Under the Desk During Independent Math

Operational definition first: 'Priya leaves her seat, cries, or places her head/body under the desk for more than 30 seconds during independent math, ending when the assignment is removed or she leaves for the nurse.'

Working hypothesis: behavior is maintained primarily by escape from a demanding, low-preference task, with a possible secondary function of adult attention (the escort to the nurse is one-to-one attention). The pattern of success in small-group reading and hands-on centers and failure only during independent seatwork supports a task/response-demand hypothesis rather than a global anxiety or skill-based explanation -- though both must be checked.

Data to collect

• Indirect: teacher and parent interview (FACTS or similar), review of prior records, nurse visit log.
• Direct: ABC recording across at least three math periods; partial-interval recording of on-task behavior; latency from task start to first episode; conditional probabilities (episodes following which task types: two-digit subtraction? timed worksheets? writing demands?).
• Structural: does the behavior occur when a peer or adult is nearby? Does it occur on days the nurse is in the building?
• Skill check: is her math computation and math vocabulary adequate, or is the work genuinely beyond her instructional level?
• Social-emotional screening with a validated tool to see whether anxiety is broad or situation-specific.

If a brief experimental or structured antecedent manipulation is safe and feasible, test whether reducing task difficulty or offering a break option reduces episodes.

Functionally, it targets the wrong thing. If the function is escape, the contingencies do not touch the antecedent: the worksheet is still present, still hard, and still delivered with no support. Response cost on recess adds a punishing consequence that could increase avoidance of school generally and may create a new escape (refusing to go to recess).

• The reprimand and escort may reinforce the behavior if adult attention is a secondary function.
• It teaches nothing. A behavior plan must build a functionally equivalent replacement (asking for a break, requesting help, using a coping strategy) and reinforce it richly.
• Missing instruction. Removing the assignment guarantees she never practices the skill, so the academic gap widens and the task becomes harder -- a self-sustaining loop.
• Legal/practice concerns. A pattern of removing recess can accumulate into a disciplinary change of placement; excluding her from Fun Friday for disability-related behavior without an FBA/BIP and without considering services is risky. Physical escort and any restraint or seclusion would require specific policy and training, and seclusion is not warranted here.

Better package: modify the task (chunk it, pre-teach, reduce length, provide a model, offer choice of order), teach and honor a break card with a brief return, reinforce engagement and help-seeking, keep responses to escalation neutral and brief, and reserve the nurse for medical needs.

• Direct behavior outcomes: frequency and duration of episodes, latency to first episode, percentage of intervals on task, and math work completion/accuracy. Graph weekly against baseline.
• Replacement behavior: rate of appropriate break requests and help-seeking -- if these do not rise, the plan is not teaching.
• Treatment integrity: a short checklist observation of the teacher (e.g., were tasks chunked, was the break card honored within one minute, was praise delivered at the planned rate). Target at least 80% of steps; low integrity means retrain before concluding the plan failed.
• Dosage and setting events: attendance, sleep or health complaints, nurse visits, and time of day.
• Academic response: brief progress monitoring on the math skill itself, since skill deficit may be driving escape.
• Social validity: ask Priya and the teacher whether the plan is feasible and acceptable; a plan nobody implements will fail regardless of design.
• Decision rule in advance: e.g., if episodes do not decline by 50% after four weeks with integrity at or above 80%, revise the hypothesis and consider a fuller FBA or a functional analysis.