist makes during a discrete trial training session. When a learner responds incorrectly — or doesn’t respond at all — how you respond in the next few seconds shapes what they learn next and what becomes habit.
Yet error correction is one of the areas where implementation fidelity drifts fastest. RBTs often default to whatever feels natural in the moment, usually some version of “just try again.” Without a consistent, intentional procedure, you can end up reinforcing inconsistency instead of building fluency.
This post covers the four main error correction strategies used in structured ABA teaching, when each is appropriate, and what to watch for when implementing them.
Why Your Error Correction Procedure Matters
When a learner makes an error in a discrete trial, you have a few seconds to deliver a correction before the window closes. The goal is not to penalize the error — it is to transfer stimulus control from the incorrect response back to the correct one, without turning the session into a correction marathon.
The strategy you choose affects:
- Acquisition rate — some correction procedures slow learning if overused
- Accidental reinforcement of errors — pausing too long or reacting inconsistently can inadvertently reinforce incorrect responding
- Prompt dependency — poor error correction teaches learners to wait for a prompt rather than respond independently
The Four Main Error Correction Strategies
1. No-No Prompt (NNP)
How it works: Present the SD. If the learner makes an error, deliver a neutral error marker (“no” or a similar brief signal), re-present the SD, and wait. If they error again, re-present and deliver a prompt at the appropriate level. Reinforce correct responding at any point.
When to use it: NNP works best for learners who are close to mastery on a target and making occasional errors despite having demonstrated the skill. It is most appropriate when accuracy is already high and you want to increase independence without introducing a more complex correction sequence.
When to avoid it: NNP is a poor fit for early acquisition. If the learner does not yet have the skill fluently, a string of “no’s” without meaningful transfer teaches frustration more than behavior. It can also create a negative association with the SD if applied too broadly.
2. 4-Step Error Correction
How it works: Present the SD → learner errors → immediately prompt the correct response using the most intrusive level appropriate for the learner → run a distractor trial (a mastered, different item) → re-present the original SD → if correct, reinforce (typically with a thinner schedule than an independent correct).
When to use it: The 4-step procedure is the most widely used error correction format in discrete trial training for good reason. It re-establishes correct responding without ending on an error, uses a distractor to prevent massed error correction, and builds in a prompt-to-independence transfer on re-presentation.
When to avoid it: In sessions with very high trial density, 4-step can break pace significantly. Some learners also learn to make errors on purpose to access the distractor trial — if you notice a pattern of errors on otherwise mastered items, review your ABC data and consider adjusting.
3. Transfer Trial (Errorless Correction)
How it works: Present the SD → learner errors → immediately deliver a most-intrusive prompt without a verbal error marker → reinforce the prompted correct response at a reduced level → fade the prompt across re-presentations.
When to use it: Transfer trial is appropriate when errorless learning is the intended approach — typically for learners in early acquisition, those with a history of reinforced incorrect responding that you are extinguishing, or learners who show significant distress at error feedback.
When to avoid it: Without careful attention to prompt fading, errorless correction can deepen prompt dependency. If you use transfer trial, ensure your prompting schedule actively fades across trials and that you are collecting data to confirm independence is increasing, not decreasing.
4. Simple Re-Presentation
How it works: Present the SD → learner does not respond (or responds incorrectly) → allow the inter-trial interval → re-present the SD, often with a slight increase in prompt level.
When to use it: Simple re-presentation is best suited to no-response errors — where the learner did not respond at all rather than giving an active incorrect response. It is also common in massed-trial formats with high-frequency targets where maintaining pace is essential.
When to avoid it: For active errors (a clearly wrong response), simple re-presentation without a correction signal or prompt can inadvertently reinforce the error by following it with another trial opportunity. Use 4-step or transfer trial for active errors.
Choosing the Right Procedure
No single error correction procedure is universally correct. Selection depends on the learner, the target, and the phase of acquisition:
| Situation | Recommended Approach |
|---|---|
| Early acquisition, new skill | Transfer trial (errorless) or 4-step |
| Skill near mastery, occasional errors | No-no prompt |
| Active error (wrong response) | 4-step or transfer trial |
| No-response error | Simple re-presentation or 4-step |
| Significant distress at error feedback | Errorless / transfer trial |
| Suspected intentional erroring | Review reinforcement schedule; reconsider 4-step distractor |
Document which procedure is specified for each target in the learner’s program book or BIP. Consistency across therapists is essential — if staff are using different procedures across sessions, skill generalization is harder and data interpretation is confounded.
Common Fidelity Drift Points
These are the most common ways error correction breaks down in real sessions:
- Reinforcing after an error before delivering the correction — even a smile or softened tone counts as reinforcement
- Skipping the distractor trial in 4-step — the most common shortcut; without it, you are massing error trials rather than distributing them
- Re-presenting at the same prompt level used during acquisition — correction prompts should start more intrusive than your current acquisition level and fade from there
- Inconsistency across staff — if the BCBA uses 4-step and the RBT uses no-no prompt, the data will be confounded and the learner may perform differently with different therapists
How Kipr Helps
Knowing error correction procedures from a training manual is different from executing them fluently in the middle of a session with a real learner. Under time pressure — especially for a new RBT in the first weeks on the job — the natural response is to default to whatever feels easiest, which is rarely the most appropriate choice.
Kipr lets you practice error correction procedures in simulation before applying them with real clients. You can run discrete trial scenarios with AI client personas and rehearse delivering 4-step corrections, transfer trials, or no-no prompts consistently — without the stakes of a live session. When a correction procedure becomes habitual through deliberate practice, it holds up under pressure far better than something you’ve only read about.
Join the Kipr waitlist for early access.