An avoidance trap? Within-person dynamics between social interaction anxiety and the tendency to avoid physical...

Physical activity and sport settings may become socially threatening when students anticipate negative evaluation of their bodies, competence, or behavior. Yet it remains unclear whether social interaction anxiety precedes exercise-related avoidance and whether such avoidance, in turn, precedes later anxiety. This study examined reciprocal longitudinal associations at both the between- and within-person levels between social interaction anxiety (SIA) and the tendency to avoid physical activity and sport (TAPAS), assessed with a 10-item self-report scale. A baseline cohort of 1,676 university students aged 18 years or older from nine universities in Sichuan Province, China, completed four assessments approximately 12 weeks apart during the 2024-2025 academic year. A random-intercept cross-lagged panel model separated stable between-person differences from time-specific within-person deviations, and longitudinal scalar invariance was supported for both measures. At the between-person level, higher average SIA was associated with higher average TAPAS (r = 0.381, p < 0.001). At the within-person level, elevations in SIA relative to students' usual levels predicted subsequent elevations in TAPAS (βs = 0.121-0.127), and elevations in TAPAS predicted subsequent elevations in SIA (βs = 0.126-0.133); all cross-lagged effects were significant at p < 0.001. Both constructs also showed significant within-person temporal continuity. The reciprocal pattern was stable across the three approximately 12-week intervals and remained substantively unchanged after adjustment for baseline sex, year of study, pre-university residence, and body mass index category. The findings are consistent with a reciprocal avoidance process across an academic year and suggest that campus mental-health and physical-activity initiatives may benefit from jointly addressing social-evaluative concerns and barriers embedded in exercise settings. Because the study was observational and relied on self-report measures, the associations should not be interpreted causally.