Effects of programmed teaching errors on acquisition and durability of self-care skills

Donnelly et al. (2017) asked what happens when you mess up a teaching step. They taught self-care chains like hand-washing to kids with developmental delays. Then they tried three common mistakes: giving the next prompt too soon, waiting too long to give praise, or skipping a step.

Every single error hurt learning. Kids took longer to master the chain. Worse, the skills they already had fell apart when the error showed up again. Even a five-second delay in praise was enough to break the chain.

Sureshkumar et al. (2024) looks like a contradiction. They used video prompts through Zoom and got big, lasting gains in first-aid skills. The key difference: their video never gave a prompt out of order and praise came right on time. The tech kept the timing perfect, so no Donnelly-style errors occurred.

Griffen et al. (2023) extends the warning. They gave staff an AI coach that buzzed the instant a prompt was late or out of order. Fidelity shot up and kids learned faster. The study shows one fix for the exact errors Donnelly proved are harmful.

Rojahn et al. (1994) adds a twist. They found a short pause AFTER feedback helped college kids. Donnelly shows a pause BEFORE feedback hurts kids with delays. Same variable—timing—but opposite effects because the learners and tasks differ.

Your prompt sequence is fragile. One late “good job” or one skipped step can undo weeks of work. Run a quick fidelity check before each session: Are prompts in the right order? Is praise within two seconds? If not, fix it on the spot. Your learner’s skill—and your time—depend on it.