Anxiety and Social Initiation in Young Children with Autism Spectrum Disorder

Background: Previous research indicates that children with autism (ASD) are less likely to initiate social interactions (Rivard et al., 2016) which may result in decreased opportunities for effective social development (Schietecatee et al., 2012). Many children with ASD also experience co-occurring anxiety, with recent studies reporting comorbidity rates as high as 40% (Kerns et al., 2016). Moreover, previous literature suggests that anxiety is associated with several social communication impairments (Taylor & Gotham, 2016), but the relation between anxiety and social initiation in school-aged children with ASD is not clear (Swain, Scarpa, White, & Laugeson, 2015).

Objectives: The purpose of this study was to examine the role of anxiety on social initiation behaviors in children with ASD compared to their TD peers. Anxiety was hypothesized to moderate the relation between developmental status and social initiation behaviors, such that children with autism and high anxiety would have the lowest level of social initiation behaviors compared to the TD sample.

Methods: Our sample included 110 children (ages 3:0 to 6:11), 71 (49 % female) TD children and 45 children with ASD (20% female). Video recordings of an 8-minute parent-child free-play task were coded using an adapted version of the Modified-Classroom Observation Schedule to Measure Intentional Communication (M-COSMIC; Clifford et al., 2010), from which percentages of child social initiations were obtained. Child anxiety ratings were collected using the Parent-Form Anxiety subscale of the Behavioral Assessment System for Children—Second Edition (BASC-2; Reynolds & Kamphaus, 2004).

Results: A moderated multiple regression analysis was conducted using the PROCESS macro for SPSS (Hayes, 2013). Child age, gender, and verbal ability were controlled for in the analysis. Findings included a significant main effect for developmental status on social initiation (B = -.225, CI₉₅ [-.423, -.026], p = .027), which supported our hypothesis that children with ASD would initiate less than TD children. There was a nonsignificant main effect for anxiety on social initiation (B = .001, CI₉₅ [-.002, .003], p = .558), contrary to our hypothesis that lower levels of anxiety would yield higher levels of social initiation. Finally, the interaction between developmental status and anxiety on initiation was nonsignificant (B = .002, p = .393), but trending in an unexpected direction. While the TD group demonstrated the highest levels of social initiation overall, children with ASD and high levels of anxiety demonstrated the highest levels of social initiation for the ASD group. Results suggest that the variables in our model accounted for 38% of the variance in initiations.

Conclusions: Consistent with previous research, children with autism initiated fewer social interactions with their parents than TD children (Humphrey & Symes, 2011; Rivard et al., 2016), but children with ASD in our study did not have higher levels of anxiety. Although nonsignificant, results suggest the frequency of social initiations in children with ASD tended to increase with higher child anxiety. Future research should investigate the function of anxiety in the social interactions of children with ASD, as moderate levels of anxiety may potentially be adaptive in young children with ASD.