A Study on the Influencing Factors of Non-Cognitive Abilities of Junior High School Students Based on Hierarchical Linear Models

Background: Non-cognitive abilities are essential for adolescent development, yet research examining their multilevel determinants remains limited. While prior studies have identified individual or school-level factors independently, few have investigated how these levels interact to shape non-cognitive development during the critical junior high school period. Methods: Using data from 7,894 students in the China Education Panel Survey (CEPS), this study employed Hierarchical Linear Modeling (HLM) to simultaneously examine individual and school-level predictors. The analysis focused on three dimensions of non-cognitive abilities (resilience, cooperation, and emotional stability) while testing cross-level interactions between student characteristics and school environment factors. Results: Individual-level analysis revealed significant effects of gender (β=-0.043), appearance (β=0.160), exercise time (β=0.002), family background (β=0.061), and parental involvement (β=0.101). School-level factors showed school type (β=-0.071) and locations (β=-0.034) significantly predicted outcomes. Notably, school location moderated the effects of migration status (β=-0.076) and appearance (β=0.037), while school ranking enhanced exercise benefits (β=0.001). Conclusion: This study demonstrates the complex interplay between individual and school factors in shaping non-cognitive abilities. The findings suggest that educational interventions should adopt a multilevel approach, addressing both student characteristics and school environment improvements, particularly in resource allocation for rural schools and physical activity programs. These results provide both theoretical insights for developmental psychology and practical guidance for educational policy-making.