Evaluating Autism Knowledge Through Pre- and Post-Training Surveys: A Psychometric and Statistical Approach

Abstract

This project focused on analyzing the psychometric properties and training impact of the Autism Knowledge Survey (AKS) across over 18 datasets. Tasks included data cleaning, merging pre- and post-training survey responses, and conducting statistical analyses in REDCap, Excel, and R Studio. Psychometric evaluation revealed high internal consistency (Cronbach’s alpha = 0.836) and identified five key factors via Principal Component Analysis: diagnosis, genetics, autism across the lifespan, social challenges, and interventions. Ordinal regression showed that experience had no significant effect on knowledge in three areas, while paired t-tests revealed limited improvement in five specific topics post-training. These insights underscore the need for targeted training content in areas such as early intervention, autism genetics, and social relatedness. Results support the refinement of training curricula and demonstrate how statistical modeling can inform future ASD education strategies.