Prediction of Students Performance Based On School Dropout Risk Factors
Leogildo Freires, Federal University of Alagoas (Brazil)
Heitor Araújo, Federal University of Bahia (Brazil)
Júlio Costa, Federal University of Minas Gerais (Brazil)
Luan Filipy Freire Torres, Universidade Federal de Alagoas (Brazil)
Gabriel Macedo, Federal University of Alagoas (Brazil)
Ane Mayra Melo Silva, Federal University of Alagoas Center for Excellence in Social Technologies (NEES) (Brazil)
Abstract
School dropout has been widely recognized as a complex and multifactorial phenomenon, with significant implications for both individual development and educational systems. In recent years, advances in early warning systems have sought to anticipate this risk by identifying factors associated with school disengagement, particularly those of a relational nature, such as interactions among students, families, and schools [1]. Study aimed to investigate the predictive effect of school dropout risk factors on the performance of Brazilian students, using a synthetic sample of 10,000 cases, derived from a real dataset of 3,678 students from four states of Brazil. Synthetic data was generated using the Gaussian Copula technique, which estimates new responses based on the correlation between the observed variables in the real dataset, reproducing its relationships. Due to ethical considerations and compliance with Brazilian data protection legislation (LGPD), this study relies on synthetic, anonymized data, however, a nationwide data collection using real-world data is currently being conducted in Brazil. This approach was considered due to the General Law of Data Protection (LGPD). Results show that the relationships of students with family, other students, and school professionals' impacts on school performance. Results indicated that the pedagogical quality (p=.004), parenting (p<.001), support structure at home (p<.001), coordination with the family (p=.048), and belonging (p<.001) impact students’ performance. These findings reinforce the central role of relational dimensions in shaping students’ academic outcomes, highlighting that school performance is not solely determined by individual or structural factors, but emerges from the quality of interactions within the educational ecosystem. Overall, results provide empirical support for the advancement of early warning systems that incorporate relational indicators, contributing to more sensitive, context-aware, and effective strategies to prevent school dropout.
|
Keywords |
School dropout, Large-scale analysis, Educational psychology. |
|
REFERENCES |
[1] Vasconcelos A.N., Freires L.A., Loureto G.D.L., Fortes G., Costa J.C.A., Torres L.F.F., Bittencourt I.I., Cordeiro T.D., and Isotani S. (2023). Advancing school dropout early warning systems: the IAFREE relational model for identifying at-risk students. Front. Psychol. https://doi.org/10.3389/fpsyg.2023.1189283 |
The Future of Education




























