By Matt Harrington, BCBA · Behaviorist Book Club · April 2026 · 12 min read

Chaining in ABA: Building Complex Skills Step by Step

Overview & Clinical Significance

Chaining is a behavior analytic procedure used to teach multi-step skills by linking discrete responses together into a unified behavioral chain. Each response in the chain serves as both a conditioned reinforcer for the preceding behavior and a discriminative stimulus for the next response, creating a seamless sequence that terminates in terminal reinforcement. The procedure is grounded in Skinner's analysis of operant behavior and has been empirically validated across decades of applied research in the Journal of Applied Behavior Analysis.

For behavior analysts working with learners who have autism spectrum disorder, intellectual disabilities, or other developmental differences, chaining is indispensable. Tasks that appear simple to neurotypical individuals — washing hands, making a sandwich, completing a morning routine — require mastery of numerous discrete steps in a specific order. Without explicit instruction using chaining procedures, many learners fail to acquire these functional life skills.

Chaining is not a single procedure but a family of related techniques, each with distinct implementation steps, empirical support, and clinical considerations. Forward chaining teaches steps in sequence from the first to the last. Backward chaining begins with the final step and works backward toward the first. Total task chaining presents and prompts the entire chain on every trial. The choice among these variants depends on learner profile, the nature of the task, and the training context.

As a supervisory tool, chaining procedures offer BCBAs a structured framework for training ABAT and RBT staff. Teaching staff to implement behavior chains mirrors the same principles used with learners: break the skill into discrete components, demonstrate each step, provide prompts and corrective feedback, and systematically fade support as competency increases. The result is a replicable, measurable staff training model that aligns with BACB supervision requirements.

Background & Context

The theoretical foundation of chaining lies in the stimulus-response (S-R) chain model, which holds that each completed response in a sequence produces a stimulus that functions both as a conditioned reinforcer for the previous behavior and as an SD for the next. This bidirectional stimulus function creates a self-sustaining chain that moves the learner progressively toward terminal reinforcement.

Early experimental work in behavioral psychology demonstrated that complex sequences could be assembled from simpler operants. Applied researchers subsequently adapted these findings for use with clinical populations. The VB-MAPP and ABLLS-R assessment tools both include chaining-relevant skills, recognizing that independent performance of multi-step tasks is a primary goal of functional skill programming.

Task analysis is the prerequisite procedure for any chaining program. A task analysis involves breaking a behavioral chain into its component steps — each step being a discrete, observable, measurable response. Quality task analyses are developed through direct observation of competent performers, literature review, or expert consultation. The level of specificity required depends on the learner's current repertoire: a learner with limited fine motor control may require a more granular task analysis for hand washing than a learner who already has those prerequisite motor skills.

Historically, backward chaining was favored in early ABA literature because the learner always experiences terminal reinforcement at the end of each training trial, which theoretically maximizes reinforcer contact. However, subsequent research has not consistently demonstrated backward chaining superiority over forward or total task methods. The empirical literature in JABA suggests that task characteristics, learner variables, and implementation fidelity often matter more than the specific chaining variant selected.

In supervisory contexts, understanding the historical and theoretical basis of chaining allows BCBAs to make evidence-informed decisions when selecting training methods for staff and for adapting procedures when initial approaches are not producing expected progress.

Clinical Implications

Selecting the appropriate chaining variant requires clinical judgment that integrates learner data with procedural knowledge. Forward chaining is often preferred for tasks where the initial steps are less challenging or where motivation is highest at the start of the chain. Backward chaining is frequently chosen when terminal reinforcement contact is a priority or when the final steps of a chain are easiest to master. Total task chaining is well-suited for learners who already have some component skills in their repertoire and who benefit from full-chain exposure on every trial.

Prompt hierarchies are embedded within chaining programs to support acquisition. Most-to-least (MTL) prompting begins with the most intrusive level needed to evoke correct responding and systematically fades toward independence. Least-to-most (LTM) prompting starts with minimal support and increases intrusiveness only when the learner does not respond independently within a specified latency. The choice between MTL and LTM should be driven by error sensitivity: for learners where error responding is likely to be reinforced or where errorless learning is clinically indicated, MTL is typically preferred.

Data collection during chaining programs should capture step-by-step performance, not just overall completion. A task analytic data sheet allows the practitioner to identify which steps are mastered, which are in acquisition, and which are not yet showing progress. This granular data supports instructional decisions such as when to advance to the next step in forward chaining, when to add a step in backward chaining, and when to modify the task analysis itself.

Generalization and maintenance must be programmed explicitly. Learners who master a behavioral chain in a clinic setting may fail to perform it at home or school unless generalization training is incorporated. Training across multiple trainers, settings, and stimulus conditions — as well as using natural reinforcement where possible — increases the probability that chains will maintain and generalize. Supervisors should ensure that RBTs receive explicit instruction on generalization programming so that skill gains transfer to the natural environment.

Ethical Considerations

BACB Ethics Code 2.01 requires that behavior analysts use scientifically supported interventions. Chaining procedures have robust empirical support across diverse learner populations, and selecting them over less-validated approaches aligns with this ethical obligation. When a BCBA chooses to implement chaining, they should be able to articulate the evidence base and explain why this method was selected for this particular learner.

Code 2.09 addresses the right of clients to effective treatment. Chaining is among the most powerful tools available for teaching functional life skills, and withholding or inadequately implementing it can deprive learners of meaningful skill acquisition opportunities. BCBAs are responsible for ensuring that chaining programs are implemented with fidelity, that data are collected systematically, and that programs are modified when data indicate insufficient progress.

When using chaining as a supervisory training tool for RBTs and ABATs, Code 4.05 applies: supervisors must provide adequate training and oversight. Simply explaining chaining conceptually is insufficient. Supervisors must model correct implementation, observe supervisees performing the procedure, and provide specific performance feedback. Competency-based training — where the supervisee demonstrates the skill to criterion — is the standard aligned with BACB expectations.

Code 1.05 addresses the cultural responsiveness of behavior analysts. When selecting and implementing chaining tasks, BCBAs should consider whether the skills targeted reflect the learner's cultural context and family values. A chaining program for self-care tasks should incorporate the specific materials, sequences, and standards that are meaningful within the learner's household. Caregivers should be trained as partners in implementation, not simply informed of the program after the fact.

Documentation of chaining programs — including task analyses, data sheets, prompt procedures, and maintenance criteria — is required under Code 2.10. Thorough documentation ensures continuity of care and supports ethical accountability.

Assessment & Decision-Making

Before implementing a chaining program, the BCBA must conduct a thorough skills assessment to identify which steps of the target chain are already in the learner's repertoire, which are emerging, and which are absent. The ABLLS-R, VB-MAPP, and AFLS are commonly used to identify functional skill deficits across self-care, academic, and vocational domains that may benefit from chaining instruction.

Prerequisite skill assessment is equally important. If a learner lacks the motor, sensory, or attending skills required for specific chain steps, those prerequisites must be addressed concurrently or prior to chaining instruction. For example, a chaining program for tooth brushing requires sufficient grip strength and fine motor coordination. If those skills are absent, the task analysis may need to be adapted or supplementary motor training may be needed.

Decision rules should be established before a chaining program begins: What criteria define mastery of each step? Typically, mastery criteria specify the level of independence (e.g., independent or with gestural prompt only), the number of consecutive correct trials, and the number of consecutive sessions meeting criterion. These criteria should be individualized and documented in the behavior intervention plan.

When a learner is not making expected progress, the BCBA should conduct a systematic analysis before changing procedures. Common barriers include: an overly lengthy task analysis, use of insufficiently potent reinforcers, inconsistent implementation across trainers, environmental distractors, or unidentified prerequisite skill deficits. Data review, direct observation, and interview with implementers are all essential before modifying the program.

Chaining programs should also include a plan for assessing and promoting generalization and maintenance. Probes in novel settings and with novel trainers should be built into the program plan from the outset, not added as afterthoughts when clinic-based mastery is achieved.

What This Means for Your Practice

For BCBAs supervising ABAT and RBT staff, chaining offers a dual-purpose framework: it is simultaneously an intervention tool for learners and a training model for staff. When you teach a new staff member to implement a chaining program, you are modeling the same principles — task analysis, prompt fading, systematic data collection, and reinforcement — that the staff member will use with learners.

Developing high-quality task analyses is a clinical skill that improves with deliberate practice. BCBAs should habitually analyze their own task analyses: Are the steps observable and measurable? Is the level of specificity appropriate for this learner? Are prerequisite skills accurately identified? Peer review of task analyses within supervision groups can reveal assumptions and gaps that are invisible to the original author.

Implementation fidelity is the bridge between a well-designed chaining program and actual learner outcomes. Even an excellent task analysis produces poor results if steps are skipped, prompts are delivered inconsistently, or reinforcement is poorly timed. BCBAs must directly observe implementation regularly and provide specific, behavior-specific feedback to staff. Vague feedback such as 'good job' does not build the precise clinical skills that lead to learner progress.

Data-based decision-making is non-negotiable. The pattern of step-level data across sessions tells a clinical story: which steps are mastered, which are plateauing, which show regression under certain conditions. Reviewing these patterns with supervisees turns data collection from a compliance exercise into a clinical reasoning exercise — and that is where professional development actually happens.

Finally, communicate chaining programs clearly to caregivers. Families who understand why each step matters and how to deliver prompts consistently become powerful generalization partners. Simple visual supports, short training sessions, and regular feedback loops with caregivers can dramatically extend the impact of chaining programs beyond the therapy session.