I.7: Make data-based decisions about the efficacy of supervisory practices.
Free practice questions for this part of the BCBA, with an explanation for every option.
After a BST package, an RBT's treatment-integrity scores rose from 55% to 92%. The BCBA wants to judge whether the supervision was effective overall. Which additional data are MOST important to review?
- A. The number of supervision hours logged.Incorrect. Logged hours show how much supervision was given, not what it achieved.
- B. The client's target-behavior data, to see whether better implementation led to client progress.Correct. Supervision is ultimately judged by improvements in supervisee performance and the related client outcomes.
- C. How long the RBT has been employed at the agency.Incorrect. Length of employment does not show whether this supervision worked.
- D. The RBT's self-rating of confidence.Incorrect. Self-ratings can add information but are not a direct measure of effectiveness.
A BCBA evaluates a feedback package with three RBTs. She introduces it at different times for each RBT, in a multiple-baseline design. RBT 1's integrity improves right after the package starts. RBTs 2 and 3 improve during their own baselines at the same time, before the package reaches them. What is the MOST appropriate conclusion?
- A. The design should be abandoned and results judged from final scores alone.Incorrect. Final scores alone cannot establish what caused the change. The supervisor should look for confounds and strengthen the design.
- B. A functional relation is clearly shown, because all three RBTs improved.Incorrect. Improvement before the intervention in untreated baselines undermines the replication needed to show a functional relation.
- C. The package was ineffective for RBTs 2 and 3.Incorrect. Their performance already improved before they received the package, so its effect on them cannot be judged.
- D. Experimental control is weak. The baseline changes suggest another variable (or interdependence between RBTs) may explain the improvement.Correct. In a multiple baseline, change in the untreated tiers threatens internal validity and suggests confounding variables or tiers that are not independent.
An RBT's integrity data have stayed around 60%, with no upward trend, over four weeks of weekly verbal feedback. What is the MOST appropriate data-based decision?
- A. Stop measuring integrity, because the data are not changing.Incorrect. Removing measurement takes away the basis for decisions.
- B. Report the RBT for an ethics violation.Incorrect. Low integrity is a performance problem to solve. It is not, by itself, an ethics violation.
- C. Keep the same feedback for another two months.Incorrect. A stable, low trend under the current approach shows that a change is needed, not more of the same.
- D. Change the supervisory approach, for example by doing a performance diagnostic and then adjusting the intervention (such as adding rehearsal, graphed feedback, or more frequent feedback).Correct. When data show no progress, supervisors should evaluate why and change their practices (Ethics Code 4.10).