Abstract
BACKGROUND: School bullying represents a critical global public health issue among adolescents. Although existing evidence suggests physical activity (PA) may reduce bullying risk, longitudinal data on bidirectional associations and gender-specific variations remain limited. This study employed a cross-lagged panel model (CLPM) to investigate the temporal dynamics between PA and school bullying (SB) and examine gender-moderating effects. METHODS: A cohort of 577 middle school students (294 boys, 283 girls; mean age = 14.31 ± 2.1 years) from seven schools across Wuhan, Shijiazhuang, and Chengdu completed three-wave longitudinal assessments over 9 months (September 2024-March 2025). Validated instruments included the School Bullying Scale (SBS) and Physical Activity Rating Scale (PARS). The CLPM analysis evaluated bidirectional predictive pathways, with gender-stratified multi-group comparisons. RESULTS: Significant bidirectional negative associations emerged: (1) PA at T1/T2 predicted reduced SB at T2/T3 (β = -0.14 to -0.26, p < 0.001). (2) SB at T1/T2 predicted decreased PA at T2/T3 (β = -0.27 to -0.38, p < 0.001). (3) Gender significantly moderated these relationships, with PA conferring stronger protective effects against subsequent SB in males (β = -0.35 vs. -0.21 for PA→SB paths). Conversely, SB triggered earlier and more pronounced PA reductions in males (β = -0.42 vs. -0.29 for SB→PA paths). CONCLUSIONS: PA and SB demonstrate stable bidirectional negative associations in adolescents, with significant gender divergence. Males exhibit greater resilience to bullying through PA engagement but heightened vulnerability to PA reduction post-victimization. These findings underscore PA-based interventions as promising bullying mitigation strategies, necessitating gender-tailored implementation approaches.