Beyond Simple Treatment Evaluation
In applied behavior analysis, demonstrating that an intervention works is only the first step. Once a treatment package has been shown to be effective, researchers and clinicians often want to know more: Which specific components of the package are responsible for the effect? What dosage or intensity produces the best results? Component analysis and parametric analysis are two research approaches that answer these different but related questions.
The BCBA exam tests your ability to distinguish between these two types of analysis, understand when each is appropriate, and identify them in research scenarios.
What Is a Component Analysis?
A component analysis (also called a dismantling study) evaluates the individual contributions of different elements within a multi-component treatment package. The goal is to determine which components are necessary for the treatment to be effective and which may be unnecessary.
How It Works
In a component analysis, the researcher first establishes that the full treatment package is effective. Then, individual components are systematically added or removed to determine their contribution. There are two primary approaches:
Dropout approach: Start with the full package and remove components one at a time. If behavior deteriorates when a component is removed, that component is likely contributing to the treatment's effectiveness. If behavior remains the same, the component may not be necessary.
Add-in approach: Start with a single component and add others one at a time. Compare the effectiveness at each stage to determine whether each additional component improves outcomes.
Example
A behavior analyst implements a treatment package for reducing aggression that includes functional communication training (FCT), extinction of aggression, and a visual schedule. The package is effective, but the team wants to know whether all three components are necessary.
Using the dropout approach, the analyst might implement the full package first, then remove the visual schedule while keeping FCT and extinction in place. If aggression remains low, the visual schedule may not be a necessary component. The analyst might then remove extinction to see whether FCT alone is sufficient.
Alternatively, using the add-in approach, the analyst might start with FCT alone, then add extinction, then add the visual schedule, evaluating the effect at each step.
When to Use a Component Analysis
- When a treatment package is effective but resource-intensive, and you want to determine if a simpler version would work equally well.
- When you want to understand the mechanism of change (which specific elements are driving behavior change).
- When training implementers on the full package is difficult and you want to identify the essential components.
- When you want to develop a more efficient treatment protocol.
What Is a Parametric Analysis?
A parametric analysis examines different values or levels of a single variable to determine the optimal level. Rather than asking "which components are effective?" it asks "how much (or what intensity, duration, or schedule) of this variable produces the best results?"
How It Works
The researcher systematically varies one dimension of the intervention while holding all other aspects constant. The different values are compared to determine which produces the best outcomes. This can be done using an alternating treatments design, a changing criterion design, or a series of phase changes.
Example
A behavior analyst knows that token reinforcement is effective for increasing homework completion. A parametric analysis might compare different reinforcement schedules: earning one token per completed assignment versus one token per three completed assignments versus one token per five completed assignments. The analysis reveals which schedule produces the best balance of effectiveness and efficiency.
Another example: A clinician uses response cost to reduce inappropriate language. A parametric analysis might compare losing 1 token per instance versus losing 3 tokens per instance versus losing 5 tokens per instance to determine the optimal magnitude of the consequence.
When to Use a Parametric Analysis
- When you know a procedure is effective but want to optimize its parameters.
- When you want to determine the minimal effective dose of an intervention.
- When comparing different reinforcement schedules, magnitudes, or delays.
- When you want to identify the point of diminishing returns for a particular variable.
Key Differences
The fundamental distinction is straightforward: a component analysis examines which parts of a treatment package are responsible for behavior change, while a parametric analysis examines how much of a single variable is optimal.
Component analysis involves adding or removing qualitatively different elements. Parametric analysis involves adjusting the quantitative value of a single element. Both are valuable for refining interventions, but they answer different questions.
How These Differ from Treatment Comparisons
A treatment comparison (often conducted using an alternating treatments design) compares two or more entirely different interventions to see which is more effective. This differs from both component analysis and parametric analysis:
- Treatment comparison: "Is Treatment A or Treatment B more effective?"
- Component analysis: "Within Treatment A, which components are necessary?"
- Parametric analysis: "What level or dose of this specific variable in Treatment A is optimal?"
BCBA Exam Tips
When an exam question describes a researcher removing or adding elements of a treatment package, the answer is likely component analysis. When the question describes varying the amount, intensity, frequency, or magnitude of a single intervention variable, the answer is parametric analysis. When two completely different interventions are compared, it is a treatment comparison.
Pay close attention to the wording of exam questions. Phrases like "which element," "necessary component," or "essential ingredient" point toward component analysis. Phrases like "optimal schedule," "how many," "what intensity," or "what magnitude" point toward parametric analysis.
Both analyses build on existing evidence of effectiveness. They are not first-line evaluations of a new treatment; rather, they are refinement strategies used after an intervention has already been shown to work.