/ Single-Case Design Graphs

Overview › Design guide › Alternating treatments

Design guide

Alternating treatments and multielement designs

Alternate two or more conditions rapidly, often session by session, and compare their data paths. A functional analysis is a multielement design.

You will be able to

  • Explain how rapid alternation shows experimental control, and when to use the adapted design to compare ways of teaching skills.
  • Judge whether a functional analysis is differentiated by comparing each test condition with play.
  • Decide whether alternating treatments graphs show control and whether functional analyses are differentiated, reaching 8 of your last 10 in the practice.
Question it answersWhich of two or more conditions produces more (or less) behavior?
Use it whenConditions can switch quickly, are easy to tell apart, and you want an answer fast. Functional analyses use this design.
Avoid it whenOne condition's effect is likely to carry over into the next, or the intervention takes many sessions to work. To compare ways of teaching a skill, which stays learned, use the adapted version.

How it shows control

Every switch between conditions is a chance to see the difference again. If one condition's data path stays clearly above or below the other's through many alternations, the difference is replicated each time. Separation without overlap is the key feature. Counterbalance the order of conditions, and give each condition a distinct cue, so the order can't explain the difference.

The logic of this design in more depth, its judgment calls and exceptions, and a knowledge check: Visual analysis Lesson 3, Part D.

Minimum standards

What experimental control looks like

Experimental control Alternating treatments graph of responses per minute. Baseline is 1.9 to 2.5 over 4 sessions. In the alternating phase, Treatment A falls from 1.2 to 0.2 while Treatment B stays between 1.6 and 2.0, with no overlap.
Clear, consistent separation. Treatment A falls below Treatment B from the first alternation and stays there. Each treatment has six sessions, and every Treatment A point is below every Treatment B point, so the difference is replicated each time the treatments alternate.
No difference Alternating treatments graph of responses per minute. Baseline is 1.9 to 2.5 over 4 sessions. In the alternating phase, Treatment A (1.2 to 1.9) and Treatment B (1.3 to 1.9) overlap and cross repeatedly.
Overlapping paths. The two treatments cross repeatedly and occupy the same range. Neither is shown to be more effective. Both are lower than baseline, but that drop happened once, at the start of the alternating phase. Like an AB comparison, it doesn't show that either treatment caused it.

Comparing ways to teach skills: the adapted design

The standard design needs behavior that can rise and fall from session to session. A skill, once learned, stays learned, so two teaching procedures can't take turns on the same targets. The adapted alternating treatments design gives each procedure its own targets instead (Sindelar et al., 1985).

Functional analysis: differentiated or not?

In a multielement functional analysis, each social test condition arranges one possible reinforcer, such as attention, escape from demands, or tangibles. The alone condition provides no social consequences at all, and a play condition serves as the control. The analysis is differentiated when responding is consistently higher in one or more test conditions than in play. Compare each test condition with play, not just with the highest condition. A condition only moderately above play still counts when its points sit consistently above play's. When more than one test condition is consistently above play, the behavior may be maintained by more than one reinforcer (multiply controlled). When only the alone condition is consistently above play, the behavior may produce its own reinforcer (automatic reinforcement). Structured criteria exist for making this judgment reliably (Hagopian et al., 1997).

Differentiated Functional analysis graph of responses per minute with five cycles of five conditions. Demand is 2.5 to 3.6 in every cycle. Play, Attention, Tangible, and Alone all stay between 0 and 0.3 and overlap.
Escape function. Responding in Demand is above Play in all five cycles. Attention, Tangible, and Alone stay at the same low level as Play, so only the escape contingency is differentiated.
Undifferentiated Functional analysis graph of responses per minute with five cycles of five conditions. Play, Attention, Demand, Tangible, and Alone all fall between 1.7 and 2.9 and overlap throughout.
No condition stands out. Responding is high and overlapping in every condition, including Play. That doesn't identify a function. High responding in all conditions, including Alone, can suggest automatic reinforcement, but this needs follow-up, such as a series of extended alone sessions.
Undifferentiated Functional analysis graph of responses per minute with five cycles of five conditions. Play, Attention, Demand, Tangible, and Alone all stay between 0 and 0.4 and overlap throughout.
Too little responding to compare. Responding stays between 0 and 0.4 per minute in every condition, test conditions and Play alike, so no condition stands out and no function is identified. The behavior may have occurred too rarely during the sessions to show a difference. Next steps can include longer sessions, or test conditions changed to match the situations where the behavior is reported to happen.

Comparing every test condition with Play, why the demand condition can be capped, and a knowledge check: Visual analysis Lesson 3, Part E.

Practice

Decide whether each alternating treatments graph shows experimental control, and whether each functional analysis is differentiated. Most items then ask a second question: when an alternating treatments graph shows control, which treatment produced less responding; when a functional analysis is differentiated, which test conditions are consistently above Play; and when it isn't, why not. An item counts only when both answers are right.

Graph it