Experimental Design
Logic of single-case designs
Behavior analysis establishes experimental control by demonstrating that the independent variable, not extraneous factors, produced change. Essential features include repeated measurement, baseline stability, and replication of the effect. Designs include reversal (ABA/ABAB), multiple baseline across behaviors, settings, or participants, alternating treatments, changing criterion, and withdrawal. Each has conditions under which it is preferred. Reversal requires that the behavior be reversible and that withdrawal be ethically defensible-a frequent exam trap is recommending reversal for a dangerous behavior.
Threats to internal validity
Know maturation, history, testing, instrumentation, regression to the mean, and sequence effects. Alternating treatments designs control for sequence effects through rapid alternation and counterbalancing. Multiple baseline designs do not require withdrawal but do require that baselines be independent and that the untreated tiers remain stable until intervention begins.
External and social validity
Generality concerns whether effects persist, transfer, and spread. Social validity concerns acceptability of goals, procedures, and outcomes to consumers.
Rule of thumb
Ask: what would the data look like if the intervention had no effect? If your design cannot rule that out, it lacks experimental control.
Worked example
A clinician wants to evaluate a token system for aggression. Reversal would require withdrawing an effective treatment for dangerous behavior, which is not defensible. A multiple baseline across settings or an alternating treatments design with a comparison condition preserves experimental control without removing treatment.