Desigualdades educacionais no ENEM 2023 : uma análise multivariada da associação entre sexo, cor/raça, tipo de escola e desempenho em Sergipe

Abstract

This undergraduate thesis investigates educational inequalities among the participants of the 2023 National High School Exam (ENEM) in the state of Sergipe, considering the variables of gender, race/color, type of school, and test performance. The microdata from INEP, comprising over 65,000 records, were used. However, a central methodological challenge was the high incidence of missing data in essential variables, such as school type, which required the application of data imputation (using the MissForest algorithm) to enable the analysis. Aware that imputed values are estimates, the interpretation of the results requires caution. The subsequent methodology included descriptive analysis, ANOVA, Tukey's HSD Test, ChiSquare Test, and Multiple Correspondence Analysis (MCA). The results revealed significant performance disparities among racial groups, with white students generally showing higher average scores in all evaluated areas. Performance differences were also observed between genders, with males outperforming in Mathematics and Natural Sciences, while females achieved better scores in the Essay. Regarding the type of school, students from private institutions outperformed those from the public system in all knowledge areas. The multivariate analysis showed that white and male students are more associated with private education, whereas black, brown, indigenous, and female students are concentrated in public education. These findings, analyzed in light of the database limitations, highlight the persistence of structural inequalities in the educational system of Sergipe and indicate the urgency for public policies that promote greater equity.