Using machine learning to predict UK and Japanese secondary students' life satisfaction in PISA 2018.

Area of Science:

• Educational Psychology
• Computational Social Science
• Student Well-being Research

Background:

• Student life satisfaction is crucial for academic success and long-term health.
• Previous research on life satisfaction has rarely utilized machine learning (ML) methods.
• This study addresses the gap by applying ML to predict student life satisfaction.

Purpose of the Study:

• To predict secondary students' life satisfaction using individual-level variables via ML algorithms.
• To identify key predictors influencing students' life satisfaction.
• To explore the utility of ML in understanding student well-being.

Main Methods:

• Utilized supervised machine learning models: Random Forest (RF) and K-Nearest Neighbors (KNN).
• Data sourced from the PISA 2018 dataset, including UK and Japan samples.
• Compared model performance and identified significant predictors.

Main Results:

• Both RF and KNN models showed better predictive performance on UK data compared to Japan data.
• The Random Forest model demonstrated superior accuracy in predicting student life satisfaction over the KNN model.
• Key predictors identified include meaning in life, student competition, teacher support, bullying, and ICT resources.

Conclusions:

• Confirms the multi-dimensional nature of life satisfaction and pinpoints critical influencing factors.
• Pioneers the use of ML techniques to investigate predictors of student life satisfaction.
• Provides a practical framework for interventions aimed at improving secondary students' life satisfaction.