Revealing the Key Factors behind Academic and Social-Emotional Resilience: A Machine Learning and IPTW Approach

• In Event: AERA Roundtable Session Wednesday 1:45 pm - Gold 3
  In Roundtable Session: Innovative Tools for Supporting Social Emotional Learning

Abstract

This study investigates key predictors of academic and socio-emotional resilience among socioeconomically disadvantaged students using machine learning and inverse probability treatment weighting (IPTW). Drawing on OECD’s Social and Emotional Skills Survey (N =54,729), we define resilience as high achievement or strong socio-emotional skills despite low SES. Predictive models identify key individual and contextual factors, including teacher–student relationships, school belonging, and behavioral engagement. Subsequent IPTW analyses revealed that positive school relationships had consistent causal effects on both types of resilience, while perceived pressure and behavioral problems reduced resilience likelihood. Results highlight the distinct and shared drivers of academic and emotional success under adversity, offering actionable insights for equity-oriented educational interventions.

Authors

• Siyuan Chen, University of California - Santa Barbara

• Xin Tang, Shanghai Jiao Tong University