Component analysis

Open the treatment package to study its parts

A treatment package may contain instructions, prompts, reinforcement, feedback, environmental changes, caregiver actions, technology, and other elements. Package-level evidence asks whether the combination produces change. The component question asks which elements contribute to that result and whether a simpler combination can preserve it.

The BACB BCBA Test Content Outline, 6th edition covers rationales for comparative, component, and parametric analyses alongside single-case design and data interpretation. It is examination content rather than a protocol or authority to practice.

A review of 30 single-case component analyses defined the method as a systematic assessment of two or more independent variables within a treatment package. Most studies in that review identified a necessary component, while fewer independently established sufficiency or evaluated every combination. A claim should match the combinations actually tested.

Necessary, sufficient, active, and additive differ

• A necessary component is one whose removal weakens the package under the tested conditions.
• A sufficient component produces the target effect alone under the tested conditions.
• An active component contributes to change, though it may need another component.
• An additive component improves the outcome when joined to another effective element.

These labels describe evidence from a particular analysis. A component can work only in combination, matter during acquisition and less during maintenance, or help one person in one setting. Similar outcomes after removal can reflect a lingering effect, ceiling performance, weak measurement, or inadequate exposure rather than proof that the component has no value.

Add-in and dropout analyses answer different questions

An add-in analysis begins with one component or a small combination and adds elements systematically. Presenting components alone can test sufficiency. Adding one element to a stable combination can test its incremental contribution. Order matters because learning, practice, fatigue, or a ceiling can change what later additions appear to do.

A dropout analysis begins with the full package and removes components systematically. If performance deteriorates after one element is removed and recovers when it returns, the pattern supports necessity. Previous exposure to the full package may create carryover, so stable performance after removal may be difficult to interpret.

The component-analysis review describes reversal, alternating-treatments, multiple-baseline, and combined arrangements. A broader single-case design guide also notes that durable learning limits dropout logic because acquired behavior may persist after a teaching component is withdrawn. Select the design according to reversibility, expected speed of effect, risk, burden, carryover, and the exact question.

Component interactions need explicit testing

Suppose components A and B produce little change alone, while A plus B produces a clear effect. Calling either component inactive would miss their interaction. The reverse can occur when one component adds effort or interferes with another.

A complete analysis of a large package can require many combinations. Prioritize contrasts that answer a real decision: whether a burdensome element is needed, whether a simpler package retains benefit, whether one element adds measurable value, or whether implementation can be made more acceptable. The evidence-based practice discussion by Slocum and colleagues cautions that outcomes from a validated package may reflect interactions that studies of isolated components do not reveal.

Counterbalancing, randomizing eligible condition order, separate stimulus sets, or planned washout may reduce some sequence effects. Each choice introduces assumptions that should be visible. When carryover cannot dissipate safely, a rapid component comparison may be unsuitable.

Predefine the package, contrast, and decision rule

Before data collection, specify:

• every component and the full package
• the component combinations that will be tested
• the outcome, opportunity, observation window, prompts, and exclusions
• condition order, exposure, phase-change rule, and carryover control
• the evidence pattern that supports necessity, sufficiency, or added value
• implementation-fidelity and observer-agreement procedures
• adverse-effect, assent-withdrawal, and stop rules

Graph each condition in time order. Report raw observations and component combinations instead of relying only on an average. Measure whether every assigned element was delivered correctly. A weak component result is uninterpretable when that component was often missing or implemented differently.

Track preference, effort, feasibility, unwanted effects, generalization, maintenance, and ordinary supports separately from the primary outcome. Greater efficiency matters only when the streamlined package remains effective, acceptable, accessible, and practical.

Clinical safeguards govern every comparison

Preserve augmentative and alternative communication, food, water, bathroom access, mobility, prescribed care, pain care, rest, emergency help, and effective safety protections across conditions. Removing an essential support to discover whether it is “active” creates an unsafe test. Avoid withdrawing a component when loss of benefit could create unacceptable risk.

The current BACB Ethics Code applies to BCBA and BCaBA certificants and people who completed an application. It addresses competence, client involvement, informed consent and assent when applicable, assessment-based intervention, risk, data, and continual evaluation. A meaningful withdrawal signal, new health concern, adverse effect, or change in risk can end the planned comparison.

A fictional component-analysis example

Rowan is a fictional adult who chooses supports for a five-step volunteer sorting routine. Three acceptable variants use the same pictorial checklist: checklist alone, checklist plus a self-advancing highlight, and the full package with highlight plus an optional vibration cue. Eighteen matched opportunities occur in six randomized blocks, with each variant presented once per block. All essential communication and safety supports remain available.

Rowan completes 18 of 30 steps independently with checklist alone, 25 of 30 with checklist plus highlight, and 24 of 30 with the full package. Staff deliver the assigned components correctly in 18 of 18 opportunities. Rowan rates checklist plus highlight as the most comfortable variant.

The separated counts support a hypothesis that the highlight adds value to the checklist under these conditions. The vibration cue shows no additional measured benefit in this small comparison. That finding does not prove the cue is universally inactive, because task sets, sequence, interaction, and sampling remain possible explanations. The team selects checklist plus highlight provisionally, then monitors everyday use, preference, and maintenance.

Common points of confusion

A comparative analysis contrasts distinct interventions. Component analysis compares elements or combinations from one package. A parametric analysis changes the value of one dimension, such as duration or magnitude. A functional analysis manipulates antecedents and consequences to test hypotheses about behavioral function. Analysts can use an alternating-treatments or reversal arrangement to study components; the design arrangement and analytic question remain separate concepts.