A Multi-level Analysis of Factors Associated with Student Performance: A Machine Learning Approach to the SAEB Microdata

TL;DR

This paper tackles the Brazil basic-education policy problem by developing a multi-level machine learning framework using SAEB microdata to predict student proficiency. It integrates four data sources (student socioeconomic characteristics, teacher profiles, school indicators, and principal management) and compares four ensemble models, with Random Forest achieving $90.17\%$ accuracy and $AUC=0.9669$. Using SHAP for Explainable AI, it shows that the school's average socioeconomic level is the dominant predictor, highlighting systemic factors over isolated individual traits. The findings offer a data-driven, interpretable tool to guide policy and school leadership in promoting educational equity, while acknowledging limitations and proposing future longitudinal and data-extension work.

Abstract

Identifying the factors that influence student performance in basic education is a central challenge for formulating effective public policies in Brazil. This study introduces a multi-level machine learning approach to classify the proficiency of 9th-grade and high school students using microdata from the System of Assessment of Basic Education (SAEB). Our model uniquely integrates four data sources: student socioeconomic characteristics, teacher professional profiles, school indicators, and principal management profiles. A comparative analysis of four ensemble algorithms confirmed the superiority of a Random Forest model, which achieved 90.2% accuracy and an Area Under the Curve (AUC) of 96.7%. To move beyond prediction, we applied Explainable AI (XAI) using SHAP, which revealed that the school's average socioeconomic level is the most dominant predictor, demonstrating that systemic factors have a greater impact than individual characteristics in isolation. The primary conclusion is that academic performance is a systemic phenomenon deeply tied to the school's ecosystem. This study provides a data-driven, interpretable tool to inform policies aimed at promoting educational equity by addressing disparities between schools.

Figures (4)

• Figure 1: The multi-level data preprocessing pipeline, from raw data acquisition to the final processed dataset.
• Figure 2: Confusion Matrix for the final Random Forest Classifier.
• Figure 3: Comparative ROC Curve for all tested models.
• Figure 4: SHAP Summary illustrating the global importance and impact of the top predictor variables (based on a 20,000 record sample).