Alternating Treatments in One Classroom: Confounded Comparison of Token Economy and Response Cost
A school-based BCBA is asked to compare a token economy with a response-cost system for 8-year-old Devon, whose disruptive comments occur 10–14 times per 45-minute class. The principal wants an answer in four weeks. The BCBA has one classroom, two teachers, and 16 days. She plans an alternating-treatments design: token economy on Mondays/Wednesdays, response cost on Tuesdays/Thursdays, and no plan on Fridays. Teacher A runs token days; Teacher B runs response-cost days because of scheduling. Baseline data are collected for only three days (18, 11, and 16 disruptions) before the comparison begins. After 12 days, token days average 6 disruptions and response-cost days average 9; however, the first response-cost day after a long weekend had 17, and token days after response-cost days are lower than token days after token days. The principal says the data “prove token economies work.”
1. Identify the design and analysis problems, including confounds and threats to internal validity.
Read the worked answer
The design is an alternating-treatments (multi-element) comparison, but it is confounded. Teacher is perfectly confounded with condition: Teacher A always runs token economy and Teacher B always runs response cost. Day of week is also confounded, and Fridays have no plan, so the “no plan” days may affect Monday data. Baseline is only three days and variable (18, 11, 16), so there is no stable comparison. Carryover and sequence effects are visible: token days after response-cost days are lower than token days after token days, and the first response-cost day after a long weekend is an outlier. There is no reported IOA or procedural integrity. Small n and one classroom limit generality. The principal’s “prove” conclusion is not supported.
2. What would you recommend before making a recommendation to the principal, and how would you interpret the current data?
Read the worked answer
Before advising adoption, I would: (a) collect a longer baseline or at least demonstrate stability; (b) randomize and counterbalance condition order; (c) have the same teacher implement both conditions or counterbalance teachers across conditions; (d) keep Friday as a control or include it in the alternation; (e) collect IOA on 30% of sessions and integrity checks; and (f) use direct observation of disruptive comments. For analysis, inspect level, trend, variability, immediacy, and overlap; do not rely on averages alone. Current data are preliminary and confounded. I would tell the principal that token economy is a reasonable candidate, but the comparison does not isolate its effect. If a decision is required now, choose based on least restrictive options, social validity, and continue evaluation with a proper design.
3. If the principal insists on adopting the token economy immediately, what consultation and ethical considerations apply?
Read the worked answer
I would not misrepresent the data as proof. I would say the token economy is promising but not experimentally demonstrated by this comparison. If the principal adopts it anyway, I would frame that as a clinical decision, not an evidence claim, and build in progress monitoring: daily disruptive comments, integrity checks, and a plan to revise if data do not improve. I would ensure the token system is feasible, acceptable to the teacher, and not punitive or stigmatizing. I would also clarify consent and communication with the family. Ethically, the consultant should avoid pressure to overstate findings, document the limitations, and continue to evaluate. If the principal forbids further evaluation, I would document that constraint and recommend a minimal data system so the decision can be revisited.