1. Factors affecting students’ performance on national assessments of mathematics in Italy : a random forest approachDaniel Doz, Bor Bregant, 2024, original scientific article Abstract: This paper investigates the impact of teacher-assigned grades in Mathematics and Italian, students’ gender, and geographical macroregion on students’ performance in the Italian INVALSI mathematics assessment, using Random Forest analysis across grades 2, 5, 8, 10, and 13. Findings revealed that the two most influential factors are mathematics and Italian teacher-assigned grades, followed by gender. Boys consistently achieve higher INVALSI scores, while girls receive higher teacher-assigned grades. Performance disparities are observed among the five Italian geographic macroregions, with students from northern and central Italy performing better. Linguistic abilities and gender show varying significance across grades. The role of geographic macroregion is more pronounced in high school. Results were confirmed using Boosting Regression, validating the findings. This study highlights the significance of teacher-assigned grades, linguistic skills, gender, and geographic disparities in predicting students’ performance on the INVALSI mathematics test, showcasing the value of machine learning modelsin addressing educational equity.
Keywords: INVALSI, mathematics, national assessment, random forest, grades
Published in RUP: 22.01.2026; Views: 773; Downloads: 7
 Full text (2,27 MB)
This document has more files! More... |
2. Predicting Italian students’ mathematics outcomes : a decision tree regression analysisDaniel Doz, Darjo Felda, Mara Cotič, 2025, original scientific article Abstract: The present paper aims to investigate the factors that influence the achievements of Italian students on the National Mathematics Assessment INVALSI. The study is a quantitative non-experimental research and utilizes the Decision Tree Method (DTM), a data mining and machine learning approach, to analyze the relationships and interactions among the variables and their influence on students’ mathematics performance. The sample for the study consists of 15,344 grade-10 students who took the INVALSI test in the school year 2021/22. Findings show that school typology had the highest relative importance, followed by students’ school grades in mathematics, socioeconomic status, geographic macroregion, gender, age, and, finally, origin. Based on these results, policymakers and educators should prioritize interventions that enhance educational environments and individual academic proficiency, particularly focusing on school type, mathematics grades, and students’ ESCS, to improve student achievements and promote deeper learning.
Keywords: INVALSI, decision tree, cross-validation, mathematics
Published in RUP: 19.01.2026; Views: 634; Downloads: 12
Full text (1,12 MB)
This document has more files! More... |
3. |
4. |
