The Factor Structure of Restricted and Repetitive Behaviors in Young Children with and without Autism Spectrum Disorder

Background: Given the heterogeneity of restricted and repetitive behaviors (RRBs) and their relevance to DSM-5 diagnostic criteria for autism spectrum disorder (ASD), empirical evidence informing which RRB factors are prevalent in young children is critical in improving early detection. While it is clear that RRBs are present at a young age (Elison et al., 2014; Guthrie et al., 2013), RRB symptoms do not discriminate young children with ASD as effectively as social communication (SC) symptoms in early screening instruments (Berument et al., 1999; Rowberry et al., 2015), and clinicians do not endorse the presence of RRBs as frequently as SC symptoms (Stone et al., 1999). Children with developmental delays (DD) and typical development (TD) also frequently demonstrate RRBs (Evans et al., 1997; Bodfish et al., 2000), leading to further challenges in determining whether certain behaviors are symptomatic of ASD at an early age.

Objectives: (1) To determine the factor structure of RRBs in two through five year olds using the Repetitive Behavior Scale–Revised (RBS-R), and (2) to examine whether the factor structure of RRBs is consistent across children with ASD, DD, and TD.

Methods: There were 363 participants selected from a larger sample of children recruited for the FIRST WORDS® Project at Florida State University, a longitudinal, prospective study investigating early detection of ASD and other communication disorders. Children were included if an RBS-R had been completed between ages 2 and 5 (mean=34.81 months, SD=7.19). The RBS-R is an empirically derived clinical rating scale for measuring the presence and severity of a range of RRBs that are characteristic of ASD.

Outcome measures of developmental level (Mullen Scales of Early Learning), adaptive behavior (Vineland Adaptive Behavior Scales, Second Edition) and autism symptoms (Autism Diagnostic Observation Schedule) were obtained to determine a best estimate diagnosis. The sample included 190 children with ASD, 99 with DD, and 74 with TD.

Confirmatory Factor Analysis (CFA) was used to evaluate previously supported three-, five-, and six-factor models using the RBS-R in toddlers (Mirenda et al., 2010). Because models were nested within each other, a difference test was computed to determine whether the most parsimonious model was preferable. Further analyses utilizing Measurement Invariance techniques will also be reported in order to examine diagnostic group differences in RRB factor structure.

Results: Results suggested that the three-factor model, consisting of stereotyped/restricted, compulsive/ritualistic/sameness, and self-injurious behaviors provided preferable fit (χ²/df=1.36, CFI=.98; TLI=.98; RMSEA=.03). Similarly, the five- and six-factor models provided good fit, but were not as parsimonious (five-factor: χ²/df=1.30, CFI=.98; TLI=.98; RMSEA=.03; six-factor: χ²/df=1.31, CFI=.98; TLI=.98; RMSEA=.03).

Conclusions: The RBS-R allows for thorough assessment of a range of RRBs in toddlers and young children. Stereotyped and restricted behaviors emerge as a single factor, a second consists of compulsive, ritualistic, and sameness behaviors, and self-injurious behaviors fall independently on a third factor. Evidence-based support regarding the structure of RRBs in children across diagnostic groups is critically important in understanding early symptomology and improving early detection of ASD.