Individuals with autism often have difficulties with social skills that interfere with their educational experience and everyday life. Social Stories is a paradigm that is commonly used to address these difficulties. Our goal is to develop adaptive technology that would assist caregivers of these individuals to create customized social skills instructional modules that augment this paradigm.
Objectives:
Refl-ex is a tool that helps individuals practice social skills by presenting them with a social scenario. Here an unexpected obstacle arises, and the system guides them through the process of overcoming the obstacle. The introduction of an obstacle is an important part of our approach, and differentiates us from existing social skills software; however, it makes the modules difficult and time consuming to author. Our current goal is to build a tool, called REACT, which will make it easier for parents of children with autism to create and share these modules.
Methods:
REACT takes parents through the steps of authoring an instructional module that describes the process of completing an everyday task (e.g. going to a movie with a friend). Throughout the authoring process, the parent is provided with suggestions for information to include in their module. Last year, we described the process we used to generate a knowledge base that included the steps a person can take to complete a particular everyday task, obstacles that can arise and potential solutions to those obstacles. This knowledge base enables REACT to provide the author with the suggestions.
Once a prototype of the tool was developed, we conducted a study in which 9 parents of children with autism (ages 8-16) were asked to create modules for their child that described the process of going to a fast food restaurant for lunch. The parents were asked to create 2 modules each that described the situation, introduced an obstacle, and presented the child with 3 solutions for overcoming the obstacle. For one module, the parents were asked to create the module on their own without any help. In the other module, the parents were given suggestions from the knowledge base throughout the authoring process. Five parents created the module on their own first, and 4 parents were given the suggestions first.
Results:
We found that the parents’ stories varied greatly. Factors such as their experience writing social stories, and the educational approaches used with their child (Floortime, ABA, etc) greatly impacted the quality of the parents’ stories. We also found that while the suggestions did give the parents ideas for information to include in the story, the parents sometimes felt compelled to use suggestions, even when they had better ideas for what to write, and to use the suggestions as is, despite the fact that they knew that they could modify the wording.
Conclusions:
The initial prototype of REACT showed great promise for enabling parents to quickly create complex and customized social skills instructional modules for their child. There is room for improvement; however, in the way the suggestions are presented to the parents.