An ABA prompt level data sheet should preserve the initial response before help, the prompt that was planned, the prompt actually delivered, the level at which the response occurred, and any correction or retry. Blocked and invalid opportunities need their own visible record as well. With those distinctions intact, reviewers can compare initial independence with correct responding after support.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

This worksheet records an authorized procedure; it does not select the prompt type, arrange a hierarchy, prescribe a delay, or authorize fading. Those clinical choices draw on assessment, evidence, the current plan, client preferences and communication, risk, contextual fit, implementation capacity, and applicable requirements.

Separate the questions hidden in one prompt code

The current BCBA Test Content Outline, Sixth Edition separately identifies stimulus and response prompting, prompt fading, chaining, observable goals, contextual fit, procedural integrity, and data-based decisions. The current RBT Test Content Outline, Third Edition addresses implementation, measurement, documentation, and professional conduct within technician scope. These outlines describe certification content; they do not create a universal prompt hierarchy or validate this worksheet.

The phrase “prompt level” can hide three different facts: the planned step in a procedure, the help actually delivered, and the least support that appeared sufficient on that opportunity. A response after a gesture does not prove that the gesture was necessary. Likewise, a model delivered too early may conceal whether an independent response would have occurred.

Define the procedure version before collection

Open a new phase when the target, initial-response window, prompt types, order, timing, correction treatment, or opportunity rule changes materially.

Prompting contractEntryClient or authorized case codeClient-selected or informed goalTarget response and procedure versionEligible opportunity and trial boundaryInitial instruction, cue, or naturally occurring conditionInitial-response definition and wait windowIndependent, incorrect, no-response, blocked, and invalid codesPrompt types and locally defined codesPrompt order, decision rule, and timingPlanned prompt level for this phaseMeaning of delivered level and apparently required levelError-correction and retry treatmentReinforcement and feedback arrangementProbe, teaching, maintenance, and generalization labelsOrdinary communication, AAC, mobility, sensory, and other access supportsClient choice, assent, dissent, break, and stop processSafety and escalation conditionsIntegrity, observer, agreement, and review planAuthorized record location, retention, and correction ruleResponsible clinician and next review trigger

Pair local prompt definitions with examples and nonexamples. Labels such as “minimal,” “moderate,” and “maximum” remain difficult to compare unless the record states the observable assistance.

Local prompt-code key

The authorized procedure needs its own key because prompt types do not necessarily fall in the same order for every response. The rows below are editable placeholders.

CodeObservable prompt or conditionDelivery rule and timingResponse counted at this level whenAccess or safety boundaryINDNo programmed prompt before the initial responseOrdinary supports remain availableVDefined verbal promptGDefined gestural promptMDefined model promptPPDefined partial physical promptConsent, assent, training, safety, and role requirements applyFPDefined full physical promptConsent, assent, training, safety, and role requirements applySTDefined stimulus promptINVOpportunity invalid or blockedExcluded from the valid initial-response denominatorNever recode as incorrect

Not every program uses every row, and the displayed order is not a recommendation. Physical contact is not a default. Prompt selection and implementation remain governed by the current plan, competence, authorization, client rights, safety, and applicable law and policy.

Blank trial-level prompt data sheet

Enter the initial response before adding later support. Two separate fields then preserve what was delivered and the level at which the response occurred.

Date/sessionOpportunityTarget/exemplarInitial response before promptPlanned promptPrompt actually delivered and timeResponse after promptApparently required levelCorrection/retryValid?Client communication/contextObserver1Yes / no2Yes / no3Yes / no4Yes / no5Yes / no6Yes / no7Yes / no8Yes / no9Yes / no10Yes / no

If a prompt is delivered before the defined initial window ends, record that implementation deviation. Do not label the response independent merely because the prompt was subtle or routinely available.

Coverage and response summary

These quantities use different denominators and should not be collapsed.

QuantityCalculationWhat it can describeValid opportunity coverageValid initial-response opportunities / planned opportunities × 100How much of the planned frame supports initial-response analysisInitial independenceIndependent initial responses / valid initial-response opportunities × 100Responding before programmed prompts in the valid subsetPrompt exposureValid opportunities receiving a programmed prompt / valid initial-response opportunities × 100How often programmed prompts were deliveredCorrect after planned promptCorrect post-prompt responses / valid opportunities that received the planned prompt × 100Post-prompt responding among prompted opportunitiesFinal correct respondingIndependent initial plus correct post-prompt responses / valid opportunities × 100Correct responding by the end of the recorded procedure, not initial independence

Also report the distribution of prompt types, timing, correction, and deviations. A single percentage cannot show whether the prompt plan was implemented as written.

Prompt distribution and implementation review

Period or conditionPlanned opportunitiesValid initial observationsInitially independentProgrammed prompts deliveredCorrect after promptIncorrect/no response after procedureEarly/late/wrong prompt deviationsClient/access notes

Prompt codeDelivered countResponse occurred at this levelCorrections or retriesInterpretation note

Prompt labels are categories unless the local procedure establishes an order for a specific response. Turning them into numbers and calculating an “average prompt level” falsely implies equal intervals. A shift in the distribution can trigger review, but it does not by itself prove fading, dependence, acquisition, or harm.

Fictional completed example

These numbers are invented for arithmetic review. Case R is not a real client, and the target, prompt procedure, timing, criterion, and decision are not recommendations.

Case R has a fictional goal involving an accessible message to ask for a missing project material. The defined response may be spoken, signed, selected, or generated with the person's usual AAC. Ordinary access to the communication system is not scored as a prompt. Eighteen opportunities were planned under procedure version 2.

Fictional evidenceCountTreatment in the reviewPlanned opportunities18Coverage denominatorValid initial-response opportunities16Two opportunities were blocked before the cue and remain invalidIndependent initial responses7Occurred before a programmed promptIncorrect or no response before prompting9Initial outcomes remain recordedCorrect after the planned prompt6Post-prompt numerator among the nine prompted opportunitiesIncorrect or no response after the recorded procedure3Final outcomes remain visibleFinal correct responses13Seven initial independent plus six correct after prompt

Valid opportunity coverage is 16 divided by 18, or 88.9%. Initial independence is 7 divided by 16, or 43.8%. Among the nine opportunities that received the planned prompt, correct responding after that prompt is 6 divided by 9, or 66.7%. Final correct responding is 13 divided by 16, or 81.3%.

The 81.3% final-correct value must not replace the 43.8% initial-independence value. They answer different questions. The six correct post-prompt responses also do not prove that the recorded prompt was the least assistance required, because the example did not experimentally test every less supportive alternative.

The two blocked opportunities remain outside the response denominator rather than becoming errors. The example does not establish mastery, prompt dependence, effective fading, causation, maintenance, generalization, social validity, or client benefit.

What prompting research can and cannot answer

The SWEEPS prompting-strategy decision-tool study synthesized recommendations into a structured aid and evaluated usability; its authors identified clinical utility as a needed next step. A comparison of prompting hierarchies for leisure and vocational skills used predetermined hierarchies within specific tasks and participants. Another comparison of prompting strategies across two task types illustrates that outcomes and strategy selection may depend on the response and learner characteristics. A portable video-modeling study separately recorded independent, supported, prompted, and error outcomes under its activity-specific definitions.

Together, these sources support precise local definitions and transparent records. They do not establish one universal hierarchy, delay, prompt type, fading rule, criterion, or expected result for every client or skill.

CASP's ABA Practice Guidelines page currently points to Version 3.0 and explains how the licensed material may be accessed. The worksheet neither reproduces that material nor certifies that a practice complies with it. Review the licensed document itself when it applies.

Preserve communication and ordinary access

The current ASHA AAC practice guidance provides context for multimodal communication, vocabulary, positioning, and access. An AAC device, interpreter, mobility support, sensory support, visual schedule, prescribed equipment, corrective lenses, or other ordinary support is not automatically a prompt. Define the target around the person's authorized access conditions.

Do not withhold communication, breaks, food, water, bathroom access, mobility, prescribed care, or emergency help to create an unprompted opportunity. Record assent, dissent, requests to pause, and changes in access. Physical or intrusive assistance requires the applicable competence, authorization, consent, assent, safety, and role boundaries; a blank template cannot supply them.

Protect records and scope

Current controlling documents are linked from the BACB Ethics Codes hub. Topics in the Ethics Code for Behavior Analysts include competence, confidential information, records, assessment, behavior-change programs, consent, welfare, and accurate data. Which provisions apply depends on the practitioner's role and facts; this template supplies neither an ethics ruling nor legal advice.

For regulated organizations, HHS guidance on Security Rule risk analysis calls for assessing risks and vulnerabilities affecting electronic protected health information while recognizing that methods vary. That guidance does not approve this sheet or any storage product. Keep records only in authorized locations and follow the applicable access, correction, audit, export, backup, retention, disclosure, and disposal rules.

Phase and correction log

DateProcedure versionOriginal entryCorrected entry or new phaseReason and source evidenceAuthorized byOpportunities affectedSummaries affected

Preserve the original initial response when correcting a later field. Label late entries. Never backdate a hierarchy or silently transform a corrected response into an independent one.

Stop and route concerns

Unclear target definitions, opportunity rules, initial-response windows, prompt descriptions, decision logic, correction procedures, authorizations, safety plans, ownership, or record destinations require clarification before the data are interpreted. Acute risk, distress, withdrawn assent, inaccessible communication, unwanted effects, major implementation departures, possible data alteration, privacy exposure, and work beyond competence belong in the appropriate clinical, emergency, organizational, privacy, security, compliance, legal, payer, or regulatory route.

A prompt-level summary is a review trigger, not an automatic fading or treatment decision. Keep the trial-level evidence, deviations, access conditions, and unresolved questions available to the qualified reviewer.

Related resources

Sources