Trials to Criterion in ABA: What It Measures and Why

Trials to criterion counts how many teaching trials a learner needs to meet a mastery goal. Learn how BCBAs use it to judge teaching speed and goal design.

Key takeaway

Trials to criterion is the number of teaching trials a learner needs before meeting a set mastery goal. It tells you how fast a skill was learned, not just whether it was learned.

Rad N Bad: Your Programming Sucks, Let's Make It Better — Individualized Goals, Layered Skills, and Generalization by Design
Watch the full CEU recording

Rad N Bad: Your Programming Sucks, Let's Make It Better — Individualized Goals, Layered Skills, and Generalization by Design

Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA · 0.5 CEU · 33 min
Watch on openceu.com →

Free learning path on this topic

PDA and Demand Avoidance

B. Kuereine Gray's three PDA talks plus Clelia Sigaud on the logic of anxiety. What PDA is and isn't, how to collaborate with caregivers, how to handle complex profiles and replacement behaviors, and why demand-avoidant learners respond the way they do.

Start the path →4 courses · 5 CEUs · Watch at your own pace

Experts who cover this

4 recordings on openceu cover trials to criterion.

Trials to criterion is the number of teaching trials a learner needs before meeting a set mastery goal. It tells you how fast a skill was learned, not just whether it was learned. Two learners can both reach 90% correct. One might need forty trials, and the other might need four hundred.

That gap matters for every BCBA, the clinician who designs the program, and every RBT, who runs sessions. It shapes how you compare teaching methods and how you write goals. It also shapes how you judge whether a skill is truly ready to leave the program. The clinicians below look at the measure from several angles. Some use it to catch slow teaching. Others use it to keep young learners from burning out.

What trials to criterion actually counts#

A trial is one chance to respond. In discrete trial teaching, that means one instruction, one response, and one consequence. Discrete trial teaching is a style where a skill is broken into short, repeated teaching rounds. Trials to criterion simply adds up those rounds until the learner meets the goal.

The goal itself is the mastery criterion. That is the rule that says when a skill counts as learned. A common rule is something like 90% correct across three sessions. Trials to criterion sits on top of that rule. It counts every trial it took to get there, including wrong answers and prompted answers.

In one telehealth study, researchers defined it this way. They counted the number of trials presented before learners met a criterion of three sessions above 90% correct. Each trial included the question, the picture, and whatever response followed. A wrong answer still counted as a trial.

This is why the measure is so useful for comparing methods. Percent correct shows where a learner ended up. Trials to criterion shows what the trip cost. A cheaper trip means less time, less effort, and fewer chances for frustration.

Why percent-correct graphs hide the real story#

Michael Carrero and Sean Yocum make a sharp point about the graphs most clinics send out. Most graphs that go to funders show percent correct or a frequency count. They rarely show how many trials a skill took. Carrero notes that many data platforms can already report trials to mastery. Clinicians just do not pull that report.

The problem shows up once a program finally hits its goal. When a target reaches 80%, the graph makes it look like everything worked. Carrero says the graph cannot tell you how long the climb was.

it might have taken them two months to get there, and that same learner with a different correction... It might have gotten there in four days, right?From the talk — Michael Carrero

He calls that lost time a gap the learner will never get back. The same end point can hide a fast path or a very slow one. Yocum adds that clinicians rarely check this at all.

You graph correct responses for months, but almost nobody measures how many trials it took to master a target under this correction versus another one.From the talk — Sean Yocum

The takeaway is simple. If you only graph accuracy, you cannot see waste. Trials to criterion makes the waste visible.

Using trials to mastery to choose an error correction#

Error correction is what a teacher does right after a wrong answer. It might be a model, a prompt, or a repeat of the trial. Carrero and Yocum argue that many clinics use one house style of correction for every learner. Over time, staff drift away from each other in how they run it. Nobody stops to ask if it fits the child.

Yocum names two problems with that setup. First, the best correction differs from learner to learner. A single house style will be quick for some kids and slow for others. Second, nobody checks. Without trials-to-mastery data, a slow correction looks the same as a fast one once both reach criterion.

There is a second warning sign besides slow progress. Yocum asks whether the learner starts avoiding work because the corrections feel like a grind. Carrero is direct about this. If a learner shows escape behavior right after corrections, that is not a coincidence. Escape behavior means actions that help a person get away from a task. Carrero says it is the learner telling you the correction is part of the problem.

So trials to criterion does two jobs here. It compares how fast each correction works. It also pairs with behavior data to show what each correction costs the learner.

Small criteria for small learners#

Kelly Brzak looks at trials to criterion from the goal-writing side. Her focus is toddlers and preschoolers, ages two to five. She argues for very small per-session criteria with this age group. She writes goals like three out of three or two out of two items, not eight out of ten.

If I set a minimum of three out of three items as a goal, as opposed to five out of five or four out of five or eight out of 10, I'm going to let my little people be more successful.From the talk — Kelly Brzak

Brzak says the payoff spreads to everyone. The child is less likely to burn out on her. The child succeeds with far fewer responses. She and her technicians succeed too. One sample goal from her talk reads like this. The client will expressively identify a minimum of three of three items per session across six consecutive sessions.

She is honest that this did not come naturally. She admits she used to push back on it. Three out of three never felt like enough to her. She thinks eight out of ten was trained into her because 80% is easy to figure. Her experience with young children changed her mind.

And anytime that I've been with little ones and I've tried to push past the two or three, that's when I start getting into task refusal and elopement and kids wanting me to leave.From the talk — Kelly Brzak

Elopement means a child running away from the work area. For Brzak, a high trial demand is a direct path to that kind of behavior.

The hidden math behind an eight-out-of-ten goal#

Brzak credits Dr. Patrick McGreevy, who developed Essentials for Living, with opening her eyes. She saw him speak many times while she was learning behavior analysis. His point was about the total trial load hidden inside a goal.

Take a common goal. Susan will tact eight out of ten colors across five consecutive sessions. A tact is a label, like naming a color you see. Brzak does the math out loud. That goal asks Susan to tact 50 colors before she can master it.

Brzak points out that most clinics do not stock anything close to fifty different colors. Unless the clinic has an unusually wide array of markers, Susan ends up naming the same few colors over and over. Brzak says this is where hoop jumping starts. Hoop jumping means the child performs just to get through the task, not because the skill is growing.

Her advice is to be mindful of trials to criterion when you pick a goal for young children. The number of trials a goal demands is part of the goal. It is not a side detail. A goal can look modest on paper and still ask a toddler for a huge pile of responses.

Writing goals that build in variety#

Brzak pairs small criteria with goal wording that forces variety. One rewrite from her talk reads like this. Susan will tact two out of two colors across eight total colors. The child only answers twice per session. But the goal still spans a wider set of colors over time. That keeps sessions fresh and lowers the chance of hoop jumping.

She also recommends running records. A running record is a log of which targets were already presented. It helps staff rotate targets so the child sees variety. Brzak notes it helps the treatment team stay varied too.

She adds a caution about colors in particular. Some two- and three-year-olds are not yet ready to tell colors apart. For them, she suggests play-based activities rather than traditional trials for that skill.

Finally, she pushes hard on caregiver involvement. She asks clinicians to include parents and caregivers in programming and in running trials. She says she cannot say that enough. She warns against handing parents a goal sheet just to sign. One idea is to discuss a different goal area each session. Another is to let the parent pick from three or four areas. She also reminds her audience to plan for generalization from the start.

Hitting the number is not the same as mastery#

Sean Yocum and Michael Carrero push the conversation past the trial count. Their concern is what happens after a learner meets criterion. A typical goal might call for 80% across three sessions. The target then moves to a maintenance schedule. When it comes back up, the learner often gets it wrong. Carrero says that means the skill was never mastered in the first place.

Yocum describes the usual setup bluntly.

the learner performs this skill accurately with the therapist present in a controlled environment, using familiar materials under ideal motivation at a predefined percentage. Let's just call it what it is. That's performance art.From the talk — Sean Yocum

He points out that perfect conditions do not exist in real life. The two hosts then lay out four conditions that should all be met. The skill occurs across people, including parents. It survives setting changes. It maintains over time. And it contacts natural reinforcement, meaning the payoff that happens in daily life.

Carrero puts the reinforcement piece first. He says mastery means you have faded reinforcement to the normal level for that setting. His example is asking where the bathroom is. The reward is not praise for asking. The reward is getting to use the bathroom.

If a skill misses any of the four, Yocum says it is still in training. This echoes a long-standing warning against treating generalization as a later phase (Stokes & Baer, 1977).

Applying a criterion to staff training#

Dr. Jessica Osos brings the same logic to supervision. Her topic is training RBTs to help with assessments. Her point is that you must define what "ready" looks like before you delegate the work. Then you need a way to confirm it.

She says a supervisor needs a mastery criterion they would feel comfortable defending to an auditor. Saying "I felt like she was ready" will not hold up. Her example of a defensible standard is concrete.

something like 90 to 100 of procedural fidelity across three consecutive role plays followed by at least one in-person observationFrom the talk — Dr. Jessica Osos

Procedural fidelity means how closely someone follows the steps of a procedure as written. Osos wants this measured during behavioral skills training, a structured way to teach staff a new skill. The RBT should meet the criterion before working with a patient on that task.

She also builds in a safety net. If someone later falls outside the criterion, that should trigger a retraining flag. For staff billing assessment time, she suggests ongoing fidelity checks. This mirrors the clinical side closely. You set a clear criterion, count the practice it takes, and keep checking after the goal is met.

Where the experts pull in different directions#

These clinicians agree on a lot. All of them want criteria that are clear and tied to real outcomes. Both Brzak and the Rad N Bad hosts care about generalization and caregiver involvement. Both also link poor teaching design to escape behavior.

The real difference is which way each one wants to move the criterion. Brzak worries that standard goals ask too much of very young learners. Her fix is to shrink the per-session demand to two or three responses. She wants fewer total trials so toddlers stay successful and engaged.

Yocum and Carrero worry that standard goals ask too little of the right things. To them, 80% across three sessions is a low bar dressed up as mastery. Their fix is to add conditions: other people, other settings, time, and natural reinforcement. That approach likely means more total practice before a skill is called mastered, not less.

Osos sits closer to the stringent side for staff. Her example calls for 90 to 100% fidelity plus an observation before an RBT works alone.

These views can coexist. A small per-session criterion can still be spread across settings and people. But the tension is real in practice. A team has to decide where its main risk lies. Is it burnout from too many trials, or false mastery from too few conditions?

For more on Brzak's approach, see Child Development Deep Dive: Early Childhood (2-5 year olds). For more on staff training standards, see Delegating with Confidence: Training and Supervising RBTs to Support Assessments.

What the research says#

Researchers use trials to criterion as a standard yardstick for teaching efficiency. A few studies show how it is applied.

Lindgren and colleagues compared discrete trial teaching delivered by telehealth and in person. Their main measure was trials to criterion for each stimulus set, a group of items taught together. They defined it as the trials presented before learners met three sessions above 90% correct (Lindgren et al., 2024).

Lindgren, N. A., Higbee, T. S., Osos, J. A., Nichols, B., & Campbell, V. E. (2024). Comparing the effectiveness of discrete trial training delivered via telehealth and in‑person on skill acquisition. Behavior Analysis in Practice, 17, 783–795. https://doi.org/10.1007/s40617-023-00855-4

Roncati and colleagues taught intraverbal skills to three children with autism. An intraverbal is answering a question or filling in words in conversation. More exposure to one prompt type made that prompt more efficient, measured by trials to criterion (Roncati et al., 2019).

Roncati, A. L., Souza, A. C., & Miguel, C. F. (2019). Exposure to a specific prompt topography predicts its relative efficiency when teaching intraverbal behavior to children with autism spectrum disorder. Journal of Applied Behavior Analysis, 52(3), 739-745. https://doi.org/10.1002/jaba.568

Hranchuk and colleagues found that modeling a response before a trial needed fewer trials to criterion than standard trials alone. This only held for children who already had the naming repertoire. That means the ability to learn a new word's meaning after just one exposure (Hranchuk et al., 2019).

Hranchuk, K., Douglas Greer, R., & Longano, J. (2019). Instructional Demonstrations are More Efficient Than Consequences Alone for Children with Naming. The Analysis of Verbal Behavior, 35(1), 1-20. https://doi.org/10.1007/s40616-018-0095-0

Finally, Fienup and Brodsky found that a less stringent mastery criterion hurt effectiveness with college learners (Fienup & Brodsky, 2017). That finding is a useful check when deciding how small a criterion can go.

Fienup, D. M. & Brodsky, J. (2017). Effects of mastery criterion on the emergence of derived equivalence relations. Journal of Applied Behavior Analysis, 50(4), 843-848. https://doi.org/10.1002/jaba.416

FAQ#

What is trials to criterion in ABA? It is the count of teaching trials a learner needs to meet a mastery goal. Every trial counts, including errors and prompted answers. It shows how efficient teaching was, not just whether the skill was learned.

How is trials to criterion different from percent correct? Percent correct shows accuracy at one point in time. Trials to criterion shows the total effort it took to reach the goal. Two learners can hit the same percent after very different amounts of practice.

Should young children have smaller mastery criteria? Kelly Brzak recommends two out of two or three out of three items per session for toddlers. She finds larger demands lead to task refusal and elopement. Other clinicians stress that meeting any number still needs to hold up across people, settings, and time.

Turn this topic into a CEU

You just studied this. Now get credit for it.

Watch Rad N Bad: Your Programming Sucks, Let's Make It Better — Individualized Goals, Layered Skills, and Generalization by Design with Sean Yocum, MA, BCBA, LBA & Michael Carrero, BCBA, LBA and earn 0.5 free BCBA CEUs. Detailed certificate, delivered the moment you finish.

Watch and earn the CEU →Free account · No card · detailed cert

Keep going with a free learning path

PDA and Demand Avoidance

B. Kuereine Gray's three PDA talks plus Clelia Sigaud on the logic of anxiety. What PDA is and isn't, how to collaborate with caregivers, how to handle complex profiles and replacement behaviors, and why demand-avoidant learners respond the way they do.

Start the path →4 courses · 5 CEUs · Watch at your own pace