农村儿童多维健康的同伴效应研究OACHSSCD

Based on the multidimensional health survey data of 7,591 rural children from Liaoning and Hebei provinces in 2024,this study develops a multidimensional health indicator system for rural children.Innovatively integrating the entropy weight method with the PAM clustering algorithm,an unsupervised machine learning approach,it measures a multidimen-sional health index and classifies children into health groups,and then examines the influence and underlying mechanisms of peer effects on rural children's multidimensional health.The results show that the average multidimensional health of peers exerts a significant positive spillover effect on rural children's multidimensional health,and that the effect strengthens as the proportion of high-health peers within the class increases.Mechanism analysis reveals that peer effects play an im-portant synergistic role along two key pathways,namely physical health-social adaptation and mental health-social adapta-tion.Moreover,the peer effect gradually weakens from the low to the high quantiles of the multidimensional health index,exhibiting a compensatory pattern.The effect is also stronger among girls,left-behind children without migration experi-ence,and those with both parents working away from home.These findings suggest three policy implications.First,promote integrated multidimensional health interventions and build a collaborative public health policy framework.Second,actively shape healthy peer environments to capitalize on the positive spillover effects of advantaged groups.Third,implement target-ed public health services to refine a stratified and categorized health support system.