An evaluation of preference for video and in vivo modeling.

Three autistic children watched short clips of a task. Then they watched an adult do the same task live.

The team counted how fast each child learned and which model the child liked more.

Sessions alternated each day so every kid tried both ways.

All three children learned the skill at the same speed, no matter which model they saw.

When asked to pick, none of the kids cared if the model was on a screen or in the room.

Wilson et al. (2020) saw something different. Their teens with autism learned cooking faster with pure video modeling than with video prompting. The 2010 study and the 2020 study look opposite, but they tested different things. The 2010 paper compared video to live; the 2020 paper compared two kinds of video support.

Piraneh et al. (2022) also found video modeling beat social stories for tooth-brushing. Again, the win went to video, yet the 2010 paper says video equals live. The gap shows video shines when the other choice is a weaker tool like a story or prompting, not when the rival is a live expert.

Richman et al. (2001) asked a similar question earlier. They saw no speed gap between self-video and other-video for conversation skills. That null result matches the 2010 null result: once you stay inside the video family, format rarely matters.

You can stop worrying about the perfect model. If a live therapist is busy, press play. If the tablet is charging, model live. Kids learn either way, so pick the option that saves you time and resources.