Why Single-Subject Designs Matter
Single-subject experimental designs (also called single-case designs) are the primary research methodology in applied behavior analysis. Unlike group designs that compare average outcomes between groups of participants, single-subject designs demonstrate experimental control by using each participant as their own control. Understanding these designs is essential for both the BCBA exam and evidence-based practice.
Baseline Logic
All single-subject designs rely on baseline logic, which involves three elements: prediction, verification, and replication.
- Prediction - Baseline data establish a pattern that allows the researcher to predict what behavior would look like if no intervention were applied
- Verification - The researcher verifies that the behavior change was due to the intervention and not other variables
- Replication - The effect is replicated to strengthen the demonstration of experimental control
Reversal (ABAB) Designs
The reversal design is the most straightforward single-subject design. It alternates between baseline (A) and treatment (B) phases to demonstrate that behavior changes systematically with the introduction and withdrawal of the independent variable.
Phases:
- A1 (baseline) - Measure behavior without intervention
- B1 (treatment) - Introduce the intervention and measure behavior
- A2 (return to baseline) - Withdraw the intervention and measure behavior
- B2 (treatment) - Reintroduce the intervention and measure behavior
Strengths: Provides a clear demonstration of experimental control through replication. If behavior improves during treatment, returns to baseline levels during withdrawal, and improves again during reintroduction, the evidence for a functional relationship is strong.
Limitations: Cannot be used when the behavior change is irreversible (such as skills that once learned cannot be unlearned). Ethical concerns arise when withdrawing effective treatment for dangerous behaviors. Not suitable when there are strong carryover effects.
Multiple Baseline Designs
Multiple baseline designs stagger the introduction of the intervention across different tiers (behaviors, settings, or participants) to demonstrate that behavior changes only when the intervention is applied to each tier.
Multiple Baseline Across Behaviors
The intervention is applied to different behaviors of the same individual at staggered time points. All behaviors are measured simultaneously during baseline, and the intervention is introduced to one behavior at a time.
Multiple Baseline Across Settings
The intervention is applied to the same behavior in different settings at staggered time points.
Multiple Baseline Across Participants
The intervention is applied to the same behavior in different individuals at staggered time points.
Strengths: Does not require withdrawal of treatment. Suitable for behaviors that are irreversible. Can demonstrate generality of the intervention across different contexts.
Limitations: If behaviors covary (change together without independent intervention), the demonstration of experimental control is compromised. Requires at least three tiers. Takes more time than reversal designs because of the staggered introduction.
Changing Criterion Design
In a changing criterion design, the criterion for reinforcement is systematically changed in a stepwise fashion. Each new criterion level serves as a mini-baseline for the next phase. The design demonstrates experimental control when the behavior closely tracks each criterion change.
Example: A student's reading fluency goal starts at 50 words per minute, then increases to 60, then 70, then 80. If the student's reading consistently matches each criterion level, the evidence supports a functional relationship between the intervention and reading fluency.
Strengths: Ideal for behaviors that should change gradually (such as increasing academic fluency or decreasing substance use). Does not require withdrawal of treatment. Demonstrates that the behavior is under the control of the programmed contingencies.
Limitations: Not suitable for behaviors that should change immediately. Provides weaker evidence of experimental control if the behavior changes are large and rapid.
Alternating Treatments Design (ATD)
The alternating treatments design rapidly alternates between two or more conditions (typically different interventions) to compare their effects on behavior. Conditions are counterbalanced across sessions to control for sequence effects.
Example: Comparing the effectiveness of two teaching methods by alternating between them across sessions and measuring skill acquisition under each condition.
Strengths: Allows direct comparison of multiple treatments. Does not require withdrawal of an effective treatment. Can identify the most effective intervention relatively quickly. No baseline phase is required (though one is often included).
Limitations: Potential for multiple treatment interference (carryover effects between conditions). Requires that conditions be discriminable to the participant. Not appropriate for comparisons involving more than three or four conditions.
Multielement Design
The multielement design is similar to the alternating treatments design but is most commonly associated with functional analysis. Different conditions (such as attention, escape, tangible, and control) are presented in a rapidly alternating fashion with distinct discriminative stimuli for each condition.
Visual Analysis of Graphed Data
Single-subject designs rely on visual analysis rather than inferential statistics. Key features to analyze include:
- Level - The average value of data within a phase
- Trend - The direction and slope of data within a phase
- Variability - The degree of fluctuation of data around the trend
- Immediacy of effect - How quickly behavior changes when conditions change
- Overlap - The proportion of data in adjacent phases that share the same range
- Consistency - Whether the pattern of change replicates across phases or tiers