An ABA prompt dependence worksheet should describe an observable pattern, not turn “prompt dependent” into a trait of the person. It should show what happened after an accessible natural cue, before a programmed prompt, after help was requested, and when the planned observation could not be interpreted.

Use this copyable ABA prompt dependence worksheet to organize a qualified review of signals, natural-cue observations, and possible support-reduction questions. It separates ordinary teaching from deliberately planned natural-cue observations, preserves invalid conditions, and keeps final correctness apart from independence. The worksheet does not tell a team to remove support.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

Clinical and access boundary: A qualified team must decide whether any observation or change is meaningful, assessed, authorized, within scope, accessible, and safe. Preserve AAC, communication, sensory, mobility, visual, hearing, literacy, language, and other ordinary access. Honor help, break, stop, refusal, freezing, pulling away, distress, pain, and fatigue under the person’s plan. This worksheet does not authorize prompt withdrawal, delayed help, physical contact, response blocking, forced compliance, or withholding an essential support. Current BACB ethics resources provide professional context, not a case decision.

Define the signal instead of labeling the person

A prompt-dependence signal is a task- and condition-specific pattern in which the defined response occurs reliably after an added prompt but occurs less often after the accessible natural cue alone. That pattern raises questions about cue clarity, prompt timing, learning history, consequences, task meaning, response effort, setting, access, and implementation. The pattern alone cannot explain why it exists.

Call the observation what it is: “independent responses occurred on 4 of 12 valid natural-cue opportunities in this routine and version.” Avoid “the client is prompt dependent.” A person can respond independently in one task and wait for help in another. Help seeking can be functional self-advocacy, not evidence of failure.

The BACB test-content outlines and Sixth Edition BCBA Test Content Outline locate prompts, stimulus control, measurement, integrity, maintenance, and generalization within behavior-analytic practice. They do not define a universal probe, delay, or support-reduction rule.

Meaningful goal, authorization, and access card

Start with why the response matters to the person. A natural-cue observation is not justified merely because staff want fewer prompts.

FieldCase-specific entryPerson’s description of the meaningful outcome. Record.Participation in selecting, changing, or declining the goal. Explain.Natural setting, routine, cue, response, and outcome. Define.Accepted response forms that meet the same purpose. List.Ordinary AAC, communication, sensory, motor, literacy, and mobility access. Preserve.Current assessment, plan, version, and authority. Identify.Existing prompts, timing, corrections, and consequences. Describe.Baseline signal and competing explanations. Summarize.Consent requirements and assent or dissent process. Clarify.Medical, motor, communication, educational, or other review needed. Route.Qualified clinical owner and collaborators. Name.Effective date, review date, and superseded version. Date.

The ASHA AAC Practice Portal describes AAC as individualized and multimodal. A speech-generating device, communication board, keyboard, visual schedule, adapted tool, or partner-supported access may be part of the natural environment. Do not remove it to manufacture an “independent” test.

Operationally define every observation state

Write definitions that another trained observer can use. “Waited for me” is not enough. Specify the cue, response, response window, prompt onset, and what makes an opportunity invalid.

ElementObservable definitionExampleExclusion or invalid conditionNatural cue. Define.Defined response. Observe.Accepted accessible response. Include.Initial response window. Time.Independent response. Bound.Correct after programmed prompt. Separate.Help request. Preserve.Decline, break, or stop. Preserve.Error or no response. Distinguish.Valid observation. Define.Invalid observation. Define.

An early undocumented prompt can erase the natural-cue observation. Missing materials, unavailable AAC, a changed instruction, interrupted routine, unplanned cue, health concern, or safety stop may make the row invalid under the prospective rule. Retain the event and reason instead of converting it to a learner error.

Separate the review question from the teaching procedure

A team may compare observations only when doing so is clinically appropriate and prospectively defined. “Natural-cue review” here means the approved condition written by the qualified owner. Natural-cue review does not mean abruptly removing support, waiting through distress, or creating failure.

ConditionQuestionCue and ordinary accessProgrammed prompt availabilityHelp responseEnd or invalid ruleOrdinary teachingWhat happens under the current plan?Natural-cue reviewWhat occurs under this approved observation condition?Maintenance checkWhat remains after the stated interval and exposure?Generalization checkWhat occurs with this person, place, material, or cue?

The assessment-based instruction tutorial discusses prompt presence, prompt type, and reinforcement contingencies as variables that may be relevant when assessing prompt-dependent behavior. That supports writing a focused question and matching observations to it. It does not establish which variable matters in a particular case.

Observation ledger

One row is one planned observation. Record the initial result before a prompt and the final result separately.

Date/timeTarget/context/versionConditionNatural cue and accessValid? reasonInitial responsePrompt delivered and timeFinal responseHelp/decline/stopLatencyBurden or unwanted effectConsequenceObserverordinary / natural-cue / follow-upyes / noindependent / help / decline / error / no response

Never overwrite the initial response with a corrected response. A response after a programmed prompt can be important progress, but it is not independent under the natural cue. A help request, alternative response, or decline belongs in its own field rather than being hidden inside “incorrect.”

Condition summary with distinct denominators

Condition and versionPlannedInvalidValidIndependent before promptCorrect after promptHelpDeclineError/no responseFinal correctNotesOrdinary teachingNatural-cue reviewnot applicable or labeledFollow-up

For each condition, mutually exclusive initial states should reconcile to valid opportunities. Final correctness is a later outcome and must not be added to the initial-state total. Report counts with percentages and name the denominator every time.

Review questionNumeratorDenominatorResultDoes not establishObservation coveragevalid observationsplanned observationsperformanceNatural-cue independenceindependent natural-cue responsesvalid natural-cue observationscausation or generalizationOrdinary-teaching independenceindependent responses before promptsvalid teaching observationstreatment effectPrompted correctnesscorrect responses after promptsvalid prompted opportunitiesindependenceFinal teaching correctnessindependent plus correct-after-prompt responsesvalid teaching observationsnatural-cue controlHelp or declinehelp or decline eventsvalid observations in the stated conditionlack of skillFidelitycorrectly implemented componentsobservable component opportunitiesappropriateness or benefit

Fictional accessible transit-card refill example

Jules is a fictional adult learning to refill a transit card through an accessible kiosk. The team defined the kiosk balance notice as the natural cue, accepted touch screen or switch access, kept the visual sequence and AAC available, and approved two observation conditions. The example tests arithmetic, not the target, prompt arrangement, or a decision threshold.

There were 48 planned observations and six invalid observations. Two had the switch interface disconnected, one opened the wrong kiosk screen, one included an early spoken cue, one was interrupted by a system restart, and one ended for a reported headache. That leaves 48 - 6 = 42 valid observations, or 42 / 48 x 100 = 87.5% valid coverage.

Eighteen valid observations used the approved natural-cue review condition. Ten were independent, four included a help request, and four ended in error or no response. The categories reconcile: 10 + 4 + 4 = 18. Natural-cue independence was 10 / 18 x 100 = 55.6%.

Twenty-four valid observations occurred during ordinary teaching. Eight were independent before a prompt, 13 were correct after a programmed prompt, one included a help request, and two ended in error or no response. These initial or final-path categories reconcile: 8 + 13 + 1 + 2 = 24. Independence before prompting was 8 / 24 x 100 = 33.3%. Final correct responding was (8 + 13) / 24 = 21 / 24 x 100 = 87.5%.

Condition resultCountDenominatorPercentageBounded meaningValid coverage4248 planned87.5%Interpretable observations under the rulesNatural-cue independent1018 valid natural-cue reviews55.6%Independent in that defined conditionTeaching independent before prompt824 valid teaching observations33.3%Initial independence under current teachingTeaching final correct2124 valid teaching observations87.5%Includes 13 prompted responses

The difference between conditions is descriptive. It does not prove that one condition caused responding, that prompts should be removed, or that the observations had equal difficulty, exposure, motivation, implementers, or context. The 87.5% final-correct value cannot replace the 33.3% teaching-independence value.

Fidelity and observation-quality record

ObservationCorrect versionCue and access presentCondition followedInitial window preservedPrompt timing exactInitial/final codes separateHelp/decline honoredInvalid rule appliedContext recordedCorrectOpportunitiesNotes

In the fictional review, 14 observations across nine components created 14 x 9 = 126 component opportunities. Reviewers scored 117 correct and nine mismatches; 117 + 9 = 126. Descriptive fidelity was 117 / 126 x 100 = 92.9%.

Inspect each mismatch. A premature prompt destroys the initial-response measure. Missing AAC changes access. A missed stop request is not made acceptable by a high aggregate. Fidelity says whether the written condition occurred, not whether it was ethical, suitable, preferred, effective, or beneficial.

Experience, burden, and competing-explanation review

DomainDirect evidencePossible question, not conclusionNeeded follow-upOwnerCue clarity and discriminationPrompt timing or predictabilityReinforcement historyTask meaning or response effortCommunication, sensory, motor, or health accessPerson’s report, preference, help, or dissentImplementer or setting differencesExposure, maintenance, or generalization

The narrowly bounded prompt-dependency FCT report and task comparison involving prompt dependency illustrate that arrangements and outcomes can differ by response, task, participant, and history. They cannot justify using those procedures here. A prompting-strategy decision tool explicitly notes that one recommendation concerning prompt delay and prompt dependence had not been directly evaluated. Preserve that uncertainty.

The learner-specific prompt assessment study and positional-prompt evaluation further support measuring the actual task and arrangement. Small studies do not create a universal remedy.

Turn the evidence into a bounded review record

Review fieldEntryExact signal supported by the observations.Alternative explanations not resolved.Conditions, exposures, implementers, and versions compared.Access, health, communication, motor, and safety findings.Person’s experience, help, assent/dissent, and requested changes.Invalid observations and fidelity mismatches.Maintenance and generalization evidence.Retain / revise / assess alternative / pause / stop question.Qualified owner, rationale, effective date, and next review.

Possible next questions include whether to clarify the cue, change the task, modify access, reassess the response, collect cleaner observations, review consequences, compare an authorized alternative, seek interdisciplinary input, pause, or stop. The worksheet does not choose among them. Preserve the prior version and do not recode old observations after a plan changes.

The Ethics Code for Behavior Analysts and CASP ABA Practice Guidelines access page supply professional context. Applicable consent, assent, scope, documentation, privacy, payer, school, employment, facility, licensure, and legal requirements still require qualified review.

Related resources

Sources

These sources differ in participant, task, setting, response, prompt, consequence, design, and outcome. They support observable definitions, direct measurement, access, fidelity, and modest interpretation. They do not prescribe support withdrawal, wait duration, prompt type, consequence, contact, exposure, mastery, generalization, maintenance, or a clinical decision.