An ABA chaining plan worksheet should connect a meaningful routine to a versioned task analysis, the chosen teaching arrangement, and evidence at both the whole-routine and step levels. A summary such as "used backward chaining, 80%" is not enough to reconstruct which steps were available, taught, prompted, completed by someone else, or excluded from the denominator.

This copyable worksheet preserves those distinctions so a qualified reviewer can examine one individualized plan. It does not rank forward, backward, or total-task chaining, and it does not convert a task-analysis score into a treatment decision. The routine, method, supports, data rules, and review thresholds all require case-specific assessment and approval.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

Tie the chain to a routine the client chose

Begin with what the completed routine makes possible for the person. The same outward sequence can have different value, burden, privacy implications, and access needs for different clients. Record accepted ways of performing each step before collecting data.

Plan identityComplete before implementationWhich meaningful routine did the client choose?Where does the routine clearly start and finish?Which task-analysis version is effective?Which settings, materials, people and variations are in scope?Which accessible step forms count?Which ordinary communication, mobility, sensory and medical supports remain?Which assessment and interdisciplinary recommendations apply?How are consent, assent, dissent, withdrawal and stopping handled?Who authored and approved the plan, and when is review due?

The BACB Sixth Edition BCBA Test Content Outline lists chaining, trial-based procedures, generalization, maintenance, observable goals, client preferences, contextual fit, procedural integrity, and data-based modification as distinct areas. It is examination content, not a client protocol. The BACB Test Content Outlines hub identifies the current outline.

Freeze a task-analysis version before selecting the method

A task analysis is a testable description of the routine, not a permanent truth. Observe the routine with the actual person, materials, setting, and access conditions. Split or combine steps when boundaries cannot be scored reliably. Issue a new version when the sequence, materials, cue, or response requirements change.

StepWhat is the natural cue?What completes the link?Which accessible forms count?Which ordinary supports remain?Which variations or exceptions apply?123456

The person should retain access to applicable speech, gesture, sign, AAC, writing, visual supports, positioning, and assistive technology. The ASHA AAC Practice Portal says people who use AAC should always have access to their communication tools or devices. Ordinary access is not a prompt level to remove.

Show how teaching and completion divide across the chain

Forward, backward, and total-task arrangements differ in which links are targeted and how the rest of the chain is completed. Write the selected arrangement as a case hypothesis. For every version, state what the learner may do, what the implementer completes, and how the opportunity ends. Also state whether the learner can elect to finish a link that is not currently designated for teaching.

Plan elementCurrent approved versionWhich chaining arrangement was selected, and why for this case?Which link or links are being taught now?Who completes untaught links, and how?Which prompt, timing and fading reference applies?What follows a taught link?What follows completion of the full routine?How are errors, interruptions and no responses handled?How are practice and context sampling scheduled?Which evidence triggers continue, hold, revise, return, stop or refer?

A small comparison of completion strategies within behavior chains examined manual guidance, trainer completion, and no completion of untrained steps under defined experimental conditions. It supports recording how untrained links are handled. It does not establish one completion strategy or chaining method for every routine.

Keep the routine record underneath every step score

Record one row for every planned routine opportunity, including invalid or interrupted attempts. A valid routine can still contain step-level implementation mismatches, prompted completions, or incomplete links. Keep those states visible instead of collapsing the row into one percentage.

Date/timeWhich version?Was the routine valid?What was each step code?Which prompt and consequence occurred?Who completed other links?Was the full routine independent?What access, assent/dissent or context mattered?Yes / No, reasonI / P / N / U / X by stepYes / No / Not scorable

Suggested codes are I, independently completed after the natural cue; P, completed after the planned extra prompt; N, not completed within the defined opportunity; U, unavailable because a required material or access condition was missing; and X, invalid because the procedure was not delivered as written. Define any local codes in the plan. A prompted step is not independently completed, and an unavailable step is not an error.

Do not mix whole-routine and step denominators

The fictional example uses a six-step library-return routine selected for practice across 10 scheduled routine opportunities. One opportunity is invalid because the return materials are unavailable, leaving nine valid routines and 9 × 6 = 54 valid step opportunities. Across those step opportunities, 31 are independently completed, 17 are completed after the planned prompt, and six are not completed. Three of the nine valid routines are independent from beginning to end.

Whole-routine reviewCountScheduled routine opportunities.10Valid routine opportunities.9Fully independent valid routines.3Other valid routines, which may contain prompted or incomplete links.6Invalid because required materials were unavailable.1

Step-level reviewCountValid step opportunities.54Independently completed.31Completed after the planned prompt.17Not completed.6

  • Valid routine coverage is 9 / 10 × 100 = 90.0%.
  • Independent step completion is 31 / 54 × 100 = 57.4%.
  • Final step completion is (31 + 17) / 54 × 100 = 88.9%.
  • Whole-chain independence is 3 / 9 × 100 = 33.3%.

These four values answer different questions. The first describes observation coverage. The second describes independent links. The third includes prompted links. The fourth asks whether the entire valid routine occurred independently. Do not average them, replace one with another, or convert 48 completed steps into a count of complete routines. The six other valid routines need their row-level patterns before anyone can say whether the chain finished. None of these values establishes mastery, treatment effect, generalization, maintenance, benefit, assent, or readiness to change the plan.

Use the ABA chaining plan worksheet to move from each summary back to its routine rows. A step percentage can conceal where the chain breaks, and a whole-routine percentage can conceal stable gains at particular links.

Inspect each implementation component before judging progress

Before interpreting performance, inspect whether the current task-analysis and chaining versions were implemented. A global integrity score can hide a repeated error in one consequential component. A recent practitioner guide to procedural fidelity recommends defining measurable components, planning observation, collecting and interpreting fidelity data, and acting on the result.

Implementation componentOpportunities observedImplemented as writtenMismatch patternAction owner/dateWere materials and the natural cue available?Was the correct task-analysis version used?Were taught links presented as planned?Were untaught links completed as planned?Did prompt timing and topography match?Did consequence delivery match?Did access and communication supports remain available?Was a stop, withdrawal or safety response honored?

A single total can still be reported, but retain each component count. If the same cue, prompt, or completion error repeats, a high aggregate value should not erase it.

Look for burden, access barriers and real-world variation

Chaining often involves everyday routines where motor demands, privacy, fatigue, sensory conditions, language, cultural expectations, and available assistance matter. A forward-chaining study of a hygiene routine illustrates a tightly defined task and participant context. Its procedures cannot be copied into a different routine without individualized review.

Review questionEvidence inspectedWhat remains uncertainContinue, hold, revise, return, stop or referQualified ownerIs the routine still meaningful to the client?Are accepted step forms accessible?Are task variations represented rather than scored as errors?What assent, dissent, withdrawal, distress or burden occurred?Are health, motor, communication or safety questions unresolved?Did performance differ by implementer, material or setting?

The live BACB Ethics Codes hub and Ethics Code for Behavior Analysts inform competence, dignity, client involvement, consent and assent when applicable, documentation, data accuracy, risk, written programs, and continual evaluation for covered certificants. They do not decide which chaining arrangement is appropriate here. The CASP ABA Practice Guidelines access page identifies Version 3.0 and its licensing conditions; no licensed guideline text is reproduced and compliance is not claimed.

A two-minute pre-session check

  • The client-selected purpose is still current.
  • Today's routine is within the approved scope.
  • The active task-analysis version is visible.
  • Required materials are available before the opportunity begins.
  • Natural cues match the written step definitions.
  • Accepted response forms remain unchanged.
  • Ordinary access supports are ready for use.
  • AAC stays available throughout the routine.
  • The teaching link matches the chaining version.
  • Untaught links have an identified completion plan.
  • Prompt timing can be implemented as written.
  • Programmed consequences are available and appropriate.
  • Data codes are understood by the implementer.
  • Invalid and unavailable states can be recorded.
  • Assent and dissent indicators are known.
  • Withdrawal can be honored without data recoding.
  • Safety and privacy conditions have been checked.
  • The implementer knows who may revise the plan.
  • A recent change has a new effective version.
  • Questions that require referral are held for review.

Keep exceptions and plan changes in the audit trail

DateWhat triggered review?Which evidence was reviewed?What input was received?Which options were considered?What was decided, and why?Which version takes effect?Who owns the decision?Continue / hold / revise / return / stop / refer

Preserve equipment failures, unavailable materials, interruptions, health changes, step-sequence disputes, withdrawal, privacy concerns, and implementation mismatches as their own records. Do not relabel them as learner errors. When the task analysis, method, taught link, prompt, consequence, or scoring rule changes, start a new version and keep earlier rows attached to the version in effect at the time.

Use this worksheet beside the current assessment, treatment plan, consent and assent process, safety plan, authorization, applicable law, payer requirements, and interdisciplinary recommendations. It is an educational drafting and review aid, not an automated chaining protocol or a replacement for qualified clinical judgment.

Related resources

Sources