An ABA graduated guidance worksheet should show what happened at each task step: independent movement, provider shadowing without touch, brief contact, fuller physical guidance, an error, or an invalid opportunity. “Completed the routine” is too broad to answer whether support was faded or whether contact remained acceptable.

Use this ABA graduated guidance worksheet to define a case-specific arrangement for a chained motor skill, preserve communication and contact boundaries, and review independence separately from implementation. The filled example tests the arithmetic without prescribing graduated guidance.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

Clinical and contact boundary: Graduated guidance may involve physical proximity or touch. A qualified team must determine whether the procedure is appropriate, within scope, authorized, medically and physically safe, accessible, and acceptable to the person. Explain the arrangement accessibly and respond to assent, dissent, refusal, freezing, pulling away, pain, fatigue, help, break, and stop communication. Never use this worksheet to authorize restraint, blocking, forced compliance, or contact that exceeds the written plan. See the current BACB ethics resources for professional context.

Graduated guidance is moment-to-moment response support

Graduated guidance is a response-prompting procedure for chained skills. The implementer stays positioned to help, supplies assistance when needed, and reduces it as the learner performs. Unlike a fixed ladder, support can change within a step as movement begins, stops, or becomes independent.

Because “partial physical” can describe very different contact, record body location, duration, boundaries, release cues, and actual delivery. Shadowing without touch is not physical prompting; brief contact is not independent merely because the step was finished.

The UNC graduated-guidance preparation module describes preparation around a defined chained skill, prompting, fading, and data collection. It is a teaching resource, not permission to touch a person or a universal protocol. Local clinical, organizational, educational, legal, and safety requirements still apply.

Meaningful purpose and authorization card

Start with the person’s goal, natural context, participation in selecting or declining the skill, existing access, and qualified decision owner. Staff convenience does not make a routine meaningful.

FieldCase-specific entryPerson’s description of the meaningful skill or priorityAuthorized plan, assessment, and current versionNatural setting, cue, materials, and outcomeTask-analysis version and total planned stepsAccessible response forms and communication supportsMotor, sensory, pain, fatigue, mobility, or medical considerationsWhy graduated guidance is being consideredAlternatives considered, including no-contact supportsConsent requirements and how assent/dissent will be checkedContact that is permitted, not permitted, or requires separate reviewQualified clinical owner and applicable interdisciplinary reviewersEffective date, review date, and superseded version

Other professionals may hold relevant information or authority depending on the task and setting. Task performance alone cannot establish medical, biomechanical, feeding, swallowing, mobility, or communication safety.

The person’s access to food, water, a bathroom, mobility, communication, relationships, rest, health care, safety, and emergency help is never contingent on completing a teaching routine.

Contact and positioning plan

Write the contact plan in understandable language. “As needed” is insufficient. If physical guidance is not permitted or welcomed, use another arrangement.

ElementPlanned definitionProhibited or hold conditionPerson’s accessible explanationImplementer positionShadowing distance and hand positionBrief-contact topographyFuller controlling-contact topographyMaximum duration or pressure boundaryIncrease-support cueDecrease-support cueImmediate release cueHelp, alternative, break, refusal, and stop responseHealth, pain, fatigue, or safety interruption

Stop contact at the written stop condition. Record tensing, freezing, pulling away, protecting a body area, or other established dissent, then follow the individualized plan. Never relabel dissent as an error or missing data to preserve a trial.

The Ethics Code for Behavior Analysts addresses rights, competence, understandable communication, informed consent, assent when applicable, confidentiality, documentation, and data use. No ethics source by itself authorizes contact; the qualified team remains responsible for applicable requirements and individual circumstances.

Version the task analysis and step-specific support

Task steps can be too large, physically difficult, culturally irrelevant, inaccessible, or material-dependent. Version the sequence. Identify where no-contact supports, adaptive equipment, or specialist input belong.

StepNatural cueObservable responseAccessible alternativePlanned starting position/supportIncrease ruleRelease/decrease ruleContact restrictionSafety/access note1234

Define each step observably. Avoid requiring eye contact, posture, speech, or speed unless the feature is genuinely necessary and justified. If an AAC selection, gesture, adapted grip, seated movement, or partner-assisted action meets the function, record it as an accepted response rather than a prompt to fade.

The related ABA chaining worksheet can hold the broader task-analysis and acquisition plan. This page focuses on the changing response support and contact evidence within that plan.

Use mutually exclusive first-support codes

Choose codes before data collection so one step cannot become both independent and physically prompted. Adapt the sample labels as needed.

CodeWorking definitionIncludeDo not inferIStep completes without provider shadowing counted as a prompt or physical contactAccepted accessible response and natural supportsMastery, generalization, preference, or causation.SProvider is in the defined shadow position, but no contact occurs before step completionProximity arranged specifically for immediate supportPhysical prompting or full independence under a definition that excludes shadowing.BCBrief contact occurs and is released under the written ruleDefined touch topography and durationIndependent completion.FCFuller controlling contact occurs within the authorized topographyThe planned controlling physical promptConsent, assent, safety, or treatment benefit.EThe first response meets the written error definitionObservable response, not a trait labelWhy the response occurred.NRNo defined response occurs within the accessible intervalSufficient opportunity and available accessRefusal, inability, or lack of motivation.INVThe opportunity cannot answer the review questionWrong materials, inaccessible cue, missing AAC, unplanned contact, interrupted step, recording failure, or other predefined reasonLearner error.

If support changes within a step, retain the sequence and apply the primary code according to the written rule. A reviewer may need both the highest contact level and the first support. Do not revise the rule after seeing the result.

Step-opportunity ledger

Use one row per planned step opportunity. Record what was planned, what happened first, the highest support delivered, contact duration when relevant, the person’s communication, and whether the opportunity remained valid.

Date/timeRoutine/stepNatural cuePlanned startValid? reasonFirst-support codeHighest supportContact site/durationStep outcomeHelp/break/stop/dissentPain/fatigue/accessConsequence/correctionImplementeryes / noI / S / BC / FC / E / NR / INV

Keep invalid opportunities in the ledger. Missing materials, absent cues, unavailable mobility aids, unplanned contact, and recording failures are implementation or access facts, not learner errors. If a step stops for safety or at the person’s request, preserve that response even when the row becomes invalid.

Summarize steps without hiding contact

Step or routinePlannedInvalidValidIndependentShadow onlyBrief contactFull contactErrorNo response/otherPerson requested changeNotesTotal

Totals should reconcile. Under mutually exclusive coding, independent + shadow only + brief contact + full contact + error + no response/other equals valid opportunities. If the clinical record instead codes support and outcome on separate dimensions, state that plainly and do not add overlapping categories as though they were exclusive.

A high completion percentage can coexist with frequent contact; declining contact can coexist with distress, inaccessible materials, or poor maintenance. No single number decides whether to continue.

Preserve the person’s experience in a separate record

Implementation data do not speak for the person. Ask accessibly and allow delayed, nonvocal, device-based, or partner-supported communication. Task completion or absent resistance does not establish assent.

Review dateRoutine/stepHow the arrangement was explainedCommunication access availablePerson’s description or responseContact welcomed, tolerated, declined, or unclearHelp/break/stop honoredBurden or unwanted effectRequested changeFollow-up owner

The ASHA AAC Practice Portal describes AAC as individualized and multimodal. Keep the communication system available during teaching. Do not guide a person’s hand to a device response and then treat the selection as their assent, preference, or independent communication.

When a person cannot provide legal consent, applicable consent requirements still belong with the authorized representative and qualified team. Assent and dissent remain separate person-centered evidence rather than a substitute signature.

Keep each denominator attached to its question

Review questionNumeratorDenominatorResultDoes not establishValid coveragevalid step opportunitiesall planned step opportunitieslearner accuracy or safetyIndependent-step rateindependent stepsvalid step opportunitiesgeneralization or causationNo-contact completionindependent plus shadow-only stepsvalid step opportunitiesabsence of burden or complete independenceContact usebrief-contact plus full-contact stepsvalid step opportunitiesinappropriate or appropriate contact by itselfPlanned independent yieldindependent stepsall planned step opportunitieswhy opportunities were invalidStep-specific supportcode at one stepvalid opportunities for that stepperformance on the whole chainContact-plan integritycorrectly implemented contact componentsobservable contact-component opportunitiespermission, assent, or benefitFull procedure fidelitycorrectly implemented componentsall observable component opportunitiesmastery, maintenance, or coverage

Carry raw counts next to percentages. Ten independent steps mean something different across 12 versus 80 valid opportunities. Retain the task-analysis version, context, contact permission, health status, and exposure history.

Fictional greenhouse watering-station example

Leila, a fictional adult, chose a greenhouse watering-station routine. The example assumes qualified review of relevance, contact permission, motor and medical considerations, AAC, assent, dissent, and alternatives. It demonstrates arithmetic only.

The team planned 12 routines with eight steps each, for 12 × 8 = 96 planned step opportunities. Six opportunities were invalid: two involved missing equipment, one used an outdated task-analysis version, one began without Leila’s communication device in reach, and two contained recording or cue errors. That left 96 − 6 = 90 valid steps, or 90 / 96 × 100 = 93.8% valid coverage.

Among the 90 valid steps, 52 were independent, 18 were completed while the provider shadowed without contact, 14 involved brief contact, and six involved the defined fuller contact prompt. The categories reconcile: 52 + 18 + 14 + 6 = 90.

First-support codeValid stepsPercentage of valid stepsBounded readingIndependent5252 / 90 × 100 = 57.8%Independent under this version and definitionShadow only1820.0%Provider proximity occurred; no contact occurredBrief contact1415.6%Defined brief contact occurredFull contact66.7%Defined controlling contact occurredTotal90100.0%Categories remain mutually exclusive

Steps completed without contact were independent plus shadow-only steps: 52 + 18 = 70, so 70 / 90 × 100 = 77.8% of valid steps occurred without contact. Steps with contact were 14 + 6 = 20, so 20 / 90 × 100 = 22.2%. Planned independent yield was 52 / 96 × 100 = 54.2%, which keeps invalid opportunities visible.

These sequential observations have no comparison condition and do not show that graduated guidance caused learning. They also cannot establish that contact was wanted, comfortable, safe, or necessary; those questions belong in the separate authorization, experience, and clinical records.

Fidelity record for changing support

ObservationCorrect version availablePosition matchedNatural cue preservedIncrease rule followedRelease/decrease timelyContact within planResponse code accurateAAC/access preservedHelp/stop honoredCorrect componentsOpportunitiesNotes

In the fictional example, 14 observations across nine components created 14 × 9 = 126 component opportunities. Reviewers scored 117 correct components and nine mismatches, which reconcile to 126. Descriptive fidelity was 117 / 126 × 100 = 92.9%.

Review mismatches individually. Late release, out-of-plan contact, missing AAC, and failure to honor stop communication are not interchangeable. An overall percentage cannot erase a serious event.

The 2024 study Teaching Caregivers to Use Graduated Guidance Using Video Modeling operationalized position, prompt changes, and caregiver implementation in three dyads. Its open-access full text shows how implementer accuracy can be measured separately. The small sample does not establish benefit, safety, acceptability, or a default procedure. Fidelity, learner outcome, and experience remain different evidence layers.

Review maintenance, generalization, and procedure fit

FieldReview entryTask-analysis and graduated-guidance versions reviewedMeaningful purpose and continued authorizationPlanned, invalid, and valid step opportunitiesIndependent, shadow-only, brief-contact, full-contact, error, and no-response patterns by stepContact-plan and full-procedure integrity limitsPerson’s communication, assent/dissent, choice, and requested changesPain, fatigue, sensory, motor, accessibility, or other unwanted effectsMaintenance interval and intervening practiceGeneralization people, settings, materials, and task variationsRetain / revise / assess alternative / pause / stop questionQualified owner, rationale, effective date, and next review

A published example of chaining a functional basketball sequence included graduated guidance within a broader package for one adolescent and one activity. It illustrates the need to name support levels, but cannot select a sequence for someone else.

Reopen the procedure when contact concentrates at one step, release is unreliable, implementers disagree, communication changes, burden increases, or the purpose no longer holds. The qualified team may revise materials or the task analysis, use adaptive equipment, select a no-contact method, seek interdisciplinary review, pause, or stop.

Document the decision and preserve the earlier version. Revision does not rewrite past performance.

Related resources

Sources

The sources differ in learner age, task, implementer, prompt topography, setting, design, and outcome. They support explicit task definitions, support codes, contact boundaries, direct measurement, fidelity, and cautious interpretation. They do not define a universal starting prompt, body location, contact force, fading pace, release rule, trial count, mastery criterion, safety conclusion, or expected effect.