Errors of Omission and Commission in ABA

In applied behavior analysis (ABA), a behavior plan only works when staff run it the way it was written. When they drift, that drift usually takes one of two shapes. They either leave out a step, or they add a step that was never called for. These two mistakes have names. They are errors of omission and errors of commission.

Both types lower treatment integrity, which is how closely staff follow the plan as written. But they do not cause the same harm. One tends to slow down learning. The other can make problem behavior worse. Knowing the difference helps you spot the more dangerous mistake first.

What each error actually means#

An omission error is a missing piece. The plan asks for something, and the staff member does not do it. Maybe a prompt gets skipped. Maybe a reward that was earned never gets delivered.

A commission error is an extra piece. The staff member does something the plan does not ask for. Maybe they reward a wrong answer. Maybe they give attention during a moment that was supposed to stay neutral.

So errors of omission are failures to implement a component.

From the talk — Dr. Jessica Osos

errors of commission are the opposite. This is when the implementer adds something to the plan.

From the talk — Dr. Jessica Osos

Why the two errors are not equal#

It would be easy to treat both errors as the same problem. They are not. The research keeps finding that they hurt in different ways and by different amounts.

Omission errors tend to slow skill building. The learner still gets correct information, just less of it. Commission errors are messier. They feed the learner wrong information, so they can teach the wrong thing on top of slowing the right thing.

In the talk, Dr. Jessica Osos pointed to research finding that omission errors reliably slowed how fast learners picked up new skills.

Commission errors and problem behavior#

The stakes climb when the goal is reducing problem behavior instead of teaching a new skill. Here, adding the wrong thing can undo the whole plan.

Think about a plan built to withhold attention after a behavior. One commission error, a bit of attention at the wrong moment, can accidentally reward the very behavior you want to shrink. That single slip can matter more than several missed steps would.

This holds for plans that reward the absence of a behavior too. In that kind of plan, a reward is supposed to come only when the behavior does not happen. A commission error sends a reward at the wrong time, right after the behavior. Studies of these plans show both error types weaken the effect. Commission errors can even push the behavior above where it started. The order of the errors can matter as well. A plan that starts strong and then drifts may look different from one that drifts in slow, steady steps.

St. Peter Pipkin et al. and colleagues, they found that commission errors were actually more detrimental when treating problem behavior specifically.

From the talk — Dr. Jessica Osos

Errors during skill teaching#

When the goal is teaching, both errors slow progress, but not by the same amount. Studies that compare them side by side tend to find commission errors do more damage than omission errors.

This shows up in tact instruction, which is teaching a learner to name or label things. In these comparisons, high integrity beats both error types. When the two errors are matched against each other, the omission version usually reaches mastery faster. Mastery is the point where a skill counts as learned. A missing reward hurts less than a reward given for the wrong response.

How much drift is too much#

Perfect is not the standard. Staff are human, and some level of error is going to happen on real sessions. The useful question is where the harm starts.

Some work has mapped this out for reward-delivery errors during teaching. Errors on a small share of trials, around fifteen percent or fewer, often did not stop learners from reaching their goal. Errors are still worth reducing, though. The goal is to keep drift low enough that it stays harmless, rather than chasing a flawless number no one can hit.

Why this matters for daily practice#

Treatment integrity is not a checkbox you tick once. It is a system you build and watch over time. Errors of omission and commission give you a shared language for that system.

When you review a session, you can ask two clean questions. Did we miss a step. Did we add a step. If problem behavior is the target, the added-step question deserves your attention first. If skill building is the target, both questions matter, and the added step still tends to be the costlier one.

What the research says#

Several behavior-analytic studies have measured these two error types directly.

A 2021 study varied how often reward delivery went wrong during teaching (Bergmann, Kodak, & Harman, 2021, Journal of the Experimental Analysis of Behavior). Errors on fifteen percent or fewer of trials were unlikely to stop learners from reaching mastery.

A 2023 study compared the two errors during tact instruction with children with autism over telehealth (Tamrazi & Wiskow, 2023, Journal of Applied Behavior Analysis). High integrity reached mastery fastest. The omission condition reached mastery in fewer sessions than the commission condition in most comparisons.

A 2020 study watched teachers use timeout during play (Foreman, St. Peter, Mesches, Robinson, & Romano, 2020, Behavior Modification). Teachers rarely used timeout as written, so omission errors were common. Even so, timeout run with reduced integrity still lowered problem behavior compared with baseline, the level before treatment started.

A 2025 study tested both errors in a plan that rewards the absence of a behavior (Hronek & Kestner, 2025, Journal of Applied Behavior Analysis). Both error types weakened the effect. Commission errors sometimes pushed responding above where it sat at baseline.

Frequently asked questions#

What is the difference between an omission error and a commission error?

An omission error is a missing step. The plan asks for something and staff do not do it. A commission error is an added step. Staff do something the plan does not call for. One leaves a gap, and the other adds something that should not be there.

Which error is worse in ABA?

It depends on the goal. When you are teaching a new skill, both slow progress, and commission errors usually hurt more than omission errors. When you are reducing problem behavior, commission errors can be the bigger threat. Adding the wrong response can reward the behavior you want to reduce.

Does a small amount of error ruin a behavior plan?

Not always. Some studies show that low levels of drift may not stop a learner from reaching their goal. That includes errors on fifteen percent or fewer of trials. The goal is to keep errors low enough that they stay harmless and to catch the costly ones early.