Effectiveness of video modeling in teaching earthquake and postearthquake evacuation safety skills for children with autism

Kurt et al. (2024) showed three autistic children short videos of kids dropping, covering, and holding on during an earthquake, then walking outside after the shaking stopped. No adult gave live prompts or rehearsal; the kids just watched the clips. The team later staged fake quakes with sound effects and checked if the children copied the video steps.

All three children learned the full sequence: drop, cover head, hold on, wait, then evacuate. They still did it correctly four weeks later and used the skills in a new room they had never practiced in.

Ferris et al. (2025) took the same idea into immersive VR. Their kids walked on a treadmill while a virtual car approached; after VR training, most could cross real streets safely. The VR study extends Kurt’s video-only method by adding motion and real-time feedback, but both report strong gains.

Northup et al. (1991) did the opposite swap: they taught non-verbal adults with severe ID to leave a group home when the fire alarm rang. Instead of videos, staff used live prompts, modeling, and candy. The outcome matches Kurt’s—everyone learned to evacuate—showing the medium can change while the teaching bones stay the same.

Fallea et al. (2025) ran an RCT comparing VR tooth-brushing lessons with standard pictures. VR won, giving a conceptual replication: screen modeling helps autistic kids whether the skill is earthquake safety or oral hygiene.

You can build an earthquake drill in minutes. Film a peer doing drop-cover-hold, play it on a tablet, then test with a surprise drill. If you already own VR goggles, try Ferris et al.’s treadmill setup for street crossing. Either way, let the screen do the first teaching so you save staff time for fine-tuning and praise.