Factor structure of the VABS-3 Comprehensive Parent/Caregiver form in autistic individuals: Poor fit of three-factor and unidimensional models

Abstract

The commonly used Vineland Adaptive Behavior Scales (VABS-3) divides adaptive behavior into three domains comprising three subdomains. The validity of this three-factor structure has not been explored in autistic samples, which are often heterogeneous with respect to language and IQ. Furthermore, although there are two comparable forms, Comprehensive Interview Form and Comprehensive Parent/Caregiver Form (a questionnaire), the original validation was based on interview data. Considering the widespread use of the VABS-3 in autism research, and the increased feasibility of online-administered questionnaires, it is necessary to establish the validity of the Parent/Caregiver form in autistic individuals across a range of abilities. This study aimed to investigate the measurement invariance of the VABS-3 Parent/Caregiver form between a minimally verbal group and verbal group of autistic people; however, poor overall fit of the three-factor structure precluded invariance analyses. Subsequent analyses suggested poor fit in both language and age groups, as well as a unidimensional model. The results of the current study suggest that neither the three-factor or unidimensional model fit the VABS-3 Parent/Caregiver Form, thereby cautioning against interpretation of domain or overall adaptive behavior composite scores in autistic individuals and further encouraging careful consideration of administration format. Lay Abstract Adaptive behavior is a broad set of skills needed to function in everyday life. The Vineland Adaptive Behavior Scales (VABS-3) is commonly used to measure adaptive behavior. It divides adaptive behavior into three domains, Communication, Daily Living Skills, and Socialization, each of which are split into subdomains. Analyses of this three-part structure of the first version of VABS used the instrument as an interview, but now it is done as a questionnaire as well. The structure has not been well supported in samples of autistic people, who often have different strengths and challenges in adaptive behavior compared with non-autistic people. Because adaptive behavior is an important concept in autism research and online-administered questionnaires are increasingly common, it is important to ensure the structure of the VABS-3 Comprehensive Parent/Caregiver Form (VABS-3:CPCF; a questionnaire) works well for autistic individuals across a range of abilities. This study aimed to investigate whether VABS-3:CPCF measures adaptive behavior similarly in verbal and minimally verbal autistic people. However, the data didn’t fit the structure in the first step of the analysis, so this could not be investigated. The next analyses also found the three-domain structure didn’t fit in different age and language groups. In addition, the data didn’t fit a structure combining all the domains into 1 (unidimensional). These results suggest that neither the three-factor or unidimensional structure fit the VABS-3:CPCF, cautioning against interpretation of domain or overall adaptive behavior composite scores in autistic individuals and further encouraging careful consideration of administration format.