Model Teachers or Model Students? A Comparison of Video Modelling Interventions for Improving Reading Fluency and Comprehension in Children with Autism

Egarr et al. (2021) compared two video modeling tricks to help kids with autism read better. One group watched a teacher read the page perfectly. The other group watched a short clip of themselves reading the page perfectly—before they had actually done it (feed-forward self-modeling).

The team used an alternating-treatments design. Each child got both types of videos on different days. They measured reading speed and how well kids understood the story.

Only two of the four children read faster after watching the self-model clips. The teacher-model clips gave shaky gains too. Overall, neither video style beat the kids’ own baseline every time.

The authors say video modeling can help some autistic readers, but it is not a sure bet.

Uccheddu et al. (2019) also saw mixed reading results: reading to a dog raised attendance and motivation, yet test scores stayed flat. Together these studies warn that fun extras (dogs, videos) may lift interest without lifting skill.

Cox et al. (2015) used the same alternating design with fourth-graders who had learning disabilities. Teacher-led lessons beat software for real reading growth, even though kids liked the app more. The pattern echoes here—liking the video does not guarantee learning.

Lemons et al. (2015) reviewed dozens of reading trials and found small or negative side effects on behavior and social skills when reading was taught alone. The new data line up: focus on reading first, then add separate plans for behavior if needed.

For BCBAs in school or clinic seats, this means probe both teacher and self-videos, track each child, and stay ready to switch. Pair the video with plain repeated reading, praise, and error correction. If scores do not move after a week, drop the video and try live modeling or direct instruction instead.