Task Analysis Chaining

職業

分類

項目管理

Consult experts: For specialized tasks (e.g., workplace routines, cooking techniques), consult someone with expertise in that domain.

Review existing task analyses: Published TAs exist for many common self-care and daily living skills. Adapt these to the individual learner and environment.

Trial-and-error refinement: Implement a draft TA, observe where the learner struggles, and modify — adding intermediate steps where gaps exist or combining steps where the learner easily completes multiple actions.

Validation

Validate the TA by having 2–3 different people perform the skill following only the written steps. If any performer cannot complete the skill using the TA alone, revise the steps. Each step should:

• Begin with an observable action verb
• Describe one discrete behavior (not multiple actions)
• Be sequenced in the correct order
• Include decision points where relevant ("If the water is too hot, turn the cold handle clockwise")

Step Size (Grain Size)

The number and size of steps depend on the learner:

• Finer grain (more steps, smaller actions): For learners with significant skill deficits, motor challenges, or limited attention spans
• Coarser grain (fewer steps, larger actions): For learners who can already perform some component skills or who learn quickly

A handwashing TA might be 8 steps for one learner and 22 steps for another. Adjust based on baseline assessment data.

Behavior Chains

A behavior chain is a sequence of responses in which each response produces a stimulus change that serves as:

1. A conditioned reinforcer for the response that produced it (the step just completed)
2. A discriminative stimulus (S^D) for the next response in the chain

The final response in the chain produces the terminal reinforcer (natural consequence of completing the task). This is what gives backward chaining its power — the learner always ends the chain by contacting the natural reinforcer.

Chaining Procedures

Forward Chaining

Teach the first step of the chain to mastery, then the first and second steps together, then the first three, and so on until the learner independently performs the entire chain.

Procedure:

1. Assess baseline: present the entire TA and score each step as independent, prompted, or not performed
2. Teach step 1 to criterion using appropriate prompting and reinforcement
3. Once step 1 is mastered, teach steps 1 + 2 (the learner performs step 1 independently, then receives instruction on step 2)
4. Continue adding one step at a time until the full chain is mastered
5. The clinician completes all untrained steps for the learner (or physically guides through them)

Advantages: Logical sequence matches the natural order of the skill. Each new step builds directly on prior mastery.

Best for: Skills where the initial steps are motivating or where the learner has some steps mastered at baseline. Academic task sequences, assembly tasks, cooking recipes.

Backward Chaining

Teach the last step of the chain first. The clinician completes all preceding steps, then the learner performs the final step independently and contacts the natural reinforcer. Then teach the last two steps, and so on.

Procedure:

1. The clinician performs (or guides the learner through) all steps except the last
2. Teach the last step to criterion
3. Once mastered, the clinician performs all steps except the last two
4. Teach the last two steps in sequence
5. Continue adding steps from the end backward until the learner completes the entire chain

Advantages: The learner contacts the natural reinforcer on every training trial from the very first session. High reinforcement density. The most recently taught step is always closest to reinforcement.

Best for: Self-care skills (dressing, toothbrushing, shoe tying), skills where the terminal reinforcer is strong, learners who need high reinforcement density.

Backward Chaining with Leaps Ahead

A modification in which the clinician skips ahead in the chain if baseline data show the learner can already perform certain steps. Rather than rigidly adding one step at a time, the clinician adds blocks of steps the learner has demonstrated competence on. Reduces training time for learners with partial repertoires.

Total Task Presentation (Whole Task)

The learner attempts every step of the chain on every trial. The clinician provides prompts as needed at each step where the learner cannot perform independently. Prompts are faded across sessions.

Procedure:

1. Present the entire task on every trial
2. Provide the necessary prompt at each step (using a consistent prompting strategy — MTL, LTM, or graduated guidance)
3. Collect data on which steps are independent, prompted, or incorrect
4. Fade prompts for each step as the learner demonstrates mastery
5. Criterion: all steps performed independently across consecutive sessions

Advantages: The learner practices the full chain every trial, experiencing the natural flow of the task. No artificial breaks in the chain. More naturalistic.

Best for: Learners who can already perform many steps in the chain, skills with many steps that the learner partially knows, when natural flow of the task is important for learning.

Selecting a Chaining Method

Data Collection for Chaining

Step-by-Step Recording

For each training session, score every step of the TA:

• + (Independent): Correct response within the response interval without a prompt
• P (Prompted): Correct response following a prompt (note prompt level)
• – (Incorrect): Error or no response even after prompting
• N/A: Step completed by clinician (in forward or backward chaining for untrained steps)

Graphing

• Graph percentage of steps independent per session
• For forward/backward chaining, graph can show cumulative steps mastered
• Visual analysis of the graph guides decisions about step additions, prompt fading, and TA modifications

Mastery Criteria

Define mastery for each step (e.g., independent on 3 consecutive trials) and for the full chain (e.g., 100% independent across 2 sessions with 2 different trainers).

Modifying Task Analyses

Modify the TA based on ongoing data:

• Add steps: If the learner consistently fails at a step, break it into smaller components
• Combine steps: If the learner consistently performs multiple steps as a unit, combine them
• Reorder steps: If the original sequence is not optimal for this learner or environment
• Add decision points: For tasks with conditional branches (e.g., "if the shirt is inside out, turn it right-side out")
• Individualize materials: Adapt to the specific items the learner will use (their clothing, their kitchen, their workplace)

Clinical Examples

Self-care (handwashing): 12–18 steps, backward chaining, graduated guidance, mastery at 100% independent across home and school bathrooms.

Vocational (assembling a product): 8–25 steps depending on complexity, forward chaining, model prompts, generalization across materials and supervisors.

Academic (long division): 6–10 steps, total task presentation with visual TA as a permanent prompt, faded to independent completion.

Community (ordering food at a restaurant): 10–15 steps, total task with video model, practice in simulated and real settings.

Key References

• Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied Behavior Analysis (3rd ed.). Pearson.
• Bellamy, G. T., Horner, R. H., & Inman, D. (1979). Vocational Habilitation of Severely Retarded Adults: A Direct Service Technology. University Park Press.
• Spooner, F. (1984). Comparisons of backward chaining and total task presentation in training severely handicapped persons. Education and Training of the Mentally Retarded, 19(1), 15–22.
• Slocum, S. K., & Tiger, J. H. (2011). An assessment of the efficiency of and child preference for forward and backward chaining. Journal of Applied Behavior Analysis, 44(4), 793–805.
• Walls, R. T., Zane, T., & Ellis, W. D. (1981). Forward and backward chaining, and whole task methods. Behavior Modification, 5(1), 61–74.