Pragmatic ABAB Design — Pragmatic Reversal Design
Pragmatic ABAB Reversal Design · Also known as: pragmatic reversal design, pragmatic withdrawal design, applied ABAB design, pragmatic single-case reversal
The pragmatic ABAB design is a single-case experimental design that adapts the classic reversal (ABAB) logic to real-world clinical and applied constraints. It alternates between a baseline phase (A) and an intervention phase (B) twice, demonstrating experimental control through repeated phase changes while allowing flexibility — such as abbreviated withdrawals or partial reversals — when full withdrawal of treatment is ethically or practically impossible.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use the pragmatic ABAB design when you need to establish a causal link between an intervention and a behavioral or clinical outcome for an individual participant — particularly in applied, clinical, or educational settings where randomized group trials are infeasible. It is appropriate when (a) the target behavior is expected to reverse when intervention is removed or reduced, (b) ethical or logistical constraints prevent full withdrawal, and (c) continuous repeated measurement is possible. Do not use it when the behavior is unlikely to reverse (e.g., acquired skills that are self-sustaining after learning), when repeated measurement is disruptive to participants, or when the goal is population-level generalization rather than individual-level causal demonstration.
Strengths & limitations
- Provides strong within-participant causal evidence through two replications of the intervention effect — the gold standard for single-case experimental control.
- Flexible enough for real clinical and school-based settings where ethical or practical constraints prevent ideal full-withdrawal ABAB protocols.
- Requires no comparison group; each participant serves as their own control, making it viable for rare conditions or small populations.
- Continuous data collection produces a rich time-series record that captures the dynamics of change in detail not available from pre/post designs.
- Transparent analysis via visual graphs allows clinicians and practitioners to engage directly with the evidence.
- Requires that the target behavior be reversible; skills or behaviors that persist after intervention removal cannot be studied with this design without sacrificing experimental control.
- The pragmatic withdrawal weakens causal inference relative to the ideal full-reversal ABAB — this tradeoff must be acknowledged explicitly.
- Findings demonstrate causal effects for the individual studied; replication across multiple participants or settings is needed before broader generalizations are warranted.
- Intensive and prolonged data collection demands resources and participant cooperation that may not be sustainable in every applied context.
Frequently asked
How is the pragmatic ABAB design different from the classic ABAB design?
The classic ABAB design fully withdraws the intervention during the second A phase. The pragmatic variant reduces or modifies the intervention — rather than eliminating it entirely — when full withdrawal is unethical, harmful, or refused by participants or caregivers. This makes the design usable in real-world settings at the cost of slightly weaker causal inference, which must be disclosed.
How many data points are needed per phase?
The conventional minimum is three data points per phase, but five or more per phase is preferred to establish a stable trend. Phases with highly variable data may need to be extended further before a phase change is warranted. There is no universal rule; the criterion is stability rather than a fixed count.
What effect-size statistics should I report alongside the graph?
Tau-U (a rank-based statistic that accounts for trend) and the Non-overlap of All Pairs (NAP) index are the most widely recommended non-parametric effect-size measures for single-case data. Both can be computed with freely available online calculators. Some journals also accept the Percentage of Non-overlapping Data (PND), though it has known limitations.
Can I generalise findings from one participant to a population?
Not directly. A single-participant ABAB study establishes a causal effect for that individual under those conditions. Generalization requires systematic replication — repeating the design with different participants, settings, or therapists across multiple studies. A series of replications, or a meta-analysis of single-case studies, supports broader generalization.
What if the behavior does not return to baseline during the withdrawal phase?
A non-reversal during withdrawal undermines the experimental logic. Possible explanations include: (a) the behavior has become self-sustaining and is no longer controlled by the intervention, (b) the withdrawal was insufficient (pragmatic dose too high), or (c) a confounding variable maintains the behavior. The researcher should discuss which explanation is most plausible and consider switching to a multiple-baseline design if reversal is not achievable.
Sources
- Kazdin, A. E. (2011). Single-Case Research Designs: Methods for Clinical and Applied Settings (2nd ed.). Oxford University Press. ISBN: 978-0195341881
- Kratochwill, T. R., & Levin, J. R. (Eds.). (2010). Single-Case Intervention Research: Methodological and Statistical Advances. American Psychological Association. ISBN: 978-1433810251
How to cite this page
ScholarGate. (2026, June 3). Pragmatic ABAB Reversal Design. ScholarGate. https://scholargate.app/en/experimental-design/pragmatic-abab-design
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Alternating Treatments DesignDisability Studies↔ compare
- Changing Criterion DesignDisability Studies↔ compare
- Interrupted Time SeriesCausal inference↔ compare
- Multiple Baseline DesignExperimental design↔ compare
- Single-Case Experimental DesignDisability Studies↔ compare