A comparison of the feedback sandwich, constructive-positive feedback, and within session feedback for training preference assessment implementation

Bottini et al. (2021) ran a small lab trial with adult learners. Each person got one of three feedback styles while learning to run a preference assessment.

The three styles were: feedback sandwich (positive-negative-positive), straight constructive-positive, or quick corrections given right in the middle of the role-play.

After the first role-play, the within-session group scored highest on fidelity checks. Their steps were cleaner and faster.

By the third role-play, all three groups looked the same. Skill scores and satisfaction ratings were even. The early edge had washed out.

Wine et al. (2019) had already shown that timing—before or after the task—made no difference. Bottini’s team adds that the order of praise and correction also stops mattering once staff have had a few practice loops.

Konstantinidou et al. (2023) reviewed dozens of staff-training papers and found most studies only check staff behavior, not client benefit. Bottini’s lab fits that pattern: they measured how well the learner ran the assessment, not whether the child’s preferences were actually identified.

Johnson et al. (2023) argue you should first pick the function you want feedback to serve. Bottini’s results support that idea—if the function is rapid first-time accuracy, within-session correction works. If the function is long-term fluency, any respectful sequence does the job.

If you are training new RBTs this week, give immediate within-session fixes for the first one or two practice rounds. You will see cleaner steps faster. After that, stop worrying about sandwich versus straight praise. Just keep the feedback coming and let practice do the rest. Rotate styles to match your supervisee’s preference and your own teaching comfort—no sequence is king after the early burst.