An evaluation of the production effects of video self-modeling.

García-Villamisar et al. (2017) filmed kids with autism and intellectual disability doing daily tasks. They used a phone held at the child’s eye level to make short clips.

The team ran a multiple-baseline design across chores like setting a table or packing a backpack. They first checked baseline accuracy, then filmed the child doing the task, then let the child watch the clip, and finally checked accuracy again.

Accuracy jumped right after the filming session—before the child ever saw the video. Later viewing added a small extra boost, but the big win came from simply making the movie.

In plain words, practicing while being filmed was more powerful than watching the finished product.

Hong et al. (2016) pooled dozens of video-modeling studies and found solid gains for daily-living skills. D et al. zoom in and show one reason why: the filming itself teaches.

Marcus et al. (2009) saw self-video beat peer-video for teaching letters. D et al. extend that idea to chores and add the twist that the camera, not the replay, drives most change.

Wilson et al. (2020) compared video modeling to video prompting and picked modeling. Their result aligns with D et al.—fewer steps and less confusion when the clip shows the whole task straight through.

If you use self-modeling, treat the filming like a rehearsal session. Prompt the learner through the task, hit record, and give immediate praise. You may see a jump in accuracy before you even press play. Save the viewing for later review or maintenance, but don’t skip the camera practice—it’s half the intervention.