Naive Thin-Slice Impressions Reveal Evidence of the Female Camouflage Effect in ASD

Background: Individuals with autism spectrum disorder (ASD) are characterized by social communication challenges and repetitive behaviors that are quickly detected by experts (de Marchena & Miller, 2017). Recent research suggests that even naïve non-experts are capable of forming accurate impressions about a variety of human dimensions using only narrow windows of experience called “thin-slices” (Slepian, Bogart, & Ambady, 2014). Growing recognition of sex differences in a variety of observable behaviors in ASD (van Ommren et al., 2017), combined with research showing that some autistic girls may “camouflage” their symptoms (Lai et al., 2017), suggests it may be more difficult for naïve interlocutors to detect ASD symptoms in females. In this study, we explore “thin-slice” ratings of interaction quality in females and males with and without ASD.

Objectives: Compare “thin-slice” ratings of matched females and males with ASD, and typically developing (TD) participants. Based on prior research, we hypothesized that autistic girls would be judged as less different than TD girls, whereas ASD boys would be rated as more drastically different than TD boys.

Methods: Seventy-three participants with ASD (N=35, 14 females) or TD (N=38, 13 females) completed a 5-minute “get-to-know-you” conversation with a novel confederate (N=19, 16 females). Diagnostic groups were matched on IQ (ASD: 107; TD: 108; t=-.43, p=.67), but the TD group was younger (ASD: 11.63 years; TD: 9.99 years; t=2.60, p=.01). Autistic girls and boys did not differ significantly on overall ADOS calibrated severity scores (t=.91, p=.37, Cohen’s d=.31) or scores on the Social Communication Questionnaire (t=.17, p=.86, Cohen’s d=.06). After each conversation, confederates completed a modified version of the Conversation Rating Scale (CRS-ext; Ratto et al., 2011). The CRS-ext includes 6 questions indexing conversational interest, warmth, flow, boredom, distance, and appropriate eye contact on a 1 to 7 Likert scale (range=6-42; higher scores indicate better conversation). A linear mixed effects model included age, IQ, sex, diagnostic group, the interaction between sex and diagnostic group, and a random effect to account for repeated confederates.

Results: There was a trend toward a significant interactive effect of sex and diagnosis on CRS-ext scores (estimate: -5.74, t=-1.85, p=.07; Figure). Comparing estimated marginal means (Tukey correction) revealed the interaction was driven by a significant sex difference in ASD participants, such that autistic girls had significantly higher CRS-ext scores than autistic boys (t=3.49, p=.005; Cohen’s d=.87). TD girls and boys also had significantly higher scores than ASD boys, but there was no significant difference between ASD girls and TD girls (t=1.48, p=.46; Cohen’s d=.36), TD girls and boys (t=.66, p=.91, Cohen’s d=.16), or autistic girls and TD boys (t=.95, p=.77; Cohen’s d=.23).

Conclusions: “Thin-slice” ratings of naturalistic conversations hold promise as a low-cost metric to gauge the impression that individuals with ASD make on naïve communication partners in everyday life. However, this method appears to be more effective for the male behavioral phenotype. These findings add to the literature showing that autistic girls are perceived as “less different” than autistic boys, which may reflect partially successful camouflaging.