3 - (IOP 25) Social Anxiety and Romantic Relationship Dynamics: Continuous Time Analysis of Dominance and Warmth During Conflict and Social Support Tasks

Social anxiety disorder (SAD) is associated with impairment in close relationship functioning, including romantic relationships. Although interpersonal theories point to dimensions of dominance and warmth as relevant factors, the momentary interpersonal dynamics that may be contributing to relationship problems remain poorly understood.The Continuous Assessment of Interpersonal Dynamics (CAID; Sadler et al., 2009) system makes it possible to examine such dynamics by measuring moment-to-moment changes in dominance and warmth. Few studies have utilized individually precise (i.e., idiographic) and prospective models to examine interpersonal dynamics relevant to mental disorders, such as social anxiety disorder (SAD). Such models are critical in the context of SAD, where we expect heterogeneity in interpersonal problems (Kachin et al., 2001) and the question of who drives the effect (e.g., person with SAD or conversation partner) is essential. The present study combines CAID data and multilevel residual dynamic structural equation modeling (ML-RDSEM) to examine contemporaneous and prospective interpersonal dynamics between primary participants with (n = 30) and without (n = 29) social anxiety and their romantic partners during conflict and social support tasks. Previous CAID studies have noted that the strength and interpretation of prospective effects may differ at alternate lag lengths (Sadler et al., 2015; Thomas et al., 2014), and called for further examination of effects at different time interviews. Fortunately, a recently developed continuous time version of ML-RDSEM makes it possible to identify the lag interval that would maximize the prospective effects by including the timing of observations as a continuous variable. In a previous study (Levitan & Rodebaugh, 2025), we applied this method to conflict task data from this sample of romantic dyads, and results suggested that longer interactions (i.e., more CAID ratings per dyad) are needed to examine the strength of cross-construct effects across the larger time intervals at which they are expected to be strongest. Therefore, we will present continuous time models examining the prospective effects of dominance and warmth between participants with and without SAD and their romantic partners across three consecutive interaction tasks. This presentation will highlight the strengths and challenges involved in this method.