Discrepancies between treatment preference and effectiveness

Morris et al. (2024) wrote a think-piece, not an experiment. They asked: what should a BCBA do when a client likes one treatment but data say another works better? The paper walks through four real-world clashes and ends with a decision tree. No new data were collected.

The authors found no magic answer. Instead they offer a four-step rule: (1) check if the client has the prerequisite skills for the effective option, (2) offer an adapted choice format, (3) teach the missing pieces, and (4) only then override preference. The paper is a call for more research, not a final verdict.

Hanley et al. (1997) ran the classic case: FCT and NCR both stopped attention-maintained SIB, yet every kid picked FCT. Preference and effectiveness lined up—no clash needed. That study is the happy scenario Morris et al. say is rare.

Auten et al. (2024) show how to spot the clash before it bites. Their 2018-2023 review says run a concurrent-chains test: let the client sample each intervention first. If the winner matches later outcome data, you’re done. If not, you now have Morris’s problem.

Cerasuolo et al. (2022) warn that child predictors are messy. Higher baseline IQ or language scores do not guarantee success with the “effective” option. So checking prerequisites (Morris step 1) is safer than guessing.

Next time a teen asks for token board while your graph says response-cost is faster, don’t pick sides yet. Run a five-minute concurrent-chains probe (Auten method), then a prerequisite check (Cerasuolo cue). If the preferred option still lags, teach the missing pieces and fade in the stronger protocol. You keep client buy-in without sacrificing gains.