To randomize multiple-baseline intervention starts, define a valid start window for every tier and enumerate combinations that preserve the planned stagger and phase lengths. Select one complete combination prospectively by chance. Store tier identities, candidate starts, combination probabilities, selected and delivered starts, and the matching analysis. Start separation, observation count, cases, autocorrelation, gradual effects, and overlap among windows can affect power.
Define tier-specific windows
Record earliest and latest start, minimum observations, clinical readiness, and reasons for different windows across people, behaviors, or settings.
Generate valid joint combinations
Cross the tier choices, then remove only combinations that violate prospectively stated stagger, overlap, or feasibility constraints.
Assign probabilities to vectors
State whether each valid combination is equally likely. A sequence of draws can create unequal joint probabilities if the procedure is not designed carefully.
Preserve tier and vector evidence
Store each tier start, the complete selected vector, draw evidence, and delivered vector. Do not keep only an unlabeled list of dates.
Review design power factors
Consider cases, observations, start-window overlap, stagger separation, autocorrelation, expected effect, and gradual emergence before finalizing the design.
Build Omar's tier-start combination table
For the randomize multiple-baseline intervention starts question, start with a locked protocol and a versioned tier-start combination table. Record Omar'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 randomized multiple-baseline starts without guessing what was possible before outcomes.
Work the schedule example for Omar
Omar's three tiers allow starts {5,6}, {8,9}, and {11,12}. Every choice in an earlier tier precedes every choice in the next, so 2 x 2 x 2 creates eight valid combinations. An equal-probability draw selects starts 6, 8, and 12 before the outcomes are collected. Show the schedule generator, inclusion tests, probability calculation, draw evidence, and selected sequence at full precision. This fictional example illustrates staggered starts across three tiers; it does not establish a universal phase length, alpha, power level, or treatment plan.
Audit Omar's assignment evidence
Omar's table lists all eight combinations, probability 1/8 each, tier definitions, start-window rationale, random draw, selected vector [6,8,12], delivered vector, observation coverage, and test input. It retains combinations that later produce inconvenient results. 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 Omar's review
Randomizing each tier independently without checking the joint combination can violate the planned order or stagger. Omar's generator validates the complete vector before it enters the draw pool. 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 Omar
Visual analysis still examines prediction, verification, replication, level, trend, variability, immediacy, overlap, and consistency for each tier. The randomization result complements those case-level patterns and does not average away a tier that fails to change. 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 Omar's participation
The randomized multiple-baseline starts never outranks Omar'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 Omar
For Omar's staggered starts across three tiers, 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 Omar's schedule before collecting outcomes
Before the three-classroom communication study begins, Omar's team runs the tier-start combination table 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 Omar's future data to change the assignment scheme.
Define Omar's stop, pause, and amendment rules
Omar's protocol names who may stop or pause the randomized multiple-baseline starts, 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 Omar 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 Omar's design review
Review the tier-start combination table with Omar, 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
- How to Randomize Condition Order in an Alternating-Treatments Design
- How to Randomize a Phase Start in an AB Single-Case Design
- How to Build a Restricted Randomization Set for Alternating Treatments
- How to Handle Missing Outcomes in a Single-Case Randomization Test
Sources
- Behavior Analyst Certification Board, Ethics Information and Ethics Codes
- Council of Autism Service Providers, ABA Practice Guidelines Version 3.0 public summary
- Behavior Analyst Certification Board, BCBA Test Content Outline, 6th edition
- What Works Clearinghouse Procedures and Standards Handbook, Version 5.0
- Randomized Single-Case Experimental Designs in Healthcare Research: What, Why, and How?
- Power of a Randomization Test in a Single Case Multiple Baseline AB Design
- Randomization Tests for Single Case Designs with Rapidly Alternating Conditions
- Type I Error Rates and Power of Two Randomization Test Procedures for the Changing Criterion Design
- Randomized Single-Case Intervention Designs and Analyses for Health Sciences Researchers
- Examining and Enhancing the Methodological Quality of Nonconcurrent Multiple-Baseline Designs
- American Speech-Language-Hearing Association, Augmentative and Alternative Communication