cumulative

D.7 Reversal, multiple-baseline, multielement, and changing-criterion designs

What the task asks: Tell the main single-case designs apart by how each arranges conditions to show experimental control, and know the strengths and limits of each.

Experimental Design6 min read5 practice questionsOfficial outline (opens in a new tab)

The big idea

All single-case designs use baseline logic (D.4), but they arrange conditions differently to produce verification and replication. A reversal design introduces and withdraws the intervention. A multiple-baseline design staggers it across behaviors, settings or people. A multielement design alternates conditions rapidly. A changing-criterion design shifts a performance criterion in steps.

Each design fits some situations and not others. The exam often describes a study and asks which design it is, or describes a situation and asks which design to use (see D.9).

Breaking it down

Reversal design (A-B-A-B)

Baseline (A), intervention (B), a return to baseline (A), then the intervention again (B). The return to baseline verifies the prediction; the second B phase replicates the effect. Ending in an intervention phase leaves the client with the treatment in place.

0481216BaselinePlanBaselinePlan24681012141618SessionsCall-outs per session
A reversal (A-B-A-B) graph of Femi's call-outs. They drop when the plan starts, climb back when it is withdrawn, and drop again when it returns: prediction, verification, replication.
  • Strength: the clearest demonstration of a functional relation with one participant.
  • Limits: withdrawing an effective treatment may be unethical when behavior is dangerous, and some behavior will not reverse (a learned skill, or behavior now kept going by natural reinforcers).
  • Variations: B-A-B when treatment cannot wait; multiple-treatment reversals such as A-B-A-C-A; and reversal techniques that replace withdrawal with a control condition, such as noncontingent reinforcement or DRO in place of contingent reinforcement.

Terms vary. Some texts call the A-B-A-B arrangement a withdrawal design and keep reversal for designs that actually reverse the contingency. Cooper et al. (2020) use reversal design for both, and the exam usually does too.

Multiple-baseline design

Baselines run at the same time for two or more tiers: different behaviors of one person, one behavior in different settings, or one behavior across different people. The intervention reaches one tier at a time, in a staggered order, after the previous tier has changed. Untreated tiers staying at baseline verify; each later tier changing when treated replicates.

050100MathBaselineSelf-monitoring050100Science050100English123456789101112131415SessionsPercent of intervals on task
A multiple baseline across settings for Ruth's on-task behavior. Each class changes only when self-monitoring reaches it, while the untreated classes hold steady. The stepped dashed line marks the staggered start.
  • Strength: no withdrawal, so it suits dangerous behavior and behavior that will not reverse.
  • Limits: later tiers wait longer for treatment, and tiers must be independent (treating one does not change the others) yet similar enough to respond to the same IV.
  • Concurrent vs. nonconcurrent: in a concurrent design all baselines run during the same calendar period. A nonconcurrent design, usually across participants, starts baselines at different times. It is practical, but it controls history less well, because the tiers do not share the same outside events.

Multiple-probe design

A variation of the multiple baseline in which baseline data come from intermittent probes rather than every session (Horner & Baer, 1978). It fits skills taught in sequence, or cases where continuous baseline testing would be impractical, reactive or frustrating for the learner. A short run of consecutive probes just before each tier is treated confirms that its baseline is still steady.

Multielement design (alternating treatments)

Two or more conditions alternate rapidly, often session by session, in a counterbalanced or random order. Each is paired with a distinct stimulus so the participant can tell them apart, and each gets its own data path. An effect shows up as separation between the paths. Functional analyses (F.6) use this design.

Video promptsPicture prompts0255075100123456789101112DaysPercent of steps correct
A multielement graph for Ana's cooking steps. Picture and video prompts alternate day by day, each with its own data path. The video path sits consistently above the picture path, so video prompts work better for Ana.
  • Strengths: fast comparisons, no withdrawal, and no baseline phase required.
  • Limits: multiple-treatment interference, conditions that are hard to tell apart, and a pace of switching unlike everyday life.

Some authors use multielement for alternating stimulus conditions and alternating treatments for comparing treatments; Cooper et al. treat the terms as near-synonyms.

Changing-criterion design

After baseline, the intervention runs in a series of subphases, each with a stepwise criterion for performance (for example, earning a reward for 5, then 8, then 11 math problems). Control is shown when behavior changes to match each new criterion. Varying subphase length and step size, and occasionally returning to an easier criterion, makes the demonstration more convincing (Hartmann & Hall, 1976).

012345Baseline2,0003,0002,0004,0002468101214161820DaysSteps per day (thousands)
A changing-criterion graph for a walking goal. The dotted line in each subphase is the criterion (2,000, then 3,000 steps, a planned return to 2,000, then 4,000). Steps track each criterion, including the step back down.
  • Strength: fits gradual change toward a goal, such as more exercise or less caffeine.
  • Limits: the behavior must already be in the repertoire, and the criterion can hold back performance that could have improved faster.

Seeing it in practice

DesignExampleHow control is shown
ReversalA plan for Femi’s call-outs, run A-B-A-B over four weeksCall-outs rise and fall with each change in condition
Multiple baseline across settingsSelf-monitoring for Ruth’s on-task behavior, started in math, then science, then EnglishEach class changes only when self-monitoring starts there
Multiple probe across skill setsTeaching Bilal three sets of sight words, with weekly probes of untaught setsEach set improves only after it is taught
MultielementPicture prompts and video prompts for Ana’s cooking steps, alternated dailyThe video data path is consistently higher
Changing criterionPoints for an adult client who walks 2,000, then 3,000, then 4,000 steps a daySteps track each new criterion

Common mix-ups

  • Multiple baseline needs no withdrawal. Verification comes from untreated tiers, not from reversing.
  • Staggering is the point. If every tier gets the intervention at the same time, there is no verification.
  • A multielement design is not a reversal. Conditions alternate rapidly rather than in long phases.
  • Changing criterion is for behavior already in the repertoire. A new skill calls for shaping, or a multiple-baseline or multiple-probe design.

Exam tip: Look at how conditions are arranged. Introduced and withdrawn: reversal. Staggered starts across tiers: multiple baseline. Intermittent baseline probes: multiple probe. Rapid alternation with separate data paths: multielement. Stepwise criteria: changing criterion.

Quick recap

  • Reversal (A-B-A-B) verifies by withdrawal. Avoid it when withdrawal is unsafe or behavior will not reverse.
  • Multiple baseline staggers the IV across behaviors, settings or people; concurrent is stronger than nonconcurrent.
  • Multiple probe uses intermittent baseline probes for skills taught in sequence.
  • Multielement alternates conditions rapidly; watch for multiple-treatment interference.
  • Changing criterion shows control when behavior tracks stepwise criteria.

Further reading

  • Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapters 8 and 9.
  • Horner, R. D., & Baer, D. M. (1978). Multiple-probe technique: A variation on the multiple baseline. Journal of Applied Behavior Analysis, 11(1), 189-196.
  • Hartmann, D. P., & Hall, R. V. (1976). The changing criterion design. Journal of Applied Behavior Analysis, 9(4), 527-532.

Key terms

Reversal design (A-B-A-B design, Withdrawal design)
A single-case design in which baseline and intervention conditions are introduced, withdrawn and reintroduced, typically in an A-B-A-B sequence. The return to baseline provides verification, and the reinstated intervention provides replication.
Multiple-baseline design
A single-case design in which baselines are measured at the same time for two or more behaviors, settings or participants, and the intervention is introduced to them one at a time in a staggered sequence. Untreated tiers provide verification, and each tier that changes when treated provides replication.
Multiple-probe design
A variation of the multiple-baseline design in which baseline data are gathered through intermittent probes instead of continuous measurement, with a short series of consecutive probes just before each tier receives the intervention. It suits skills taught in sequence and cases where continuous baseline testing would be impractical or aversive.
Multielement design (Alternating treatments design, ATD)
A single-case design in which two or more conditions alternate rapidly, often from session to session, each paired with a distinct stimulus and plotted as its own data path. Differences between the data paths show the effect of the conditions.
Changing-criterion design
A single-case design in which, after baseline, an intervention is applied in a series of steps, each with a new criterion for performance. Experimental control is shown when behavior changes to match each new criterion.
Multiple-treatment interference (Sequence effects, Carryover effects)
A confound in which a participant's behavior in one condition is influenced by experience with another condition in the same study, either through the order of conditions (sequence effects) or through the lingering effects of an earlier condition (carryover effects). It is a concern in multielement designs and in reversal designs with more than one treatment.

Check yourself

5 questions on D.7. Choose an answer to see whether you're right and why.

  1. Question 1A researcher measures a student's disruptive behavior in three classes. After stable baselines in all three, the researcher introduces a token system in the first class, two weeks later in the second, and two weeks after that in the third. Which design is this?
    Show the answer

    Answer: A. Baselines run at the same time in three settings, and the intervention starts in each one at a different time. That staggered introduction is the defining feature of a multiple-baseline design. Nothing is withdrawn, so it is not a reversal design.

  2. Question 2A behavior analyst alternates two error-correction procedures across sessions for a learner, in random order, with each procedure always run at a different table. The data path for one procedure is consistently higher. Which design is this, and what is its main risk?
    Show the answer

    Answer: A. Rapid alternation of conditions, each paired with a distinct stimulus and shown as its own data path, defines a multielement (alternating treatments) design. Because the learner experiences both procedures close together, one can influence responding under the other. It is not a reversal design, because no condition is withdrawn and later reinstated across phases.

  3. Question 3An adult client wants to cut down from eight cups of coffee a day. The plan sets a limit of six cups for a week, then four, then three, with a reward each day the limit is met. Consumption closely matches each limit. Which design is this?
    Show the answer

    Answer: B. The intervention stays the same while the performance criterion shifts in steps, and control is shown when behavior tracks each new criterion. A multiple-baseline design would stagger the intervention across behaviors, settings or people instead.

  4. Question 4In a multiple-baseline design, what provides verification?
    Show the answer

    Answer: D. While one tier receives the intervention, the tiers still in baseline show what would have happened without it. If they stay unchanged, the prediction is verified. Withdrawal is how a reversal design verifies, and starting all tiers on the same day would remove verification entirely.

  5. Question 5A team plans to teach a learner four related skills in sequence. Testing every untaught skill every session would take a long time and expose the learner to repeated failure. Which design best fits?
    Show the answer

    Answer: A. A multiple-probe design keeps the staggered logic of the multiple baseline but samples untaught skills with intermittent probes instead of continuous testing. A reversal design is a poor fit because learned skills are unlikely to return to baseline.

Adds a step to your study record. Saved in this browser only.

Type to search all 104 guides and the glossary.