Designing an Engaging Story-stem taken from the MCAST test

The story-stem approach is a reliable and valid method in child psychiatry for investigating children’s social understanding, such as in family/peer relationships. Research is on-going into automating the use of story-stems as the approach takes a lot of time and there are often few administrators trained to administer them. A key part of automation is detecting if a child is engaged in the story-stems; if the child is not engaged, then automated techniques will not work. The study focuses on measuring child engagement using observed facial expressions and specifically on understanding the effect of multimedia digital story-stems on their engagement levels to create a better and more engaging digital story-stem. This ensures that psychiatrists who assess story stems with videos will know whether the child involved is engaged with the story-stems and can be assessed properly. The spontaneous facial expressions of 40 children were collected while watching digital story-stems using an RGB video camera and their engagement levels can be automatically classified using observed facial expressions. Meanwhile, we look at different styles of voice (gender and expressiveness) and video format (animation vs live-action recorded) in the story-stems to see how this affected child engagement. The distribution of child engagement across different media types was compared and the results showed that children were more engaged in a story-stem narrated by a female expressive voice. While comparing the two video formats, although children were more engaged in the current version of animated videos, there were still some factors, such as doll appearance, to be investigated for the design of story-stems.