Autism spectrum disorder (ASD) is a neurodevelopmental disorder recognized in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; APA, 2013) by deficits in social communication and the presence of restricted and repetitive behaviors and interests. Despite the clinical significance of restricted and repetitive behaviors in the presence of ASD, research regarding the assessment and treatment of this domain is minimal, when compared to that of social impairments present within the disorder. The Repetitive Behavior Scale-Revised (RBS-R; Bodfish et al., 2000) is a rating scale designed to capture the broad range and scope of repetitive behaviors in individuals with ASD. Among the few measures available for this complex construct, researchers have frequently used the 43-item RBS-R as an outcome measure in intervention studies. In some instances, practitioners have used the RBS-R as a progress-monitoring tool (Boyd et al., 2011; Chowdhury et al., 2010; Esbensen et al., 2009; Rojahn et al., 2013; Schertz et al., 2016; Stratis & Lecavalier, 2013; Wolff et al., 2016). Despite its relative popularity, RBS-R validity research has yet to yield a consensus regarding the RBS-R factor structure. Such consensus is needed to clarify the number of subscales and their constructs, as well as establish item-to-subscale assignments, and thereby inform the scoring of the instrument. Previous factor analytic studies (Bishop et al., 2013; Bodfish et al., 2000; Lam & Aman, 2007; Mirenda et al., 2010; Russell et al., 2019; Hooker et al., 2019; Sturm et al., 2022), were conducted using caregiver ratings on the original English version of the RBS-R in ASD samples. Results yielded a variety of factor solutions (i.e., solutions consisting of four, five, or six factors). Though variable and inconsistent, such analyses have produced somewhat more relative support for a five-factor solution among caregiver ratings of repetitive behavior (Bishop et al., 2013; Hooker et al. 2019; Lam & Aman, 2007; Mirenda et al., 2010; Sturm et al., 2022). Additional factor analytic studies conducted using non-English versions of the RBS-R in other cultural contexts (Georgiades et al., 2010; He et al., 2019; Kästel et al., 2020), using caregiver ratings of ASD samples also yielded inconsistencies in the number of factors present (i.e., two, four, and six factors). Only one published factor analytic study involved the use of school staff ratings of the RBS-R in an ASD sample. This study (Martínez-González & Piqueras, 2017) utilized the Spanish version of the instrument in Spain and the authors retained a six-factor solution. This has left questions about the range of factor solutions examined within and across studies, as well as possible differences in factor structure due to rater type. The lack of consensus, in addition to the lack of school raters using the English version of the RBS-R, warrants independent factor analyses, to determine the most appropriate, useful, and valid factor structure of the RBS-R among school staff raters. The present dissertation involves two independent studies. Study one used exploratory factor analysis of the RBS-R items with an ASD sample (N = 234 cases, ages ranging from three to 21 years) rated by special education staff. The EFA resulted in a five-factor solution that demonstrated some similarities to existing factor models, though presented clear differences not previously observed in the literature. Study two involved applying confirmatory factor analysis (CFA) in a second independent ASD validation sample (N = 233 cases) from the same agency used in study one. CFA procedures were used to test the absolute and relative fit of the RBS-R factor solution retained from study one across five fit indices and to compare the fit of that model to others available in the literature. The fit of the factor model from study one was directly compared to the fit of existing RBS-R models found in ASD samples. Results from the CFA suggested that the five-factor model from study one adequately fit the sample data. Results from the AIC and BIC indices indicated that the five-factor model from study one was the best fitting model compared to the other existing models. Findings from studies one and two underscore the possibility that the current six-factor version of the RBS-R is potentially not the most viable model for the ASD population when using ratings made by teachers or other special education staff. Findings also suggest the presence of a new, unique factor resulting from dividing the items from the original RBS-R self-injurious behavior subscale. Further, findings leave room for future studies to continue to consider the factor structure and fit of models of the RBS-R based on teacher or other educational staff ratings in samples of those with ASD.
Read
