Criterion-Referenced Assessment in ABA Explained

Criterion-referenced assessment judges a skill against a set standard. You decide ahead of time what mastery looks like. Then you check if the learner meets that mark. It does not compare the learner to other people at all.

This matters for BCBAs, RBTs, teachers, and parents. It tells you if a person can do a real skill. It does not just say how they rank against peers. That focus on the skill itself makes it useful for planning.

Set the criterion from the skill#

Mark Malady explains where the standard comes from. You look at what a skill truly needs to be done well. From that, you set a clear mastery mark for each part.

criterion reference that you identified a logical consistency of what was required for mastery of those items. And then you identified the mastery criteria for them. From the talk. Mark Malady

So the standard is built on logic, not on a crowd. You break the skill down and ask what mastery requires. Then that becomes the bar the learner aims for.

One target, one performance mark#

Malady goes deeper on how the mark works. Each target in your assessment gets its own performance criterion. If the learner hits it, they have that skill.

criterion-based assessments, on the other hand, is about a performance criteria that is set for each individual target inside of your assessment. And the idea being that if you can perform to that, then you have mastered over that skill set. From the talk. Mark Malady

This keeps the focus tight and clear. You are not guessing if a child is "behind." You are checking one skill against one plain mark.

How it differs from norm referencing#

The other main style is norm referencing. That style compares a person to a group of peers. It asks how this child stacks up against other children the same age.

Malady lays out the difference in simple terms. One style looks at the person's own goals. The other looks at what the crowd does.

Relative evaluation skills are compared to the person's goals and values. So this is contrasted with normative comparison where we might look at, hey, this person is eight years old. What other eight year olds do? From the talk. Mark Malady

Both styles have a place. But criterion referencing keeps the client at the center. It asks what this learner needs, not what the average learner does.

Why BCBAs lean on it#

This style fits ABA work well. Our goals are usually real, useful skills. We want a child to ask, wait, share, or dress. A set standard tells us plainly if that skill is there.

It also makes progress easy to see. You know the mark before you start teaching. Each week you check how close the learner is. There is no need to compare the child to a class of peers.

This keeps the plan personal for each learner. Two children can have very different goals and needs. Each one is judged against their own set of marks. That respects where each learner actually is right now.

The big risk: teaching to the test#

Criterion-referenced tools have one main trap. You can start to teach only the test items. Then the score goes up but the real skill does not.

they identify that one potential misapplication is that you can teach to the test. So you need to be really mindful that your programming is not just based on the item inside of your test. From the talk. Mark Malady

The fix is a quality check on your programming. Ask if the skill still helps in real life. The test is a guide, not the goal.

A simple test is to watch the skill outside the session. Can the child use it at home or on the playground? If yes, your teaching is on track. If not, you may have taught only the test items. Malady covers this trap more in genArete: Learner-Centered Skill Assessment.

Be clear about where the mark came from#

Malady raises a fair concern about many tools. Their standards are often vague. You may not know who set the mark or why.

We need to be clearer on who the benchmarked learners were to set the criterion and what that criterion allowed the performance to be used to do. Right now, most of them are extremely vague or ambiguous on where the criterion came from or what its purpose was. From the talk. Mark Malady

This is a call for better care in assessment. Know who the standard was built on. Know what a passing score is meant to prove. A clear purpose makes the whole tool stronger.

What the research says#

The evidence for these tools is mixed. One systematic review looked at six common ABA assessments. It found conflicting support for their use. Many of the tools people rely on had thin proof of reliability and validity (Validity and Reliability Evidence for Assessments Based in Applied Behavior Analysis: A Systematic Review). This matches Malady's warning to know where a standard came from.

Newer work tries to make planning more standard. One team built a tool to help decide treatment hours, called the POP-C. In testing, it lined up with both norm-referenced and criterion-referenced measures of skill (Toby, L. M., Hustyi, K. M., Hartley, B. K., Dubuque, M. L., Outlaw, E. E., & Logue, J. J. (2023). Development and Preliminary Validation of the Patient Outcome Planning Calculator (POP-C): A Tool for Determining Treatment Dosage in Applied Behavior Analysis. Behavior Analysis in Practice, 17(2), 601-614. https://doi.org/10.1007/s40617-023-00861-6). Tools like this show why clear, tested standards matter.

FAQ#

What is a criterion-referenced assessment? It judges a skill against a set standard you pick ahead of time. If the learner meets that mark, they have the skill. It does not compare them to other people. The focus stays on the skill itself.

How is it different from a norm-referenced test? A norm-referenced test compares a person to a peer group. It asks how they rank against others the same age. A criterion-referenced test asks if they meet a fixed skill mark. One looks at the crowd, the other at the standard.

What does "teaching to the test" mean here? It means training only the exact items on the assessment. The test score rises but the real skill may not grow. To avoid it, check that your teaching helps in real life. Use the test as a guide, not the whole goal.