How Social Others Form First Impressions of Adults with Autism Spectrum Disorder

Background: First impressions are rapidly formed and assert long-term influences on social preferences and behavior. Whether first impressions of adults with autism spectrum disorder (ASD) differ from those made of typically-developing (TD) adults, and the factors underlying their formation, is not well understood. Highlighting aspects of social presentation in ASD that affect evaluation by potential social partners may ultimately be used to mitigate social interaction impairments in ASD by educating social others about behavioral differences in ASD and inform ASD adults about these perceived differences. 

Objectives: The ongoing project obtained first impressions from TD adults observing real-world behavior of adults with ASD, as well as TD comparison participants. We examined how specific information channels (i.e., visual cues, audio cues, and speech content) drove impression formation of adults with ASD across a range of personality traits and whether these impressions were associated with intentions to socially engage with the observed individual.

Methods: 147 TD observers, blind to the diagnostic status, rated twenty ASD adults and twenty TD comparison adults matched on gender, age, ethnicity and IQ engaging in a social presentation task. The first 10s of dialogue produced by participants was extracted to produce stimuli depicting five different presentation modalities: (a)audio content only, (b)video content only, (c)a written transcript of speech content, (d)a static frame, and (e)the full 10s clip with audio-visual. Observers gave first impressions across six character traits (attractiveness, intelligence, trustworthiness, likability, awkwardness, dominance) and indicated their intent to socially engage with the participant. 

Results: A 6(traits) X 5(modality) X 2(group) mixed-model ANOVA revealed the ASD group was rated as more awkward, less attractive, less dominant, and less likeable than the TD group (all p’s<.01), with no difference on trustworthiness or intelligence. A three-way interaction between group, modality, and trait (F(20,760)=3.06, p<.001) showed that group differences largely persisted across the audio, video, static frame, and audio-visual modalities, with no group differences found for the transcript condition except for on awkwardness. Likeability was the strongest correlate with intent to socially engage for both groups (r=.88, ASD; r=.89, TD); however, awkwardness more strongly related to observers intent to socially engage with the ASD group (r=.560) compared to the TD group (r=.242), a significant difference using an r-to-z transformation (p=.004). Future factor analytic approaches (i.e., Correspondence Analysis) will examine how first impression ratings of each trait is distributed across the specific presentation modalities, and whether patterned profiles distinguish the ASD and TD groups. 

Conclusions:  Adults with ASD are generally perceived less favorably compared to TD matched comparison adults, except for ratings of intelligence and trustworthiness. These patterns were consistent across audio, video, static image, and audio-visual conditions, but not for the transcript of the speech content. This failure to find differences in the transcript conditions suggests that the content ASD adults choose to discuss does not seem to drive initial impression formation, rather all audio/visual social presentation cues appear to be similarly influential. Findings are pertinent for exploring situations where ASD adults are interacting with a novel social partner (e.g., job interview).