Error Correction in ABA: 5 Steps & 4 Common RBT Mistakes

If a learner responds incorrectly during an Applied Behavior Analysis (ABA) session, the therapist running the session has to quickly decide whether to correct the response right away or let it go and move on to the next trial.

That decision is what error correction in ABA is all about, and the research behind it is consistent: catching a wrong response early keeps it from becoming the learner's default answer.

The trick is having every Registered Behavior Technician (RBT) make the same correction decisions consistently. It’s why Board Certified Behavior Analysts (BCBAs) write detailed error correction plans into treatment programs. Here are some common correction techniques, plus a breakdown of the mistakes that have two RBTs running the “same” procedure differently.

What is error correction in ABA?

Error correction in ABA is a set of steps that an RBT follows immediately after a learner gives a wrong answer. The aim is to get them to the correct response before the trial ends.

A BCBA develops these steps and writes them into the program for each skill a learner will build, so every RBT running the program corrects errors the same way.

How does error correction affect skill acquisition?

Clear, immediate error correction facilitates skill acquisition by keeping the wrong answer from becoming the learner's go-to response.

When an RBT corrects the learner right after the error, the learner is more likely to give the correct response the next chance they get, research suggests.

If the mistake is repeated without feedback, it can settle in as the more familiar answer, and replacing a familiar answer takes more work than teaching the right one from the start.

A randomized trial comparing error correction to an errorless approach found that learners who received error correction produced more independent correct responses over time. Still, challenging behavior stayed low with each approach.

What are the steps in an error correction procedure?

Error correction procedures aren't all built the same way. Several named versions exist, and BCBAs pick the one that fits the learner and the skill.

A common teaching technique in ABA is discrete trial training (DTT), where a skill is broken into short, repeated instruction-and-response trials. Here’s an example of a 5-step error correction procedure during DTT:

1. Withhold reinforcement: Don't praise, nod, or give any signal that the wrong response was correct. A neutral "No" or "Try again" is enough.
2. Interrupt and redirect: Keep the pause for correction brief so the learner isn’t left sitting with the wrong answer.
3. Model or prompt the correct response: Show the learner what the right answer looks like, using the prompt level that the program calls for.
4. Reinforce the correct response: Once the learner produces the correct answer, even with a prompt, reward it.
5. Run a transfer trial: Present the same or a similar instruction again, this time without the prompt, so the learner has a chance to respond correctly on their own.

The transfer trial shows whether the learner can do the skill without help.

Error correction procedures BCBAs use often

The right error correction procedure depends on the learner and the skill being taught, but some common variations include:

• Model-lead-test: Show the answer, practice it together, then test the learner’s independent responses.
• No-no-prompt: The learner gets two independent attempts, then they’re given a prompt on the third try.
• Active student response procedures: The learner repeats or restates the correct answer instead of just watching it modeled.
• Flexible prompt fading: To encourage correct responses, the level of help changes based on how the learner is responding during that specific trial.

Which procedure works best changes from learner to learner, even when the same procedures are compared under identical conditions. And a correction procedure that works well for one skill may not work as well for the next target in the same program.

After a detailed assessment of the learner, a BCBA develops a skill acquisition plan that spells out which correction procedure to use for each target, rather than applying one across the board. 

Still, some data suggests that certain approaches may be more effective. In a 2013 study, for instance, flexible prompt fading took fewer trials and less total session time than a standard error correction procedure to reach the same mastery criterion.

What’s the difference between error correction and errorless learning?

The main difference between error correction and errorless learning is when the prompt shows up.

Errorless learning prompts the learner before they have a chance to respond incorrectly, so errors rarely happen. Error correction lets the learner respond first. The RBT only gives a prompt if that response is wrong.

Neither approach is inherently more effective. Errorless learning removes the chance to practice a wrong answer, which may be especially helpful early on, when motivation may be low and correct responses unlikely.

Error correction gives the learner more independence: more chances to respond without a prompt, which can also help with motivation.

One potential approach is to start a new target with errorless prompting, then move to error correction once the learner has made progress.

Common mistakes to avoid during error correction

The four mistakes below cover most of what goes wrong with error correction in practice.

Reinforcing the error

Even if you’ve said "No," a warm tone or a smile after an error sends a confusing signal. Accidentally reinforcing a wrong answer is particularly common during DTT, and it can slow down or even prevent skill acquisition.

Skipping the transfer trial

Without a transfer trial, which takes place without a prompt, there's no way to know if the learner can do the skill independently. Every step before it only tests whether the learner can respond with help.

Low fidelity

If one RBT gives three prompts and another gives one, there’s a low level of fidelity to the treatment plan. It means that the data won't reflect the program, only variations of it.

Vague notes

If session notes don’t contain details about responses and corrections, a BCBA reviewing the data can't tell if a stall in progress comes from the learner or from how the trial was run.

Run consistent error correction with Passage Health

Consistency is what makes error correction data useful, and that comes down to training and documentation as well as clinical skill.

Passage Health is an all-in-one platform that helps your team practice error correction the same way, across every session and RBT. Here’s how: 

• Programs designed how they need to run: BCBAs write error correction steps directly into customized plans instead of adapting a generic template. You can set permissions by role to control who sees, edits, or approves each document.
• Real-time data collection: RBTs log data in our mobile app, which automatically syncs to the platform, so there’s no re-entering data at the end of the day. Programmable session timers keep rate, interval, and duration targets on track, according to the session’s phase.
• Consistency checks built into the workflow: The platform automatically flags gaps, like a missing transfer trial, so a BCBA can catch any inconsistency before it affects reporting or reauthorization.
• Billing tied to session data: Claims draw straight from your clinical records. Save a signature once and reuse it, and set start and end dates for each funding source, so out-of-date sources are automatically filtered out.
• Clear views of progress: Customizable treatment reports and graphing pull session data together to show which procedures are working. For practice owners, distinct dashboards track utilization and performance trends across the whole team.
• Enhanced scheduling: Create and edit events on the go, right from the app. Filter color-coded calendars by billing code, certification, or care team, whether you need a view of your practice’s month or a single team member’s day.
• Simplified referrals: Our Client Intake feature tracks prospective clients through your pipeline, collects their information, and manages forms, so no one gets lost in an email thread or spreadsheet.

Passage Health is one connected system, so data stays consistent between clinical, scheduling, and billing teams. Built-in integration with Frontera AI means your staff can spend less time drafting documents and more time with learners.

We add new features every quarter, and we keep the interface simple enough for a new RBT to pick up quickly. Every new practice also gets 1:1 onboarding support.

Book a demo to see how Passage Health can help your clinical team run error correction the same way across every session.

Frequently asked questions

What does error correction mean in ABA therapy?

Error correction in ABA therapy means giving a learner feedback after a wrong response, then guiding them to the right one during a trial. A BCBA gives detailed instructions about this process as part of the treatment plan.

What are the main types of error correction procedures in ABA?

Some established error correction procedures include: model-lead-test, no-no-prompt, flexible prompt fading, and active student response. A BCBA will select a correction approach based on their assessment of the learner's skill level, including the speed of responses, as well as the type of skill being taught.

Why is error correction important in ABA therapy?

Error correction is important in ABA therapy because an unrecognized mistake can turn into the learner's default response. Flagging errors right away keeps them from settling in and creates more opportunities for the learner to answer correctly.

How quickly should an error be corrected during a session?

An error should be corrected right after it happens. The longer the gap between the mistake and the feedback, the harder it is for the learner to connect the two.

References

Cariveau, T., La Cruz Montilla, A., Gonzalez, E., et al. (2019). A review of error correction procedures during instruction for children with developmental disabilities. Journal of Applied Behavior Analysis, 52(2), 574–579. https://doi.org/10.1002/jaba.524 

Carroll, R. A., Joachim, B. T., St. Peter, C. C., et al. (2015). A comparison of error-correction procedures on skill acquisition during discrete-trial instruction. Journal of Applied Behavior Analysis, 48(2), 257–273. https://doi.org/10.1002/jaba.205 

Carroll, R. A., Kodak, T., & Fisher, W. W. (2013). An evaluation of programmed treatment-integrity errors during discrete-trial instruction. Journal of Applied Behavior Analysis, 46(2), 379–394. https://doi.org/10.1002/jaba.49 

Foran-Conn, D., Hoerger, M., Kelly, E., et al. (2021). A comparison of most to least prompting, no-no prompting and responsive prompt delay procedures. Behavioral Interventions, 36(4), 1024–1041. https://doi.org/10.1002/bin.1808

Leaf, J. B., Cihon, J. H., Ferguson, J. L., et al. (2020). Comparing error correction to errorless learning: A randomized clinical trial. The Analysis of Verbal Behavior, 36(1), 1–20. https://doi.org/10.1007/s40616-019-00124-y 

Leaf, J. B., Leaf, R., Taubman, M., et al. (2014). Comparison of flexible prompt fading to error correction for children with autism spectrum disorder. Journal of Developmental and Physical Disabilities, 26(2), 203–224. https://doi.org/10.1007/s10882-013-9354-0