Longitudinal dynamics between other-gender peer contact quality and gender-stereotypical trait norms in coeducational physical education: a four-wave random-intercept cross-lagged panel study.

Coeducational physical education (PE) may reproduce or challenge gender norms depending on the quality of students' interactions with other-gender peers. This four-wave longitudinal study examined between-person and within-person prospective associations between other-gender peer contact quality in PE (CQ) and endorsement of gender-stereotypical trait norms (GST). Participants were 1,376 students from 20 primary and secondary schools in Chengdu, China, assessed from March 2025 to January 2026. After longitudinal measurement invariance was established, a random-intercept cross-lagged panel model separated stable between-person differences from time-specific within-person deviations. The CQ and GST random intercepts were negatively correlated (r = -0.287, p < 0.001). In the freely estimated model, higher-than-usual CQ was associated with lower-than-usual GST at the next wave (β = -0.197, -0.133, and -0.151; all p ≤ 0.001), whereas the reverse paths were nonsignificant (β = 0.013, 0.001, and 0.030). Equality constraints across intervals were compatible with the data, scaled Δχ2(8) = 1.869, p = 0.985. Direct Wald tests showed that the CQ-to-GST paths were more negative than the corresponding reverse paths, χ2(3) = 44.574, p < 0.001. The pattern remained in a 44-indicator multiple-indicator RI-CLPM and after removal of the CQ equality item. Full metric and scalar invariance held across boys and girls, and no statistically detectable sex-group differences emerged in either cross-lagged direction. The findings indicate a temporally asymmetric prospective association in the pooled sample. Causal, developmental-stage, and classroom-level interpretations remain limited by the observational design and unavailable age, grade, and clustering identifiers.