An ABA high-probability request sequence worksheet should show more than whether a target request was followed. It should preserve what the team asked, why that request belonged in the plan, how candidate high-probability requests were identified, what the person communicated, whether the target opportunity was actually available, and whether the sequence created strain or pressure.

The copyable ABA high-probability request sequence worksheet below keeps sequence delivery, component responses, target responses, implementation integrity, and unwanted effects in separate records. It does not provide a standard number of requests, a timing formula, or a compliance target. The fictional museum example demonstrates denominator control without setting a benchmark for a real client.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

Clinical boundary: A high-probability sequence does not make an inappropriate target appropriate. Use it only within an authorized, individualized process led by qualified professionals. Ordinary AAC, speech, gesture, sensory, motor, medical, safety, relationship, and access supports remain available. Honor and review a pause, refusal, help request, or change in participation. None of those responses permits escalating prompts. The BACB ethics resources provide current professional context, but this worksheet cannot establish consent, assent, necessity, effectiveness, coverage, or legal compliance.

Define the arrangement before interpreting momentum

The research literature commonly describes a high-probability sequence as several requests associated with a high probability of responding followed closely by a request associated with a lower probability of responding. In the foundational behavioral-momentum study, the arrangement was evaluated across five experiments and was associated with changes in response probability, latency, and task duration. Those results do not mean the physics metaphor is an observed mechanism in every applied case. A clinical record should name the events that were arranged and the responses that occurred.

This distinction matters because people often use “behavioral momentum” to describe any easy-before-hard sequence. A high-probability request sequence requires current case evidence that the first requests actually have a higher response probability under relevant conditions. It also requires a clear low-probability target and a defined relationship between the requests. Alternating familiar and acquisition teaching trials is task interspersal. That arrangement answers a different instructional-design question.

The Sixth Edition BCBA Test Content Outline separates measurement, assessment, intervention, integrity, generalization, maintenance, and data-based decisions. The BACB's test-content outline page identifies the current certification outlines. These materials help organize the review; they do not select a sequence for an individual.

Purpose and target-opportunity card

The target request should have a meaningful, authorized purpose beyond producing agreement with an adult. State what access, participation, skill, or routine is at issue. Define when the opportunity begins, what makes it valid, the response forms that count, the observation window, and what the person can do instead. If a target is unsafe, inaccessible, medically contraindicated, outside the plan, or unnecessary, omit it. A request sequence does not cure that defect.

FieldCase-specific entryMeaningful purpose and authorized plan referenceTarget request and plain-language explanationConditions required for a valid target opportunityInitial response definition and case-defined latencyAccepted response variations and access supportsHelp, break, refusal, alternative, and change responsesConditions that make the target inappropriate or invalidDirect measure and observation methodQualified owner, version, effective date, and review date

Do not score an unavailable activity, missing material, absent communication system, or changed instruction as client nonresponse. Keep those events as system or validity data. A request presented after the plan's observation window can also be informative without being recoded as an on-time initial response.

Candidate high-probability request evidence

“Easy,” “mastered,” “preferred,” and “high-probability” are not interchangeable. A mastered response might be effortful today. A preferred activity is not automatically an instruction the person wants repeated. Team recollection may generate candidates, but recent direct observations under defined conditions should determine whether the label fits the current version. Record the valid presentations and initial responses behind the probability instead of copying a percentage forward without its counts.

Candidate requestExact wording or signalContext sampledValid presentationsInitial responsesResponse probabilityEffort/access notesClient feedbackInclude, hold, or excludeEvidence date

The study of medical examination requests for two children with autism used repeated observations to identify candidate high-probability requests and separately measured responses to the target requests and procedural integrity. Its participant characteristics, thresholds, number of requests, timing, target procedures, and consequences belong to that study. They are not defaults supplied by this worksheet.

A study comparing variant and invariant high-probability requests reported different patterns across the studied arrangements. That result is a reason to record the actual request pool and order, not a rule that novelty is always better. If responding to the supposed high-probability requests declines, the label should be reviewed rather than preserved by assumption.

Versioned sequence map

Write the sequence so another authorized team member can distinguish correct delivery from a similar-looking string of demands. Record how the person learns what is coming and what follows each component response. Separately state what happens after no response or dissent and when the target request is omitted.

Sequence elementCurrent versionOperational descriptionAccess or supportConsequence or next eventException or stop conditionEvidence sourceOwnerAdvance information and choiceHigh-probability request 1High-probability request 2, if plannedHigh-probability request 3, if plannedInterrequest timing definitionTarget request and observation windowReinforcement or ordinary outcomeHelp, refusal, break, or alternative pathHold, revise, fade, or stop question

The analysis of stimuli associated with a target instruction documented a sharp decrease in responding to high-probability instructions for one participant when features associated with the later low-probability instruction were present. The investigators stopped the sequence and examined the relevant stimuli. This is valuable counterevidence to the idea that a high-probability label remains stable regardless of context.

Event ledger: record what happened before interpreting it

One row represents one planned target opportunity, not one component request. Use a linked sequence ID when the component-level record is needed. Preserve the first response, later supported completion, and communication separately.

Date/timePlan versionTarget opportunity and contextValid? reasonSequence due?Sequence delivered as written?Initial target response/latencyLater supported outcomeHelp/refusal/alternativePressure or unwanted effectObserveryes / noyes / no / not applicableyes / noyes / no / not applicableyes / noyes / no / not applicable

Linked component-response record

Sequence IDComponent requestPresented as written?Initial response/latencyConsequence delivered as written?Access issueDissent or strainValid for component analysis?

Do not overwrite the initial target response when the person later completes the routine with support. Both events can matter, but they answer different questions. Likewise, success on component requests does not turn a missing, altered, or inaccessible target opportunity into a valid target trial.

Pressure and acceptability are direct review domains

A sequence can feel predictable and supportive to one person and manipulative or exhausting to another. Record whether the target was explained, the person could decline component requests without penalty, the low-probability request was not hidden, ordinary breaks remained available, and staff did not use rapid delivery to crowd out communication. Pressure can appear as changed affect, withdrawal, repeated checking, delayed communication, efforts to escape, or direct feedback. Invite the person's description in an accessible form; relevant caregivers may also notice effects that an event count misses.

Review dateCommunication formPerson's description or observed responseTarget disclosed and understood?Choice/refusal honored?Ordinary access preserved?Burden, stigma, or relationship effectRequested changeFollow-up owner

The parametric single-case and social-validation study compared one-, three-, and six-request versions with five children and included participant and caregiver choice. The preference pattern is evidence from those participants and options; it does not establish a universally preferred sequence length. It does support making choice and acceptability visible rather than treating target-response data as the whole outcome.

The ASHA AAC Practice Portal describes AAC as individualized and multimodal and states that people should retain access to their communication tools. A request sequence must leave time and means to communicate yes, no, wait, help, discomfort, a break, or an alternative. Device use is not a component request to be earned.

Separate denominators answer separate questions

Review questionNumeratorDenominatorResultDoes not establishValid target-opportunity coveragevalid target opportunitiesplanned target opportunitiesclient performance or effectivenessSequence-delivery fidelityvalid opportunities with the sequence delivered as writtenvalid opportunities when sequence was duecomponent or target responseComponent responseinitial case-defined component responsesvalid high-probability component opportunitiestarget response or preferenceTarget responseinitial case-defined target responsesvalid target opportunitiescausation or assentPlanned-opportunity yieldinitial case-defined target responsesall planned target opportunitieswhy opportunities were invalidOverall implementation integritycorrectly implemented componentsobservable implementation-component opportunitiesbenefit, acceptability, or mechanism

Retain the counts beside each percentage. A target-response percentage cannot reveal whether the sequence was delivered correctly. A high component-response percentage cannot show whether the later request was reasonable. Neither percentage measures assent, safety, dignity, or relationship quality, so those review fields should never be auto-filled from response data.

Fictional community museum example

In this fictional record, a client and team were reviewing a supply-room routine tied to a chosen museum volunteer role. Fourteen target opportunities were planned. One was cancelled when the scheduled activity changed, and another lacked the required labeled bins. Those two opportunities were invalid for target-response analysis, leaving 12 valid opportunities.

  • valid target-opportunity coverage: 12 / 14 × 100 = 85.7%;
  • sequence delivery as written: 10 / 12 × 100 = 83.3%;
  • high-probability component response: 27 / 30 × 100 = 90.0%;
  • target response among valid opportunities: 9 / 12 × 100 = 75.0%; and
  • planned-opportunity yield: 9 / 14 × 100 = 64.3%.

The 30 component opportunities came from three defined component requests in each of the 10 sequences delivered as written. Twenty-seven met the fictional initial-response definition. The target response occurred within the case-defined latency in nine of the 12 valid target opportunities. The record does not use the two missing sequences as a causal comparison: two observations are too little context, assignment was not experimental, and other conditions may differ.

Implementation was examined separately in 16 observations across eight defined components, creating 128 observable opportunities. One hundred twenty-one were correct and seven were mismatches, which reconciles to 128. Fidelity was 121 / 128 × 100 = 94.5%.

Evidence layerFictional resultBounded readingPlanned coverage12 valid of 14 plannedTwo system conditions prevented a testSequence delivery10 of 12 due opportunitiesTwo valid targets lacked the planned sequenceComponent response27 of 30 componentsDoes not prove the requests stayed preferredTarget response9 of 12 valid targetsDescribes this definition and period onlyPlanned yield9 of 14 planned targetsKeeps system loss visibleImplementation integrity121 of 128 componentsSeven mismatches require reviewPressure and unwanted effectsdescriptive record requiredNo percentage above answers acceptabilityGeneralization or maintenancenot establishedNo cross-setting or durable claim follows

Begin the review with the two invalid opportunities, two omitted or altered sequences, three component nonresponses, three target responses outside the definition, seven fidelity mismatches, and the person's feedback. The arithmetic does not choose continue, revise, fade, or stop.

Implementation-integrity record

ObservationCurrent version availableValid target recognizedAdvance information/choice providedComponent requests accurateTiming accurateConsequences accurateAAC/access preservedDissent/help honoredCorrect componentsOpportunitiesNotes

Define each component before observing. A mismatch may reflect unclear wording, an inaccessible response, scheduling pressure, changed materials, training gaps, or version drift. It is not automatically a client variable.

Review, comparison, and stopping record

FieldReview entryVersions, settings, people, and date range reviewedPurpose and current justification for the target requestCurrent direct evidence for candidate high-probability requestsValidity loss and environmental barriersSequence, component-response, target-response, and fidelity resultsClient feedback, assent/dissent, pressure, burden, and unwanted effectsHelpful supports and ordinary access that must remainComparison condition or design limitsGeneralization, maintenance, or fading evidence still missingHold / revise / retain / compare / fade / stop questionQualified decision owner, rationale, and next review date

The Ethics Code for Behavior Analysts places client rights, involvement, informed consent, assent when applicable, confidentiality, competence, documentation, and data use within the responsible professional's duties. A completed worksheet is not evidence that those duties have been satisfied. The CASP ABA Practice Guidelines access page provides current context for individualized treatment planning, implementation, and evaluation; it does not endorse this form or a particular request sequence.

Use comparison language only when the design supports it. A sequence-associated improvement is not necessarily caused by the proposed mechanism, and a result from one target does not establish transfer. Reassess when component responding declines, the target changes, the person reports pressure, communication access is compromised, staff cannot implement the plan, the opportunity is no longer meaningful, or benefit is not demonstrated. Ending an unsuitable sequence can be the appropriate clinical decision.

Related resources

Sources

These sources vary in participants, settings, targets, request pools, consequences, timing, measures, and designs. They support explicit definitions, direct observation, version fidelity, client communication, acceptability review, and bounded interpretation. They do not supply a universal purpose, target, request count, timing, response threshold, consequence, fading plan, or forecast for a particular person.