Missing data single-case randomization test procedures must preserve each missing time position and reason and apply one prospectively justified rule across the observed and reassigned schedules. Do not delete, impute, or reclassify points according to which treatment label makes the result favorable. Report missingness by phase and assignment position, effective denominators, sensitivity analyses, and whether the chosen statistic and software remain valid for the design.

Distinguish missing from measured zero

Use explicit states for collected value, missing, invalid, interrupted, and ineligible. Preserve time, condition, reason, and source.

Predeclare the analysis rule

Specify whether the statistic uses available observations, a design-valid imputation, or another method and apply it identically across assignments.

Track denominators by assignment

Reassigned phase labels can change available baseline and intervention counts. Store those counts with every statistic.

Run meaningful sensitivity checks

Compare a small set of prospectively defensible missing-data rules, report all results, and avoid selecting by favorability.

Repair the source of missingness

Investigate access, device, staffing, scheduling, definition, or burden failures and determine whether missingness is related to condition or outcome.

Build Yara's missing-position analysis grid

For the missing data single-case randomization test question, start with a locked protocol and a versioned missing-position analysis grid. Record Yara's design, setting, randomization unit, complete schedules, constraints, assignment probabilities, draw, selected schedule, delivered schedule, statistic, tail, tie and missing-data rules, software, and reviewers. Link each schedule position to the original time-series record. The file should let an independent analyst reconstruct randomization with missing outcomes without guessing what was possible before outcomes.

Work the schedule example for Yara

Yara's plan has 16 measurement positions, with outcomes missing at positions 6 and 13 for documented access failures. The analysis grid keeps those positions missing under every permissible reassignment and calculates the predeclared statistic from available observations. Phase counts can differ across schedules, so each denominator is stored rather than assumed constant. Show the schedule generator, inclusion tests, probability calculation, draw evidence, and selected sequence at full precision. This fictional example illustrates two unavailable outcomes in a randomized time series; it does not establish a universal phase length, alpha, power level, or treatment plan.

Audit Yara's assignment evidence

Yara's grid lists all 16 positions, two missing flags, reasons, condition as assigned, condition delivered, inclusion rule, per-assignment phase counts, statistic, and sensitivity results. It verifies that no value was replaced with zero and no missing position disappeared from the audit trail. The audit also checks timestamps, protocol amendments, allocation concealment when applicable, cancellations, replacements, missing positions, phase labels, probability totals, software version, and whether the analysis generator matches the design generator. Unresolved discrepancies stay on hold.

Prevent the assignment error in Yara's review

Dropping missing observations only when they weaken the result is outcome-driven analysis. Coding an unavailable measure as zero can confuse no observed response with no observation. Yara's locked missingness rule prevents both errors. Randomization means chance operated at the named assignment step under the documented scheme. Varied, alternating, staggered, or response-guided schedules are not automatically randomized.

Integrate design and visual evidence for Yara

A methodologist assesses whether missingness, varying phase denominators, or informative observation failure undermines the planned test. The team also repairs the access or measurement system and reports what evidence remains useful for Yara's care and research question. The WWC Version 5.0 handbook is a research-review standard, not a universal clinical protocol. Review the graph, level, trend, variability, immediacy, overlap, consistency, measurement quality, assignment fidelity, original-unit magnitude, and every planned replication.

Protect Yara's participation

The randomization with missing outcomes never outranks Yara's welfare. Keep augmentative and alternative communication, interpreters, mobility, food, water, bathroom use, prescribed care, health support, rest, relationships, and emergency help available. Use accessible consent and assent processes and honor withdrawal or distress. If safety, medical need, access, or choice changes the schedule, qualified people act and document the design consequence.

Use design-specific sources for Yara

For Yara's two unavailable outcomes in a randomized time series, the BACB ethics hub and CASP public summary provide professional context, and the BCBA Test Content Outline identifies examination content on measurement and single-case design. A randomized SCED overview covers phase and alternation schemes. Research on multiple-baseline power, rapid alternation, and changing-criterion randomization shows design-specific behavior. A health-sciences methods paper discusses randomized phase, alternation, and case-placement options. Nonconcurrent multiple-baseline research addresses assignment and design-quality issues. ASHA supports continuous AAC access.

Rehearse Yara's schedule before collecting outcomes

Before the community navigation study begins, Yara's team runs the missing-position analysis grid with fictional schedules and outcomes. The rehearsal checks the earliest and latest permissible assignment, probability totals, selected-schedule trace, delivered-schedule fields, statistic, missing-data behavior, and a hand-worked reference result. Reviewers store test cases, expected outputs, code version, and correction log. The dry run can repair mechanics before outcomes; it cannot use Yara's future data to change the assignment scheme.

Define Yara's stop, pause, and amendment rules

Yara's protocol names who may stop or pause the randomization with missing outcomes, which health, safety, assent, access, staffing, technology, or feasibility events trigger action, and how urgent care proceeds. It separates an immediate protective action from a later research amendment. Any amendment receives a new version, date, rationale, approval path, and prospective assignment set; the original schedule and data remain intact. The team also tells Yara how to withdraw or raise a concern through an accessible route. These rules make the humane response predictable and keep a necessary change from being hidden as ordinary schedule variation.

Close Yara's design review

Review the missing-position analysis grid with Yara, the responsible clinician, and a statistician or methodologist familiar with the exact design. Preserve protocol, schedule set, probabilities, draw, selected and delivered sequence, raw data, graph, code, outputs, deviations, inference limits, client input, and decisions. Keep the page draft and noindex until every named review is complete.

Related resources

Sources