The Vanishing Praise Effect
A clinic supervisor reviews data from a newly hired RBT working with a 9-year-old who engages in aggression maintained by escape from academic demands. The behavior plan specifies differential reinforcement of alternative behavior (DRA) with a 30-second break contingent on requesting a break, plus extinction of aggression (escape no longer follows aggression). Baseline aggression averaged 6.2 per hour. After four weeks, aggression is 5.8 per hour, but break requests have risen from 0.5 to 9 per hour. The RBT reports the child now requests a break "constantly" and that staff have begun occasionally denying requests when they judge them excessive. Session notes show the RBT delivers the break about 40 seconds after a request on average, sometimes finishing a prompt first. The supervisor must decide what to change and how to document the decision.
1. Interpret the data. Is the intervention failing? What is the most likely account of the near-flat aggression data, and what evidence would you gather before changing the plan?
Read the worked answer
The intervention is not clearly failing on the DRA dimension: break requests rose eighteen-fold, showing the alternative response is being acquired and reinforced. The flat aggression data are consistent with several competing accounts that must be distinguished before changing anything.
First, treatment integrity: a mean 40-second delay between request and break, plus occasional denial, constitutes intermittent and delayed reinforcement of the alternative response. If the child sometimes requests and does not get a break, requesting is on a leaner schedule than the plan intends, and aggression may persist because escape occasionally still follows it (adventitious reinforcement). Second, the denial practice may itself reinstate an escape contingency for aggression if staff ever terminate a demand to stop escalation.
Evidence to gather: (a) direct observation or video review of request-to-break latency and denial frequency; (b) a conditional probability analysis of what follows aggression within 30 seconds; (c) whether break requests occur across settings or only during high-demand tasks; (d) whether the child's requests are functionally appropriate (asking versus protesting). Only after integrity and contingency data are collected should the plan be revised. The supervisor should document the integrity concern, retrain, and set a decision rule (for example, if integrity reaches 90% and aggression does not decline within two weeks, conduct a functional analysis of the alternative response).
2. The RBT's informal practice of denying "excessive" break requests is not in the plan. Analyze this from an ethics and supervision standpoint and describe the supervisory response.
Read the worked answer
The RBT is implementing a procedure not authorized by the plan, which raises both ethical and clinical problems. Ethically, behavior analysts are responsible for the services they design and supervise, must ensure technicians are trained and competent, and must not allow unapproved modifications that could harm the client. Clinically, ad hoc denial undermines the DRA contingency, may create an extinction burst for requesting, and could increase aggression if escape is intermittently available.
The supervisory response should be sequential and documented:
- Clarify the rationale to the RBT: explain why consistent, immediate reinforcement of requesting is the mechanism of change, and that "excessive" requesting is expected early in DRA and typically decreases as requesting becomes efficient.
- Retrain to criterion using modeling, role-play, and a fidelity checklist covering latency, denial, and consequence delivery.
- Increase observation frequency (for example, two direct observations per week with feedback) until integrity is stable at or above 90%.
- If the RBT believes the plan is impractical, create a formal channel: the RBT reports concerns, the supervisor evaluates data, and any change is made in writing with caregiver consent.
- Document the corrective action, including dates, content of retraining, and integrity data.
If denial continues after feedback, the supervisor must consider reassignment or a formal performance improvement plan, because continued unauthorized modification places the client at risk.