To display missing and invalid data on ABA graphs, assign each state before plotting: observed value, true zero, missing, invalid, late, corrected, not due, or zero opportunity. Use visible gaps or coded markers and annotate relevant reasons. Do not convert unknown values to zero or connect a line across a period when the display would imply observed continuity.

Create a state dictionary

Define observed, zero, missing, invalid, late, corrected, not due, and zero opportunity with examples.

Use gaps intentionally

Leave unknown or invalid values unplotted. Add a marker or annotation when the reason changes interpretation.

Preserve corrections

Show the validated current value and retain original value, author, date, reason, and correction history outside or beside the graph.

Report missingness separately

Calculate valid observations divided by observations due and list aged open records. A clean line should not hide data availability.

Make graph states accessible

Use shape, line pattern, labels, or a companion table so states do not depend on color alone. A gap means no valid numeric point is available; it does not mean zero. If a marker identifies a missing or invalid state, place it outside the outcome scale or in a clearly labeled status lane so readers do not mistake it for a measured value. Define every symbol in the legend and provide the state in accessible text.

Use the formula and its units for Gus

For Gus's missing-and-invalid graph specification: graph state is categorical before it is numeric; only valid observed values and true zeros enter the value series. Write the unit beside every result and keep the original numerator, denominator, time, opportunities, response definition, and condition. State rounding only after the unrounded calculation is preserved.

Work the arithmetic for Gus

Gus's series contains 12 scheduled sessions. At a cutoff after session 8, sessions 1 through 8 are due and sessions 9 through 12 are not due. Sessions 1 through 5 have valid values, session 6 is missing, session 7 is invalid because of a technology failure, and session 8 has a late corrected value that has now been validated.

Six of eight due sessions therefore have valid current values, so observed-data availability is 6/8 = 75%. Missingness is 1/8 = 12.5% and invalidity is 1/8 = 12.5%. The four future sessions remain outside those denominators. These percentages describe record availability, not Gus's self-advocacy performance.

Draw the series without inventing continuity

Plot sessions 1 through 5 normally. Leave the outcome series blank at sessions 6 and 7, display their distinct states in the status lane or annotations, and resume with the validated value at session 8. Do not draw a solid segment from session 5 to session 8 if readers could interpret it as measured continuity. Session 8 receives a correction marker linked to the original value, corrected value, author, timestamp, reason, and validation status.

The clinical note should say what is known: "Six of eight due observations are currently valid. Session 6 is missing, session 7 is invalid after a technology failure, and session 8 was corrected after late validation." The graph does not support filling the two gaps, estimating an unobserved trend, or treating the technology failure as a client outcome.

Apply the graph-state specification to a decision

Gus's qualified clinician identifies the decision before selecting the summary or display. Compare at least one alternative unit or classification, inspect client access and direct experience, and record what evidence would change the interpretation. A correct calculation can still answer the wrong question.

Audit data states for Gus

Gus's record distinguishes valid observed values, true zeros, missing observations, invalid events, no opportunities, not-due records, late entries, corrections, and incompatible conditions. Give unresolved items an owner, age, and next action. Preserve the original evidence and the reason for every exclusion or change.

Protect access and context for Gus

Gus's measurement keeps AAC, interpreters, mobility, food, water, bathroom use, prescribed care, health support, rest, relationships, and emergency help available. Record ordinary supports, partner response, setting, schedule, burden, and withdrawal. The data system should describe care without creating unsafe or inaccessible test conditions.

Ask five calculation questions for Gus

Use these questions in the missing-and-invalid graph specification:

  • What exact response, event, opportunity, time exposure, unit, and condition does the value represent?
  • Which raw numerator, denominator, clock, source, and calculation reproduce it?
  • Which zero, missing, invalid, not-due, corrected, censored, or incompatible states stay separate?
  • Which client experience, access, integrity, agreement, burden, and context qualify interpretation?
  • Which clinical decision, alternative display, next evidence, owner, and review date follow?

Keep raw counts and units beside every derived value.

Use current sources within their scope for Gus

For Gus's missing-and-invalid graph specification, the BCBA Test Content Outline includes measurement, validity, reliability, sampling, graphing, assessment, and data-based decisions as examination content. It is not a case protocol. The BACB Ethics Code addresses competence, client involvement, consent and assent when applicable, documentation, risk, and evaluation for covered behavior analysts. The CASP public summary supplies high-level autism-treatment scope without publishing licensed detailed procedures.

For Gus, ASHA says AAC users should always have access to their communication tools or devices. The WWC Version 5.0 handbook supplies research-review standards, not universal clinical formulas, thresholds, or decision rules. None of these sources turns a derived metric into proof of effectiveness, safety, medical necessity, compliance, or causation.

Close Gus's measurement review

Ask Gus and the responsible clinician to review the missing-and-invalid graph specification through accessible communication. Record the raw data, calculation, unit, display, assumptions, missingness, context, interpretation limits, selected decision, next evidence, and review date. Reopen the analysis when the response definition, exposure, opportunity, setting, measurement system, client priority, or access changes.

Related resources

Sources