Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Identifying the post-pandemic determinants of low performing students in Latin America through Interpretable Machine Learning methods

Marcos Delprato

TL;DR

It is found that a student with the highest probability of being a not achiever speaks a minority language and had repeated, has no digital devices at home, comes from a poor family and works for payment half of the week, and the school the student attends has wide disadvantages such as bad school climate and poor teaching quality.

Abstract

Introduction. The high prevalence of students not achieving basic learning competencies in Latin America (LAC) is concerning, even more so considering the region's deep structural inequalities and the larger post-pandemic learning losses. Within this scenario, the paper aims to contribute to the identification of the determinants of bottom and low performers (below level 2). Methodology. Based on 2022 data from the Programme for International Student Assessment (PISA) for 10 LAC countries, and using a stacking model integrating binary classification models as well as by applying Shapley Additive Explanations (SHAP) analysis for interpretability, we identify critical factors impacting on the student performance across low performers groups. Results. We find that a student with the highest probability of being a not achiever speaks a minority language and had repeated, has no digital devices at home, comes from a poor family and works for payment half of the week, and the school the student attends has wide disadvantages such as bad school climate, weak Information and Communication Technology (ICT) infrastructure and poor teaching quality (only a third of teachers being certified). For countries' estimates, we find quite homogeneous patterns regarding the contribution of top ranked factors, with repetition at primary, household wealth, and educational ICT inputs being top ten ranked covariates in at least 8 out of the 10 total countries. Discussions. The paper findings contribute to the broad literature on strategies to identify and to target those most left behind in Latin American education systems.

Identifying the post-pandemic determinants of low performing students in Latin America through Interpretable Machine Learning methods

TL;DR

It is found that a student with the highest probability of being a not achiever speaks a minority language and had repeated, has no digital devices at home, comes from a poor family and works for payment half of the week, and the school the student attends has wide disadvantages such as bad school climate and poor teaching quality.

Abstract

Introduction. The high prevalence of students not achieving basic learning competencies in Latin America (LAC) is concerning, even more so considering the region's deep structural inequalities and the larger post-pandemic learning losses. Within this scenario, the paper aims to contribute to the identification of the determinants of bottom and low performers (below level 2). Methodology. Based on 2022 data from the Programme for International Student Assessment (PISA) for 10 LAC countries, and using a stacking model integrating binary classification models as well as by applying Shapley Additive Explanations (SHAP) analysis for interpretability, we identify critical factors impacting on the student performance across low performers groups. Results. We find that a student with the highest probability of being a not achiever speaks a minority language and had repeated, has no digital devices at home, comes from a poor family and works for payment half of the week, and the school the student attends has wide disadvantages such as bad school climate, weak Information and Communication Technology (ICT) infrastructure and poor teaching quality (only a third of teachers being certified). For countries' estimates, we find quite homogeneous patterns regarding the contribution of top ranked factors, with repetition at primary, household wealth, and educational ICT inputs being top ten ranked covariates in at least 8 out of the 10 total countries. Discussions. The paper findings contribute to the broad literature on strategies to identify and to target those most left behind in Latin American education systems.

Paper Structure

This paper contains 31 sections, 16 equations, 13 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Machine learning in educational research
  4. Review of machine learning studies– learning surveys and students at risk
  5. Data
  6. Data coding– rationale
  7. Summary statistics
  8. Students and family covariates
  9. School covariates
  10. Methods
  11. Binary classification models
  12. Logistic regression
  13. Logistic Lasso with CV penalty
  14. Random forest
  15. Gradient boosting
  16. ...and 16 more sections

Figures (13)

  • Figure 1: Analytical steps
  • Figure 2: Area under the ROC curve, different ML models
  • Figure 3: Model: level 0 versus level 1 (bottom performing group). Beeswarm dot plot and heatmap of individual SHAP values for each student$/$famly and school covariate across observations
  • Figure 4: Model: level 0 versus level 1 (bottom performing group). Student$/$famliy and school covariates individual SHAP values for students with the lowest and highest average SHAP values contributions
  • Figure 5: Model: level 1 versus level 2 (low performing group). Student$/$famliy and school covariates individual SHAP values for students with the lowest and highest average SHAP values contributions
  • ...and 8 more figures