Can a Family Ask How ABA Data Were Corrected?

An ABA data correction needs provenance

Record the original value, corrected value, author, correction time, service or observation time, reason, supporting source, approval route, and downstream items reviewed. If policy uses an addendum or amendment, connect it to the original without hiding the sequence.

Different roles own different decisions

The person who observed the event may clarify what occurred. An authorized record role may enter the correction. A supervisor may validate implementation or documentation. A qualified clinician determines the effect on clinical interpretation. Technology can flag a mismatch without choosing the clinical meaning.

Review every downstream use

A corrected value may affect a graph, phase summary, mastery result, progress report, supervision review, authorization request, or claim support. List each affected artifact and its disposition. Preserve an unaffected result as reviewed rather than leaving it invisible.

The RBT Ethics Code requires accurate implementation and documentation. The BACB Ethics Code addresses accurate reporting, documentation, correction, and continual evaluation for covered behavior analysts.

Measure correction completeness

Suppose an import error affects 7 of 30 records. The practice reviews all 30, corrects 7, confirms 23 unaffected, and regenerates four graphs. Report cohort review as 30 of 30 and affected-record correction as 7 of 7. Excluding the 23 unaffected records would hide whether the cohort search was complete.

Source scope matters

The BCBA Test Content Outline covers measurement, data display, validity, and interpretation as examination content. The CASP public summary supports individualized evaluation within its autism-treatment scope. Neither source supplies a universal record-correction procedure, so the practice must follow its actual legal, payer, professional, and system requirements.

Clarify what kind of change occurred

Several actions can look like a “correction” in a report:

• fixing a transcription or calculation error
• adding information that was unavailable at the original entry time
• clarifying an ambiguous note without changing the observed fact
• changing a graph label or phase boundary
• replacing a defective export while preserving source records
• revising a clinical interpretation after new evidence

These actions have different owners and histories. A new interpretation does not rewrite the original observation. A display repair does not establish that the stored source value was wrong. Ask the provider to name the type of change.

Follow one value through the system

Choose a specific example and trace it from observation to final report. Confirm the original entry, any paper or device evidence, the stored value, the calculation, the graph point, the progress summary, and the current corrected version. This often identifies whether the defect was an entry problem, mapping problem, calculation problem, or interpretation problem.

Suppose Lian's source sheet records 3 independent responses out of 5 opportunities, while the app displays 3 out of 4. The correction should explain whether one opportunity was dropped by the app, excluded under a documented rule, or entered incorrectly. Changing the displayed percentage from 75% to 60% without resolving that question leaves the provenance incomplete.

Ask for scope, not only the changed cell

A family can ask whether the practice searched for the same issue in other dates, goals, people, devices, or exports. A single corrected entry may be the first evidence of a broader fault. The practice should define the search cohort, account for every record in it, and document why the boundary is reasonable.

If the error affected an authorization request, school communication, discharge summary, or other recipient, ask whether the corrected version was sent and the prior one identified. “Correct in the app” is only one part of reconciliation.

Use a concise written request

A practical message can say:

Please explain the correction to the entry dated [date] for [measure]. I would like to understand the original value, corrected value, reason, source evidence, author and approval, how the history is preserved, and whether any graphs, reports, or decisions were reviewed again.

The request can also ask for an accessible discussion with the qualified clinician. A formal records-access process may apply if the family wants copies rather than an explanation.

Know what a complete response looks like

A strong response identifies the record and change, separates fact from interpretation, gives the relevant dates and roles, preserves the earlier version, accounts for affected outputs, and states the follow-up decision. It can acknowledge uncertainty where source evidence is incomplete.

A response is incomplete when it offers only a new graph, attributes the change to “the system” without a cohort review, or silently removes the original value. Families can ask for escalation through the provider's record, privacy, compliance, or clinical process when the responsible role cannot answer.

Separate disagreement from factual correction

A family may disagree with a clinician's interpretation even when the observed value is accurately recorded. That situation calls for discussion, additional context, or the applicable amendment process rather than silently changing the observation. The record can preserve both the original professional entry and the family's statement when governing rules allow.

Conversely, a documented transcription error should not remain uncorrected merely because it appeared in a signed report. The practice can follow its authorized late-entry, amendment, or correction process while preserving authorship and timing.

Check the revised decision, not only the revised number

Ask whether the corrected evidence changed mastery, progress, risk, generalization, maintenance, dosage, transition, or another plan decision. If the conclusion stayed the same, the clinician can explain why. If it changed, the record should show the new rationale and what the team told the person and family.

For example, correcting one opportunity from independent to prompted changes a session from 4 of 5 to 3 of 5. That may change the displayed percentage without changing a conclusion based on several weeks of evidence. Another correction might cross a mastery threshold and require a fuller review. The effect depends on the decision rule and complete record.

Close with a prevention question

Ask what will reduce recurrence. The answer might involve a clearer definition, form validation, staff training, device testing, required reason codes, export checks, or review before reports are released. A useful corrective action has an owner and a test.

Families do not need access to confidential personnel matters to understand the control. They can reasonably ask how the practice verified that the corrected record is accurate and how it will know whether the same problem happens again.

Questions families can use

Ask what triggered the review, how the affected cohort was defined, what changed, who made and approved the entry, which evidence supports it, how the original remains visible, which reports or decisions were reassessed, and how the team verified the correction.