A behavioral system for assessing and training cardiopulmonary resuscitation skills among emergency medical technicians.

Williams et al. (1986) built a CPR course for emergency medical technicians. The course broke every chest-compression and airway step into tiny, visible actions.

Trainees watched a demo, practiced on manikins, and got instant yes/no feedback. The system counted correct moves and showed graphs after each round.

A quasi-experiment compared this behavioral class with the usual lecture-plus-demo that local squads had used for years.

EMTs who learned through the behavioral system scored higher on every skill check. Their compressions were deeper, faster, and in the right spot.

Months later the behavioral group still passed most items while the lecture group had slipped back toward beginner levels.

The study copies the same four-step recipe first mapped by Smith et al. (1975): tell, show, practice, praise. E et al. simply added finer measures and longer follow-up.

Davison et al. (1984) tried the same recipe with child-psychiatry staff and also beat lectures, but they stopped measuring after training. E et al. proves the edge can last.

Gray et al. (2026) later moved the recipe online and got 90% fidelity without a classroom. The CPR paper shows why that works: clear task analyses and live feedback do the heavy lifting, not the room itself.

If you train staff or caregivers, swap some talk time for brief demos, rehearsal, and on-the-spot data. You do not need manikins; any skill you can see and count will do. Graph the results each session and send trainees home when the line hits mastery, not when the clock runs out.