ABA glossary
Reversal design
Also called A-B-A-B design, withdrawal design
Definition
A single-case design in which the intervention is introduced, withdrawn, and reintroduced, so behavior is compared across alternating baseline and treatment phases, most often A-B-A-B. Also called a withdrawal design.
In plain words
A reversal design tests whether a behavior changes when, and only when, the intervention is in place. Baseline (A) shows what behavior looks like without the intervention; the intervention (B) is introduced and behavior changes; the intervention is withdrawn (A) and behavior returns toward baseline; the intervention is reintroduced (B) and the change appears again. Each return to a previous condition tests the prediction made from the earlier phase, which is what makes the design convincing.
Example from a session
The BCBA is evaluating DRA for Jalen's calling out. In baseline, calling out averages 15 per session; during DRA it drops to 4; when the RBTs are told to return to the baseline routine for a few sessions, it climbs back to 13; when DRA resumes, it drops again. The RBTs follow each phase's instructions exactly and enter data on time, because the phase changes only show an effect if each condition is run as written.
For BCBA candidates and supervisors
Reversal designs show prediction (baseline projected forward), verification (behavior returns in the second A phase), and replication (the effect repeats in the second B phase). They are a poor fit when the behavior is unlikely to reverse (learned skills), when withdrawing an effective treatment is unethical (dangerous behavior), or when the change would be irreversible. Some authors reserve "reversal" for designs that apply the treatment to an alternative behavior during the reversal phase and use "withdrawal" for simply removing it. A-B designs lack the verification and replication needed to show experimental control.
Often confused with multiple-baseline design, multielement design
A reversal design removes the intervention to show control. A multiple-baseline design never removes it; it shows control by introducing the intervention at different times across tiers. A multielement design alternates conditions rapidly, often session by session, rather than in long phases.
Exam tip
If the behavior is a skill that will not "unlearn," or the behavior is dangerous, the answer is usually not a reversal design. Look for A-B-A-B and "return to baseline" as the identifying features.
Can you spot reversal design in a scenario?
Practice BCBA D.7 with exam-style questions and instant feedback.