Social anxiety disorder (SAD) is a common psychological disorder associated with broad interpersonal impairment. Most previous studies have examined nonverbal behavior in SAD using human coders. However, one recent study utilized a machine-based analysis of nonverbal behavior and dyadic synchrony in SAD (Asher, Kauffmann, & Aderka, 2020). In the present study, we compared human and computer assessments of nonverbal behavior in social anxiety to enhance our knowledge about their commonalities and unique differences in capturing nonverbal behavior in the context of SAD. Specifically, the present study included 152 individuals: 38 individuals diagnosed with SAD and 114 individuals without SAD. Participants formed 76 opposite-sex interaction dyads comprising either two individuals without SAD (n = 39 control dyads) or one individual with SAD and one individual without SAD (n = 37 SAD dyads). All participants underwent a getting-acquainted task and were videotaped during the conversation. Half of the interactions were small talk interactions and half were closeness-generating interactions that required significant self-disclosure. We found that both types of coding were associated with self-reported social anxiety but that machine-based coding was superior in capturing social anxiety in closeness-generating contexts. Implications for research on nonverbal behavior in SAD are discussed.
Keywords: Interaction; Nonverbal synchrony; Observer coding; Social anxiety disorder.
Copyright © 2022 Elsevier Ltd. All rights reserved.