D.6: Critique and interpret data from single-case experimental designs.
Free practice questions for this part of the BCBA, with an explanation for every option.
A BCBA graphs a client's correct mands per session. Across 5 baseline sessions, mands rise steadily from 2 to 10. The intervention starts in session 6, and mands keep rising at the same rate to 18 by session 10. What is the most defensible interpretation?
- A. The intervention is ineffective and should be stopped immediatelyIncorrect. The data do not show that the intervention fails either. They simply cannot separate its effect from the existing trend.
- B. A functional relation cannot be concluded, because the baseline trend already predicted the improvement seen in interventionCorrect. When baseline is trending in the therapeutic direction, the same improvement during intervention is what the baseline predicted, so no effect is shown. Ideally, the analyst waits for a stable baseline or uses a different design.
- C. The data show an immediate change in level, which confirms experimental controlIncorrect. There is no change in level or trend at the phase change. Responding continues on the same path.
- D. The intervention is clearly effective because mands nearly doubledIncorrect. The increase continues an already-improving baseline trend, so it cannot be credited to the intervention.
In a multiple-baseline-across-settings design, a self-management intervention starts in the classroom. On-task behavior rises right away in the classroom, and also rises at the same time in the cafeteria and the library, where baseline conditions are still in effect. How should this be interpreted?
- A. Experimental control is strongly shown, because behavior improved in all three settingsIncorrect. Multiple-baseline logic needs the untreated tiers to stay stable as verification. Simultaneous change in them undermines that verification.
- B. The design is a valid changing-criterion design, because performance changed in stepsIncorrect. No stepwise criteria were set. This is a multiple-baseline design, and the pattern is not stepwise matching to a criterion.
- C. The other two settings should be dropped from analysis, and the classroom data treated as an AB demonstration of controlIncorrect. Dropping the inconvenient tiers misrepresents the data, and an AB comparison cannot show experimental control anyway.
- D. Experimental control is weakened, because the untreated tiers changed before the intervention was introduced to themCorrect. The covariation suggests either that the tiers were not independent (behavior generalized) or that an outside variable affected all settings, so a functional relation cannot be shown with confidence.
A BCBA uses an A-B-A-B design to evaluate teaching a client to initiate greetings with peers. During the return to baseline (second A), greetings stay at the high intervention level for all 8 sessions, and peers are seen responding warmly to them. What is the best interpretation?
- A. Responding may now be maintained by natural contingencies, so the reversal does not show experimental control; a design that does not need reversal would be betterCorrect. Once newly taught social behavior contacts natural reinforcers, it may not reverse. Irreversible behaviors are better evaluated with designs such as multiple baseline.
- B. The data prove the behavior was never under the control of the interventionIncorrect. Failure to reverse does not prove the intervention was ineffective. The behavior may have contacted natural reinforcement (a behavioral trap).
- C. Observers must have stopped collecting data accurately during the second baselineIncorrect. Nothing in the scenario suggests a measurement problem. Irreversibility is an expected outcome for some skills.
- D. The intervention clearly caused the behavior, because responding stayed highIncorrect. Without a return toward baseline levels, the reversal design provides no verification, so a functional relation is not shown.