Can an ABA Provider Change a Data Definition Mid-Goal?

A definition can need repair

A team may discover that different staff count different response forms, opportunities are not observable, the time window is unclear, ordinary supports were excluded, or the measure no longer answers the client's question. Leaving a defective definition unchanged can produce more data without producing better evidence.

The CASP public summary supports individualized assessment and treatment planning for people diagnosed with autism. It does not prescribe a single measurement definition.

Control every ABA data definition change

Record the original definition, problem, examples, client input, proposed version, author, approval, effective date, affected programs, staff training, and validation. Preserve the original entries. A version change should appear in the graph, table, or interpretation whenever it affects comparability.

Do not silently relabel old data under the new rule unless the source evidence genuinely supports rescoring and the correction remains traceable.

Qualified interpretation matters

The BACB Ethics Code addresses competence, assessment, client involvement, intervention, data evaluation, documentation, and supervision for covered behavior analysts. The BCBA Test Content Outline covers measurement systems, operational definitions, graphing, interpretation, and measurement validity as examination content.

An administrator can update a system field after approval. That role should not decide the clinical definition.

Compare phases cautiously

A broader response definition can increase counts without any change in the person's behavior. A shorter observation window can lower opportunity totals. A new observer, setting, device, or prompt rule can change results at the same time. The clinician should state which explanations remain possible.

Families can ask for separate summaries before and after the version boundary when pooling would mislead.

A practical example

The original help-request definition counts speech only. The revised version counts speech, AAC, and a consistent signed request. Before the change, 3 of 10 opportunities meet the old definition. After the change, 8 of 12 meet the accessible definition. The team does not call this a direct improvement comparison because the measured response class changed.

Treat the definition like a controlled version

Give each version an identifier and effective date. Store its observable response, opportunity, window, unit, supports, examples, nonexamples, invalid states, and author. Link every data point to the version used.

The change record should explain what problem the revision solves. “Make the graph look better” is not a valid rationale. Reasons may include improved accessibility, clearer observer agreement, alignment with the person's chosen response, a corrected time window, or a different clinical question.

Validate before using the new definition broadly

Test whether two trained observers can apply the revised rule to the same examples. Confirm that the data form and export preserve the new categories. Check that all affected staff can find the current version and that old templates no longer create unlabeled data.

The validation can reveal unintended consequences. A broader response class might include ambiguous movements. A shorter window might exclude a person's typical communication pace. The team can refine the definition before it becomes the basis for a high-stakes decision.

Do not erase the earlier phase

Old observations remain evidence of what was measured under the old rule. Display a phase line, separate panel, or version label. If the same source record can genuinely be rescored under both definitions, document who rescored it, when, and how agreement was checked.

Avoid rewriting the entire history under the new definition based on memory. The absence of source detail should remain visible as a limitation.

Separate access correction from measured improvement

Adding AAC, gesture, or another effective response form can repair an exclusion in the original definition. That change may immediately raise the counted total without any sudden change in the person's behavior. The better definition is still valuable because it recognizes communication that was previously ignored.

Explain this clearly to the person and family. The revised baseline may become the appropriate starting point for future decisions. The record can acknowledge that earlier results understated accessible responding.

A version change can affect more than the graph

Review treatment plans, progress reports, mastery criteria, supervision tools, authorization documents, family materials, app forms, and staff training. A payer or school report may need the version and date to understand why values shifted.

If a prior decision relied on a defective definition, the qualified clinician should decide whether to reassess it. Administrative staff can coordinate updated documents without rewriting clinical rationale.

A second example involving the opportunity rule

Kai's goal initially counts a choice opportunity whenever two items are visible. The team learns that several trials occurred before Kai's AAC page loaded. Version two requires the device to be available, the partner to pause for eight seconds, and both options to be accessible.

The number of eligible opportunities decreases, while the validity of each opportunity improves. Percentages before and after should remain in separate phases. The family can ask whether the new condition better matches everyday choice and how many observations are needed for the next decision.

Record the family-facing explanation

A concise note should state what changed, why, when, who approved it, how the person participated, whether old data remain comparable, and what happens next. This allows future staff and reviewers to understand the graph without guessing at the phase boundary.

Know when the old definition should be retired

After the effective date, staff should use the new version for the named program unless the plan explicitly runs a comparison. Remove obsolete forms from active workflows, preserve them in version history, and confirm that dashboards and exports label the new data correctly.

Mixed-version collection creates a false continuous series. If it occurs, identify the affected dates and collectors, correct only what source evidence supports, and mark unresolved records. The clinician can decide whether enough clean post-change evidence exists for the next decision.

Review the change with the person and family

Show examples of what now counts, which supports remain available, and how the graph will appear. Ask whether the revised response and opportunity reflect daily life. Invite correction when the definition misses a familiar communication form or creates an unrealistic condition.

This conversation is especially important when the revision changes a mastery threshold or progress report. The family should understand that a numerical jump at the boundary can reflect measurement change, actual performance change, or both.

Questions families can use

Ask why the definition changed, who authored it, how the client participated, which examples count, whether old data can be rescored, when staff were trained, how the boundary appears, and which decisions should use only the new version.