The Development and Validation of the Dutch Contextual Assessment of Social Skills (CASS): An Independent Observational Outcome Measure of Social Skills in Adolescents with Autism Spectrum Disorder

Background: Social skills interventions are considered to be efficacious if social skills are improved as a result of an intervention. Nevertheless, the objective assessment of social skills is hindered by a lack of sensitive and validated measures. To measure the change in social skills after an intervention, questionnaires reported by parents, clinicians and/ or teachers are commonly used. Yet, observations are the most ecologically valid method of assessing improvements of social skills after an intervention. For this purpose, The Program for the Educational and Enrichment of Relational Skills (PEERS) was developed for adolescents, in order to teach them the age-appropriate skills needed to participate in society (Laugeson, Frankel, Mogil, & Dillon, 2009). It is an evidence-based intervention for adolescents with ASD that taught ecologically valid social skills techniques.

Objectives: The current study aims to describe the development and psychometric evaluation of the Dutch Contextual Assessment of Social Skills (CASS), an observational outcome measure of social skills for adolescents with Autism Spectrum Disorder (ASD) (Ratto et al., 2011).

Methods: 34 adolescents (M = 14.68, SD = 1.41, 71% boys) with ASD performed the CASS before and after a social skills intervention (i.e. PEERS or the active control condition). Each adolescent completed a 3-minutes conversation with a confederate. The conversation was prompt as a natural introduction between two-unfamiliar, similarly ages, and opposite sex peers who meet for the first time. The adolescent and the confederate completed a brief questionnaire about the conversation (Conversation Rating Scale).

Results: Results indicated sufficient psychometric properties. The Dutch CASS has a high level of internal consistency (Cronbach's α coefficients = 0.84). Data supported the convergent validity (i.e. significant correlated with the Social Skills Improvement System (SSiS). The Dutch CASS did not significantly correlate with the autistic mannerism subscale from Social Responsiveness Scale (SRS), thus proved the divergent validity. Based on scorings made by raters who were kept blind to the time points, reliable change index were computed to assess the change in social skills. With regard to the content validity, only the learning objectives of the first two meetings of PEERS about conversational skills relatively matched with rating domains of the CASS. Due to this underrepresentation, we found an existing observational measure (TOPICC) that covers some of the other learning objectives of PEERS. TOPICC covers 22% of the learning objectives of PEERS about conversational skills, meanwhile CASS is 45%. Unfortunately, 33% of the learning objectives of PEERS was not covered by CASS or TOPICC.

Conclusions: Recommendations are made to improve the psychometric properties and content validity of the Dutch CASS.