An ABA graph audit traces every point to its source record and verifies the definition, unit, denominator, formula, axes, spacing, phase labels, missingness, annotations, accessibility, and version. Recalculate a sample independently and preserve corrections. Clinical review should begin from a display whose source lineage and construction choices are visible rather than from a polished image with unknown assumptions.

Trace every point

Link date, session, source record, raw value, numerator, denominator, formula, definition version, and condition.

Verify display mechanics

Check axes, scale, spacing, line connections, phase lines, criteria, symbols, labels, legends, and rounding.

Inspect missingness and annotations

Confirm every blank, exclusion, event note, and correction has a source and state.

Test accessibility and export

Review contrast, direct labels, data-table access, screen-reader text, print view, and file version.

Use a point-level source audit

Select every point for a small graph or a risk-based sample for a longer series. Trace date, session ID, raw value, numerator, denominator, formula, definition version, condition, and validity state. Compare the plotted value with the locked record and record the reviewer and result. Expand the sample when any mismatch appears.

Audit the complete population for phase boundaries, goal lines, annotations, missing states, and manual overrides. These elements may occur rarely but can change the interpretation of the entire graph.

Reconcile Uma's four affected points

Uma finds two source-value mismatches, one absent opportunity denominator, and one point under the wrong phase date. If those categories describe four distinct points, 4 of 16 = 25% require correction. If one point has more than one issue, report four findings and the smaller count of affected points separately.

After repair, the second reviewer reproduces sixteen of sixteen points from locked records. That confirms construction reproducibility for the corrected version. The clinician still evaluates measurement validity, clinical importance, client experience, and the decision supported by the data.

Audit the display as a system

Check y-axis origin and range, x-axis spacing, point order, line connections, phase labels, criterion sources, aggregation, rounding, and filters. Confirm that a zero is measured, a blank is classified, and excluded observations remain visible in the source table. Compare the interactive view, downloaded image, printed page, and mobile layout.

Provide a readable data table, direct labels, sufficient contrast, noncolor cues, and concise alternative text that names the measure and major structure without claiming a conclusion the graph cannot support.

Preserve release evidence

Store the source extract or stable reference, graph code or configuration, audit checklist, corrections, reviewer, approval, and released file identifier. When the graph changes, issue a new version and retain the old one. Link any corrected clinical note or report to the updated graph through the applicable documentation process.

Review ABA graph audit before release

The graph-release audit for Uma identifies the clinical question, source record, unit, denominator, time basis, display rule, qualified decision owner, and next review before construction. Preserve the raw data and the exact graph version shown to reviewers.

Work the graph example for Uma

Uma audits sixteen plotted points. Two fail source-value matching, one lacks its opportunity denominator, and one sits under the wrong phase date. Four of 16 points, or 25%, need correction. After repair, all sixteen trace to locked records and a second reviewer reproduces the display.

Audit what the display leaves out for Uma

For Uma, record missing observations, invalid points, small or zero denominators, phase uncertainty, context changes, access conditions, and any series hidden by aggregation. A clean picture gains value only when those states remain available for review.

Protect access and client meaning for Uma

During Uma's reproducible graph review, keep AAC, interpreters, mobility, food, water, bathroom use, prescribed care, health supports, rest, relationships, and emergency help available. Ask the person through accessible communication whether the outcome and display reflect something useful. Immediate safety action never waits for graph completion.

Use sources within their role for Uma

For Uma, the BACB ethics hub and CASP public summary provide professional context. The BCBA Test Content Outline includes graphing, visual analysis, and data-based decisions as examination content. The WWC handbook page and Version 5.0 handbook address research review, so routine care should avoid borrowing a causal claim without suitable design.

For the reproducible graph review question, Lobo and colleagues discuss single-case design and analysis, while Wolfe, Barton, and Meadan describe systematic visual-analysis protocols. McTiernan and colleagues show that linear and ratio scales can produce different visual impressions. Zonneveld and colleagues study graph-construction training. ASHA supports continuous access to AAC tools.

Close the graph review for Uma

Review the graph-release audit with Uma and the responsible clinician. Record the display choice, source lineage, graph version, interpretation limits, decision, owner, next data requirement, and review date. Reopen the audit when the definition, measure, axis, condition, criterion, context, access, or software changes.

Before release, confirm every correction is complete, the independent reproduction matches, and the accessible exports display the same data and phase structure. Clinical review should cite the released version rather than an editable working graph.

Related resources

Sources