When the Token Board Stops Working
A 7-year-old named Devon attends a clinic-based program four afternoons a week. His team introduced a token board six weeks ago: one token per completed 5-minute work block, ten tokens exchanged for 10 minutes of tablet time. Data show work completion rose from 40% to 85% in the first three weeks, then fell to about 55% and has stayed there for three weeks. Staff report Devon now completes the first two blocks, then asks for tokens for behaviors like sitting down or handing over a pencil. When staff say tokens are only for work, he drops to the floor and the session pauses. The BCaBA supervising the case has 12 hours per week across nine clients and has not observed Devon in two weeks.
1. What is the most likely function of Devon's recent pattern, and what evidence in the setup supports or fails to support that hypothesis?
Read the worked answer
The pattern is most consistent with escape from the work demand, with a secondary possibility of access to the tablet through the token exchange. The support: completion dropped specifically after the first two blocks, the disruption reliably produces a session pause (escape), and the behaviors he requests tokens for are low-effort alternatives to work. What is not supported yet is a definitive conclusion - the BCaBA has not observed in two weeks, so the antecedent conditions (task difficulty, staffing changes, sleep, medication) are unverified. A defensible next step is a brief descriptive or functional analysis: record what happens in the 30 seconds before the first off-task request, whether task difficulty changed in week four, and whether pausing the session actually removes the demand or merely delays it. The token request itself is ambiguous - it could be a mand for the tablet (tangible) or a mand to avoid work. Distinguishing these requires data on what happens after the request, not staff impressions.
2. Staff propose adding a response-cost component: removing a token for each off-task request. Evaluate this proposal.
Read the worked answer
Response cost could be defensible, but not as proposed here. Three problems. First, the behavior is likely maintained by escape, and removing tokens does not remove the escape contingency - the session pause still follows the request, so the procedure targets the wrong consequence. Second, Devon currently has few tokens at the moment he requests; a cost that drives the balance to zero or negative risks ratio strain and may increase escape behavior rather than suppress it. Third, response cost layered onto a token system that is already losing effectiveness adds a punishment component without a functional assessment, which is ethically and clinically premature. A better sequence: (1) verify the function with direct observation; (2) if escape-maintained, modify the demand (shorter blocks, interspersed easy tasks, high-probability request sequence) and teach an appropriate break mand that is honored before problem behavior; (3) keep the token exchange predictable and rich enough that the tablet remains a credible reinforcer. If response cost is later added, define it precisely, cap losses, and monitor for collateral effects such as increased aggression or decreased manding.
3. The BCaBA has limited hours. What are the minimum supervisory actions needed before any program change?
Read the worked answer
Minimum actions: (1) Direct observation of at least one full session, including the transition into block three, since that is where the data break occurs. (2) Review of the raw data - not the summary percentages - to confirm the drop is real and not a graphing or coding artifact. (3) Staff interview focused on what changed in week four (new staff, new tasks, schedule change). (4) Written treatment-plan update specifying the target behavior, the hypothesized function, and the decision rule for the next two weeks. (5) Training and feedback for direct-line staff on the revised protocol, with a fidelity check. The BCaBA does not need to run the analysis personally, but must ensure the protocol change rests on observed data rather than staff report, and must document the rationale. Delegating observation to a trained RBT with a structured checklist is acceptable if the BCaBA reviews the results before changing the plan.