Understanding Students' Subjective Well-Being: Combining Machine Learning and Classical Statistics

• In Event: Learning and Education

Abstract

Much of the existing educational research focused on students’ academic performance with comparatively fewer studies heeded students’ well-being. This study aimed to examine the relative importance of individual, microsystem, and mesosystem factors in predicting students’ well-being. Ecological system theory was used as the overarching framework. The data from the 2018 PISA, which included 6,037 Hong Kong students, were analyzed using machine learning and hierarchical linear modeling. Results demonstrated the overwhelming importance of individual and microsystem factors in understanding well-being. Specifically, results of machine learning revealed that school belonging, resilience, parental support, and fear of failure were the most important predictors of student well-being. Hierarchical linear modeling further confirmed the importance of these influential factors. Practical and theoretical implications were discussed.

Authors

• YI WANG
  Presenting Author

• Ronnel B. King, Chinese University of Hong Kong
  Non-Presenting Author

• Shing-On Leung, University of Macau
  Non-Presenting Author