Stop Collecting Data Nobody Uses: A Minimum Viable System for School BCBAs

A school BCBA can build a beautiful data system that is completely useless by Tuesday.

It usually starts with good intentions. We want a defensible baseline. We want enough detail to see patterns. We want frequency, duration, antecedents, consequences, replacement behavior, fidelity, setting events, and maybe a notes box just in case.

Then we hand the form to a teacher who is simultaneously teaching 26 students, responding to an IEP accommodation, finding a missing Chromebook, and trying to keep a transition moving.

The form comes back half-complete—or not at all—and we call it a data problem.

Often it is a design problem.

A current public discussion among BCBAs asked how much baseline data is enough before teaching begins. The answers varied, but the tension underneath the question matters in schools: more data is not automatically better data, especially when collecting it delays support or depends on a system adults cannot sustain. Teachers have been making the same point publicly for years: progress-monitoring expectations can feel impossible during real instruction.

The answer is not to stop measuring. It is to build the smallest system that can answer the next important question.

I call that minimum viable behavior data.

Start With the Decision, Not the Data Sheet

Before choosing frequency, duration, latency, interval recording, or a rating scale, finish this sentence:

We will use these data to decide whether to ________.

Examples:

• continue the current transition support for another week;
• change the task length;
• teach a faster help response;
• move from brief consultation to an individualized assessment;
• check whether behavior changed after the schedule shift;
• determine whether the plan is not working or is not being implemented.

If the team cannot name the decision, do not ask staff to collect the data yet.

This one step prevents “collect everything” systems. It also helps the BCBA explain why a measure matters. A teacher is more likely to record a 10-second rating when they know the team will review it Friday and use it to choose Monday's support.

The Five Parts of a Minimum Viable Data System

1. One observable question

Do not begin with “How was behavior today?” Begin with a question tied to an important routine.

• Is the student beginning independent writing within two minutes?
• Is the student using the taught break response before leaving the area?
• Is disruption decreasing during the first 15 minutes of math?
• Are adults providing the pre-correction before the transition?

The question should describe something the team can see and something the intervention is designed to change.

2. One measure matched to the question

Choose the lightest measure that will answer it.

A direct behavior rating can be a practical option when precise continuous measurement is not necessary. The adult observes a defined period, then rates a clearly defined behavior immediately afterward. Research describes these ratings as efficient and flexible for school progress monitoring—but they are estimates, not magic. The behavior, observation window, anchors, and rater expectations still need to be clear.

Use more intensive measurement when the risk or decision requires it. Severe behavior, safety planning, an FBA, a legal process, or a treatment evaluation may require direct observation and more rigorous measurement. “Minimum viable” means sufficient for the decision, not merely convenient.

3. One routine and one recording moment

“Collect data all day” is not a plan.

Attach collection to a routine that already has a natural ending:

• immediately after arrival;
• when the writing block ends;
• after lunch transition;
• at the end of the class period;
• during the BCBA's scheduled observation.

Specify who records it and where. If three adults assume somebody else has the clipboard, you do not have a data system.

Try this level of clarity:

For the 15-minute independent writing block, the teacher circles 0–10 for academic engagement immediately when the timer ends. The form stays on the writing clipboard. The BCBA collects it Friday.

That is runnable. “Track on-task behavior daily” is not.

4. A review date and a decision rule

Data without a review date become paperwork.

Set the review when the system begins. For routine classroom progress monitoring, that might be after five instructional days or after a specified number of opportunities. Then agree on a simple rule.

For example:

On Friday, we will review the five writing-block ratings. If engagement is improving and the adult support was used on at least four days, continue for one more week. If engagement is flat despite implementation, revise the support. If implementation occurred on fewer than four days, repair feasibility before changing the student plan.

The exact threshold should fit the case; it is not a universal clinical standard. The important part is separating three conclusions:

1. the support appears to be helping;
2. the support was implemented and needs revision;
3. there is not enough implementation to judge the support.

Those are different problems. A graph alone cannot solve that distinction.

5. One small check on adult implementation

If student behavior is measured but the key adult action is not, the team may intensify a plan that never occurred.

Do not respond by adding a 20-item fidelity checklist. Identify the one or two adult behaviors most likely to make the plan work.

• Was the visual ready before the direction?
• Was the break response available?
• Did an adult provide the pre-correction?
• Was reinforcement delivered after the replacement response?

This can be a yes/no item completed by the implementer, a brief BCBA observation, or a permanent product. The purpose is improvement, not surveillance.

The 10-Minute Data Design Meeting

Build the system with the person expected to use it.

Ask these five questions:

1. What decision do we need to make?
2. What is the easiest credible sign of progress?
3. Which routine gives us the best chance to observe it?
4. When can you realistically record it?
5. When will we look at it together?

Then run a usability test. Hand the form to the teacher or para and ask them to explain what they would do during the actual routine. If the explanation requires repeated clarification, simplify the form. If recording competes with student safety or instruction, change the collection method or assign the observation elsewhere.

This is where collaboration becomes technical. Research with special education teachers has found that relationships with BCBAs break down when recommendations feel impractical, condescending, or blame-oriented. More recent school-BCBA scholarship makes ecological validity and shared team knowledge central to effective support.

“I cannot collect that while teaching” is not refusal. It is information about the environment your system must fit.

What to Remove This Week

Audit one current data form and remove anything that does not answer the next decision.

Look for:

• duplicate fields already captured elsewhere;
• ABC narratives when the team only needs progress monitoring;
• all-day collection when the problem occurs in one routine;
• multiple topographies that lead to the same decision;
• rating scales without behavioral anchors;
• forms nobody reviews on a predictable schedule;
• precision that the collection conditions cannot support.

Do not lower the standard of evidence. Raise the standard of usability.

A Simple Example

A student regularly leaves their seat and talks with peers during independent math. The original form asks the teacher to tally every instance across the day, record antecedents and consequences, and note task difficulty.

The teacher completes it twice in two weeks.

The team rebuilds the system:

• Decision: whether a choice of task order plus a help card improves independent math participation.
• Student measure: engagement rating from 0–10 at the end of the 15-minute block.
• Routine: independent math only.
• Adult measure: yes/no—choice offered and help card visible before work begins.
• Review: Friday after five opportunities.

Now the team can tell whether the support occurred and whether participation changed. If closer analysis becomes necessary, the BCBA can conduct targeted direct observations. The lightweight system does not replace assessment; it tells the team what to examine next.

The Standard I Use

A school behavior data system should pass four tests:

1. Useful: It answers a decision the team will actually make.
2. Doable: It survives the real routine without displacing instruction or safety.
3. Interpretable: Different people understand what the score means.
4. Reviewed: Somebody looks at it on a known date and closes the loop.

If a measure fails one of those tests, collecting more of it will not rescue the system.

The goal is not the most impressive clipboard. The goal is enough trustworthy information to make the next good decision—and a system the people closest to the student can actually sustain.

This week, choose one form. Write the decision at the top. Remove one measure nobody uses. Define one recording moment. Put the review date on the calendar.

That is not less analytical.

It is behavior analysis designed for school.

References

• Miller, F. G., Crovello, N. J., & Chafouleas, S. M. (2017). Progress monitoring the effects of daily report cards across elementary and secondary settings using Direct Behavior Rating: Single Item Scales. Assessment for Effective Intervention, 43(1), 34–47. https://doi.org/10.1177/1534508417691019
• Reilly, A. M., Crowell, G. E., Thoele, J. M., DeAngelo, S., & Bak, M. Y. S. (2025). School-based transdisciplinary teaming to maximize behavioral supports. Behavior Analysis in Practice. https://doi.org/10.1007/s40617-025-01054-z
• Squires, M., Cutrer-Párraga, E. A., Morris, J. R., Miller, E. E., & Hansen, B. D. (2024). Navigating collaboration: Factors influencing special education teachers' relationships with BCBAs in diverse school contexts. Behavior Analysis in Practice. https://doi.org/10.1007/s40617-024-01009-w
• Truman, K. J., Van Norman, E. R., & Kilgus, S. P. (2026). An estimation of trend variability and change sensitivity of Direct Behavior Rating–Single Item Scales. Journal of Positive Behavior Interventions. https://doi.org/10.1177/10983007251386077