Weight control behavior typologies and multilevel determinants among young adults in a Chinese university: a socio-ecological perspective

Globally prevalent overweight/obesity and unhealthy weight control behaviors among young adults lack refined typologies and multi-level mechanism research. Using a socio-ecological framework, this study analyzed individual, interpersonal, and community-level factors and multi-path mechanisms among 9,870 university students (5,234 with weight reduction behaviors). Logistic regression and LCA were applied to the full sample (non-weight-control group as reference), and SEM to explore high-risk weight control mechanisms.LCA identified three latent weight-control patterns: diet and exercise-oriented pattern (49.0%), pharmacological weight-loss pattern (16.0%), and meal replacement-dependent pattern (35.0%). Logistic regression showed that weight perception bias (OR = 2.89, 95% CI 2.67–3.13), weight-related anxiety (OR = 2.52, 95% CI 2.343-2.723), female gender (OR = 0.70, 95% CI 0.609-0.811) were significant factors associated with weight control engagement, while peer support (OR = 0.91, 95% CI 0.856-0.958) and campus counseling services (OR = 0.84, 95% CI 0.753-0.936) served as protective factors. For SEM, the model showed acceptable fit given the large sample size (χ2/df = 19.975, RMSEA = 0.06, CFI = 0.937, TLI = 0.921). Path analyses revealed that social support had a significant total effect on weight control behaviors (β = 0.448, P = 0.003), with a direct effect of β = 0.42 (P = 0.003). Higher social support directly links to heterogeneous weight control behaviors. Thus, subtype-specific, multi-dimensional interventions are needed to curb unhealthy weight-loss practices and improve campus public health.