Profiling low-proficiency science students in the Philippines using machine learning.

Area of Science:

• Educational Psychology
• Science Education
• Data Science

Background:

• Filipino students consistently demonstrate low performance in international science literacy assessments.
• The Programme for International Student Assessment (PISA) 2018 results placed Filipino learners second to last among 78 participating countries in science literacy.
• Identifying factors contributing to low achievement is crucial for effective science education reform in the Philippines.

Purpose of the Study:

• To utilize machine learning to analyze student questionnaire data from PISA 2018.
• To identify the key factors associated with the poorest-performing Filipino students in science literacy.
• To inform targeted interventions and policy recommendations for science education reform.

Main Methods:

• Application of machine learning approaches, specifically the random forest classifier, to PISA 2018 student data.
• Utilizing Shapley Additive Explanations (SHAP) to determine the importance of various predictive variables.
• Analysis focused on identifying students vulnerable to very low science achievement.

Main Results:

• The random forest classifier model demonstrated high accuracy and precision in identifying low-performing students.
• Fifteen variables were identified as most significant in predicting low science proficiency.
• Key factors included metacognitive awareness of reading strategies, social experiences, aspirations, pride in achievement, and family/home factors (parental characteristics, ICT access).

Conclusions:

• Personal and contextual factors significantly influence science literacy achievement, extending beyond traditional instructional and curricular elements.
• The findings underscore the need for a holistic approach to science education reform in the Philippines.
• Implications for policy and program development are suggested to address identified contributing factors.