Experimental Design

7% of the BCBA exam · blueprint Test Content Outline 6th ed. (since 2025)

Short answer

Domain D is 13 of 175 scored questions (7%). It covers internal and external validity and their common threats, the baseline logic of single-case designs (prediction, verification, replication), and choosing between reversal, multiple-baseline, alternating-treatments/multielement and changing-criterion designs, including the ethical constraints that rule out withdrawing an effective treatment for dangerous behavior.

Why it matters on this exam

Design questions are scenarios: given this behavior and this setting, which design demonstrates experimental control without harming the client?

Key concepts

Baseline logic: prediction (stable baseline), verification (return to baseline reproduces the original pattern), replication (re-introducing treatment reproduces the effect).
Reversal (ABAB): strongest demonstration, but wrong for irreversible behavior (acquired skills) and unethical where withdrawal exposes the client to harm.
Multiple baseline: staggered introduction across behaviors, settings or participants; requires functionally independent tiers.
Alternating treatments/multielement: rapid alternation with discriminative signals; good for comparing interventions; vulnerable to multiple-treatment interference.
Changing criterion: stepwise criterion changes; control shown when behavior tracks each new criterion.
Threats: history, maturation, testing, instrumentation, attrition; for external validity, the limits of generality across people, settings and behaviors.

Common mistakes

Choosing a reversal design for a newly taught skill. You cannot unteach it.
Running a multiple baseline across behaviors that belong to the same response class, which destroys the staggered logic.

Worked example

Severe self-injury drops sharply under a function-based treatment. A reversal would re-expose the client to injury, so a multiple baseline across settings demonstrates control without withdrawing treatment.

Original practice questions

Experimental DesignDifficulty 2/5

A behavior analyst has taught a client to tie their shoes and wants to demonstrate experimental control. Why is a reversal design inappropriate?

AReversal designs cannot be used with children because withdrawal is always unethical
BThe behavior must be dangerous before a reversal design can be justified
CReversal designs require at least three participants to replicate the effect
DThe skill is irreversible, so withdrawing teaching will not return responding to baseline

Show answer and rationale

Correct: D. Reversal logic depends on behavior returning to baseline levels when treatment is withdrawn. Acquired skills do not reverse, so verification is impossible. A multiple baseline across behaviors or participants is the appropriate alternative.

C — Reversal designs are within-subject and need only one participant.
A — Age is not the constraint.
B — Danger is a reason to avoid reversal, not a requirement for it.

Experimental DesignDifficulty 3/5

Which condition is required for a multiple baseline across behaviors to demonstrate experimental control?

AThe behaviors must be functionally independent
BThe behaviors must belong to the same response class
CAll tiers must begin intervention simultaneously
DBaselines must be equal in length

Show answer and rationale

Correct: A. If behaviors are functionally related, intervening on one changes the others and the untreated baselines no longer serve as controls. Tiers must start intervention in a staggered fashion, and baseline lengths are intentionally unequal.

B — Same response class means covariation, which destroys the design logic.
C — Simultaneous introduction removes the staggered comparison.
D — Unequal baseline lengths are what create the staggered demonstration.

All 2 Experimental Design questions →

2 flashcards for Experimental Design

Practice this topic

Blueprint-mapped questions with full rationales for every choice.

Start practice

Primary references

BCBA Test Content Outline (6th ed.), in effect since 2025 — Behavior Analyst Certification Board (BACB) · primary · retrieved 2026-09-22, verified 2026-09-22

Last reviewed: 2026-10-18Next review by: 2026-12-17Reviewed by: Openverse editorial team, Authoring and fact verificationHow we verify facts