πŸ”¬ BCBA Research Methods β€” TCO 6th Ed.

Single-Subject
Reversal Β· Multiple Baseline Β· ATD Β· Changing Criterion

Master the four core single-subject experimental designs with interactive quizzes, a side-by-side comparison, a design selector tool, and visual analysis memory hooks.

🎯 Take the Practice Quiz
What Are Single-Subject Designs?

Single-subject experimental designs (SSEDs) allow behavior analysts to demonstrate a functional relation between an independent variable (IV) and a dependent variable (DV) using the individual as their own control.

πŸ”¬
Functional relation is demonstrated when the IV predictably produces changes in the DV that would not have occurred otherwise. In SSEDs, replication within the same participant (or across tiers/participants) provides experimental control.
πŸ”„
Reversal / ABAB

Withdraw & Reintroduce

Strongest single design for demonstrating experimental control by reversing to baseline conditions.

Phases: A β†’ B β†’ A β†’ B Control via: Withdrawal & reinstatement of IV Limitation: Not for irreversible behaviors
πŸ“Š
Multiple Baseline

Stagger & Replicate

Most widely used in ABA. Introduces intervention at different times across behaviors, settings, or participants.

Types: Across behaviors / settings / participants Control via: Staggered replication Limitation: Requires β‰₯3 tiers; no withdrawal needed
⚑
Alternating Treatments

Rapidly Compare Conditions

Rapidly alternates between two or more conditions to compare their relative effects on behavior.

Also called: Multi-element design Control via: Rapid alternation & counterbalancing Limitation: Multiple treatment interference
πŸ“ˆ
Changing Criterion

Step-by-Step Shaping

Gradually shifts the criterion for reinforcement or punishment to shape behavior toward a terminal goal.

Best for: Shaping; gradual behavior change Control via: Behavior tracks each criterion shift Limitation: Requires predictable, gradual change

The Core Principle of All SSEDs

Experimental control is demonstrated through replication. A functional relation is established when the DV changes predictably each time the IV is introduced, withdrawn, or shifted β€” and does not change when the IV is absent.

πŸ’‘
A phase = Baseline. B phase = Intervention. Every SSED uses this notation. A "B" phase always represents a specific condition β€” the first intervention is B, the second is B' or C, and so on.
Take the Quiz β†’
How Each Design Works

The logic, phases, replication strategy, and key limitations of each design.

πŸ”„ Reversal Design (ABAB) Strongest Control

Phase Sequence
Baseline (A) β†’ Intervention (B) β†’ Withdrawal (A) β†’ Reintroduction (B)
Replication Method
Within-participant: behavior reverses toward baseline when IV is withdrawn, recovers when reinstated
Experimental Control
Demonstrated when DV changes in the predicted direction each time the IV is introduced or withdrawn
Minimum Phases
4 phases (ABAB); each phase requires β‰₯3 stable data points
A (Baseline)
β†’
B (Intervention)
β†’
A (Withdrawal)
β†’
B (Reintroduction)
⚠️
Key limitation: Not appropriate when (1) the behavior is dangerous or self-injurious, (2) behavior change is irreversible (skills acquired cannot be "unlearned"), or (3) withdrawing treatment would be unethical. In those cases, use Multiple Baseline.
πŸ’‘
Variations: BAB (start with intervention when baseline is unethical), ABAB' (second intervention modified), ABABAB (extended replication for stronger control), ABCBC (comparing two interventions within a reversal).
What to look for on the exam "The teacher implemented a token economy, then removed it for two weeks, then reinstated it" = ABAB reversal design. If the behavior improves during both B phases and worsens during both A phases, a functional relation is demonstrated.

πŸ“Š Multiple Baseline Design Most Common in ABA

Three Types
Across behaviors Β· Across settings Β· Across participants
Replication Method
Staggered introduction of IV across β‰₯3 tiers; each tier acts as a concurrent control for others
Experimental Control
Demonstrated when only the tier receiving the IV changes; untreated tiers remain stable
Withdrawal Required?
❌ No β€” this is its main advantage over the reversal design
Behavior 1
Baseline
Intervention β†’
Behavior 2
Baseline
Intervention β†’
Behavior 3
Baseline
Intervention β†’
πŸ“
Across behaviors: Same participant, same setting, multiple behaviors targeted sequentially. Across settings: Same participant, same behavior, multiple settings. Across participants: Same behavior, same intervention, multiple participants β€” provides the strongest external validity.
⚠️
Key threat: In multiple baseline across behaviors, if behavior 2 changes when only behavior 1 is treated (behavioral covariation), experimental control is weakened. This is less of a concern in across-participants designs.
What to look for on the exam "A BCBA taught on-task behavior first in math class, then in reading, then in science, introducing the intervention at different times" = multiple baseline across settings.

⚑ Alternating Treatments Design (ATD) Compare Treatments

Also Called
Multi-element design; simultaneous treatment design (older term)
Structure
Two or more conditions rapidly alternated (often session-by-session or within sessions)
Counterbalancing
Conditions are counterbalanced across time, therapists, and settings to control for sequence effects
Best Use
Comparing treatment efficiency; identifying which treatment produces faster/greater change
Baseline
β†’
Intervention A
Intervention B
A
B
A
B
⚠️
Multiple treatment interference: The main limitation of the ATD. Exposure to one treatment may affect responding during another. This limits the generalizability of conclusions about each treatment in isolation. Not ideal when treatments cannot be kept distinct.
πŸ’‘
Advantages: No withdrawal needed. Can compare treatments quickly. Works well with behaviors that fluctuate or don't have stable baselines. Can include a no-treatment control condition.
What to look for on the exam "The BCBA compared video modeling and social stories by alternating them each session, counterbalanced by time of day" = alternating treatments design.

πŸ“ˆ Changing Criterion Design Shaping Behavior

Best For
Gradually increasing or decreasing a behavior toward a terminal goal (shaping, fluency, habit reduction)
How Control is Shown
Behavior tracks each criterion change β€” rises (or falls) to meet the new criterion, then holds steady until the next shift
Strengthening Control
Vary step sizes (some larger, some smaller); reverse criteria occasionally; use sufficiently large steps
Examples
Increasing academic fluency, reducing cigarettes/day, increasing exercise duration, reducing bite size
Baseline
β†’
Criterion 1
β†’
Criterion 2
β†’
Criterion 3
β†’
Criterion 4
πŸ“
Experimental control: If the behavior changes only when the criterion changes (and not between criterion shifts), a functional relation is supported. Control is strongest when step sizes are varied and behavior tightly tracks each new criterion.
⚠️
Limitation: If criterion steps are too small, spontaneous behavior change could account for results. Not ideal for behaviors that need to be eliminated entirely (use reversal or multiple baseline instead).
What to look for on the exam "The student earned a break after completing 5 math problems; the criterion was raised to 10, then 15, then 20, with performance tracking each increase" = changing criterion design.
Visual Analysis of Graphed Data

The BACB requires visual analysis (not statistical analysis) to determine if behavior change is meaningful. Six dimensions are evaluated within and across phases.

πŸ“

Level

The mean or median of data points within a phase. Compare across phases to see magnitude of change.

πŸ“‰

Trend

The direction data are moving (accelerating, decelerating, zero-celerating). Is it going up, down, or flat?

〰️

Variability

The spread of data around the trend line. High variability = less stable data, harder to draw conclusions.

⏱️

Immediacy

How quickly does the behavior change after the phase change? Faster = stronger evidence of functional relation.

πŸ”

Overlap

The proportion of data points in the intervention phase that overlap with data in the baseline phase. Less overlap = stronger effect.

πŸ“

Consistency

Whether patterns repeat across similar phases. Consistent change across replications strengthens the functional relation conclusion.

πŸ’‘
Stability criterion: A phase should have at least 3 data points (often more) and demonstrate a stable trend before the next phase begins. Some guidelines suggest no more than 20% variability from the mean for a phase to be considered "stable."
Side-by-Side Comparison

Filter by design to highlight a column, or view all four together.

Criterion πŸ”„ Reversal (ABAB) πŸ“Š Multiple Baseline ⚑ ATD πŸ“ˆ Changing Criterion
Withdrawal Required? βœ… Yes β€” core of the design ❌ No ❌ No ❌ No
Replication Method Within-participant (behavior reverses & recovers) Across tiers (staggered introduction) Rapid alternation of conditions Behavior tracks each criterion shift
Experimental Control Strongest β€” DV changes with each IV intro/withdrawal Strong β€” only treated tier changes Moderate β€” convergent data across alternations Moderate β€” behavior matches each criterion step
Primary Use Evaluating a single treatment's effect Evaluating a treatment without withdrawal Comparing 2+ treatments Shaping/gradual behavior change
Key Limitation Unethical if behavior is dangerous or irreversible Behavioral covariation across tiers (across behaviors) Multiple treatment interference Weak control if steps are too small
Minimum Tiers/Phases 4 phases (A-B-A-B) 3 tiers minimum 2 conditions minimum Baseline + β‰₯3 criterion phases
Ethical When Baseline is Dangerous? ❌ No β€” can't withdraw if behavior is harmful βœ… Yes (use across participants variant) ⚠️ Depends on design βœ… Generally yes
Good for Skills Acquisition? ⚠️ Only if skill can be "unlearned" βœ… Yes βœ… Yes βœ… Yes (fluency building)
External Validity Within-participant replication; requires cross-study replication for generalization Highest (across participants) β€” built-in replication across individuals Limited by multiple treatment interference Within-participant; limited generalization
Counterbalancing? N/A N/A βœ… Required β€” controls sequence effects N/A
Real-World Vignettes

Apply your knowledge to clinical and research scenarios.

πŸ”„ ReversalToken Economy for On-Task Behavior
"A BCBA measured a student's on-task behavior during baseline (A). She introduced a token economy (B), and on-task behavior improved to 85%. She then removed the tokens for 2 weeks β€” on-task behavior dropped to 30% (A). When tokens were reinstated (B), behavior returned to 85%."
DesignReversal (ABAB) β€” IV withdrawn then reinstated
Functional Relation?Yes β€” behavior changed predictably with each IV manipulation
Why not Multiple Baseline?Only one behavior/setting; withdrawal was ethical and possible
Exam TrapThe second A phase (return to baseline) is essential β€” without it, you can't rule out history or maturation
πŸ“Š Multiple BaselineTeaching Social Skills Across Settings
"A BCBA used social skills training to increase appropriate peer greetings. She introduced the intervention first in the classroom (week 3), then on the playground (week 6), then in the cafeteria (week 9). Behavior improved in each setting only after intervention began there; untreated settings remained at baseline levels."
DesignMultiple baseline across settings β€” same behavior, same participant, 3 settings
ControlEach setting serves as a control for the others while untreated; change only when IV introduced
Why not Reversal?Social skills acquisition is not easily reversed β€” the skill is retained once learned
Exam KeyStaggered introduction of IV + β‰₯3 tiers + no withdrawal = Multiple Baseline
⚑ ATDComparing Video Modeling vs. Social Stories
"A researcher compared video modeling (VM) and social stories (SS) on a child's initiating play with peers. Sessions using VM and SS were alternated each day (counterbalanced by morning/afternoon). After 20 sessions, VM consistently produced higher initiation rates than SS."
DesignAlternating Treatments Design (ATD) β€” two conditions rapidly alternated
CounterbalancingAlternated morning/afternoon to control for time-of-day effects
ConclusionVM is the more efficient treatment; functional relation with each treatment individually is harder to isolate
Exam KeyRapid alternation + counterbalancing + comparing 2 conditions = ATD
πŸ“ˆ Changing CriterionIncreasing Reading Fluency
"A student read 20 words per minute (wpm) at baseline. The teacher set a criterion of 30 wpm for reinforcement; the student reached 30 wpm within a week. The criterion was raised to 40 wpm, then 55 wpm, then 70 wpm. The student's performance matched each new criterion within days of it being set."
DesignChanging Criterion β€” behavior tracks each criterion shift; terminal goal is 70 wpm
ControlEach time the criterion changes and behavior changes to match, functional relation is supported
Why not ABAB?Fluency gains are not reversible; you can't "unlearn" how to read faster
Exam KeyGradual step-by-step criterion shifts with behavior tracking each step = Changing Criterion
⚠️
Common exam traps: (1) A BAB design (starts with intervention) is still a reversal design β€” just used when you can't collect baseline first. (2) Multiple baseline across participants is NOT the same as a group design β€” each participant is still their own unit of analysis. (3) In an ATD, data are NOT averaged across conditions β€” you compare the separate data paths visually.
Practice Quiz

10 questions with per-design performance breakdown. Identify the correct experimental design in each scenario.

Question 1 of 10

Reversal (ABAB)
β€”
Multiple Baseline
β€”
ATD
β€”
Changing Criterion
β€”
Design Selector Tool

Answer 3 questions to identify the most appropriate single-subject design for your scenario.

What is the primary research goal?
Consider what question you are trying to answer with the design.
Can the target behavior safely reverse to baseline levels if the intervention is withdrawn?
Think: is the behavior dangerous? Is the skill irreversible once learned?
Can you replicate the intervention across at least 3 tiers (behaviors, settings, or participants)?
Multiple baseline requires introducing the same IV in a staggered fashion across β‰₯3 independent tiers.
Can the two conditions be kept distinct and alternated rapidly without carryover effects?
If conditions bleed into each other (e.g., a medication takes days to wash out), multiple treatment interference becomes a serious concern.
Will the behavior change gradually, step by step, toward a terminal goal?
The changing criterion design requires that each criterion step be sufficiently large to distinguish from spontaneous change.
Memory Hooks & Mnemonics

Click each card to flip it and reveal the mnemonic.

πŸ‘† Tap a card to flip

πŸ”„
Reversal
When is an ABAB design appropriate?
"In and Out β€” prove it works by taking it away"
If behavior improves with B and worsens when you remove it (A again), you've shown the IV causes the change β€” that's experimental control.
🚫
Reversal Limit
When can you NOT use a reversal design?
"Dangerous + Irreversible = No Withdrawal"
SIB, aggression, or acquired skills cannot be safely reversed. Use Multiple Baseline instead. On the exam: if the behavior is harmful or a skill, reversal is wrong.
πŸ“Š
Multiple Baseline
What makes the Multiple Baseline work?
"Stagger and Wait β€” only the treated tier should change"
You introduce the IV at different times across β‰₯3 tiers. If untreated tiers stay flat while treated ones improve, history and maturation are ruled out.
🧲
MB Covariation
What is the main threat to Multiple Baseline across behaviors?
"Behavioral Covariation = untreated behavior also changes"
If treating Behavior 1 accidentally improves Behavior 2 (before you treat B2), you can't demonstrate that YOUR intervention caused B2's change. Strongest control = across participants.
⚑
ATD
What is the ATD best for and its key risk?
"Pick a Winner β€” but watch for bleed-over"
The ATD rapidly alternates treatments to find which works best. The risk: multiple treatment interference β€” exposure to one treatment may contaminate responding during the other.
πŸ“ˆ
Changing Criterion
How does the Changing Criterion demonstrate experimental control?
"Behavior follows the criterion like a shadow"
Each time the criterion steps up (or down), behavior tracks it. If behavior only changes when the criterion changes β€” not between shifts β€” the IV is controlling the behavior.
πŸ“‰
Visual Analysis
What 6 things do you analyze in graphs?
"Level Β· Trend Β· Variability Β· Immediacy Β· Overlap Β· Consistency"
BACB uses visual (not statistical) analysis. More change in level, clearer trend, less variability, faster immediacy, less overlap, and consistent patterns = stronger functional relation.
πŸ”¬
Functional Relation
What does it take to demonstrate a functional relation?
"Predict, Verify, Replicate"
You predict what will happen when the IV is introduced, verify it does, then replicate the effect. Three demonstrations of the effect in the same experiment establish a functional relation.

Design Selection Quick Reference

If the exam says…Think…Design
"Withdrew the intervention," "removed tokens," "returned to baseline"IV was pulled backReversal (ABAB)
"Dangerous behavior," "SIB," "skill acquisition," "can't unlearn"No reversal allowedMultiple Baseline
"Staggered introduction," "3 tiers," "across settings/behaviors/participants"Sequential start timesMultiple Baseline
"Alternated conditions," "compared two interventions," "counterbalanced"Rapid switchingATD
"Raised the criterion," "step by step," "increased fluency goal," "shaping"Moving targetChanging Criterion
"BAB design," "started with intervention" (no baseline first)Still a reversal familyReversal variant
"Multiple treatment interference"Limit of ATDATD (limitation)
πŸŽ“ BCBA Exam Prep Platform

Ready to Pass the BCBA?
Get Everything You Need in One Place.

These concept pages are just the start. FlashGenius gives you a complete BCBA prep toolkit β€” practice tests, flashcard decks, cheat sheets, and topic quizzes built for how behavior analysts actually study.

🎯 Practice Tests
πŸƒ Flashcard Decks
πŸ“„ Cheat Sheets
πŸ“Š Topic Quizzes
🧠 Memory Hooks
πŸ“š Study Guides
πŸš€ Start Free on FlashGenius View All BCBA Resources β†’
Free to register Β· No credit card required Β· Trusted by BCBA candidates