Loading...
Publication
Modelação estatística aplicada a admissão dos estudantes na Universidade de São Tomé e Príncipe
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.08 MB | Adobe PDF |
Authors
Abstract(s)
O presente estudo investiga alguns fatores que influenciam a admissão de estudantes na Universidade de São Tomé e Príncipe (USTP) por meio da aplicação de técnicas avançadas de análise estatística, de forma a proporcionar insights que possam conduzir a melhorias nas práticas de selecção e a formulação de políticas mais inclusivas.
Dada a heterogeneidade das disciplinas que compõem os cursos do ensino secundário, esta análise contemplou duas situações distintas: uma para os alunos de Letras (História como nuclear) e outra para os alunos de Ciências (Matemática como nuclear). Os resultados revelaram que variáveis como a idade, género, notas em disciplinas específicas (Biologia, Física/Química, Língua Portuguesa, Matemática, História, Direito, Psicologia/Sociologia), e a escola de origem dos candidatos possuem efeitos estatisticamente significativos tanto na média final do ensino secundário quanto na probabilidade de admissão em diferentes departamentos da universidade. A análise indicou que mais de 75% da variação na média do ensino secundário pode ser explicada pelas variáveis incluídas no modelo de Regressão Linear Múltipla. Com base nos dados do período em estudo, verificou-se que no grupo dos alunos admitidos em cursos de Ciências a média final diminui em média 0,211 pontos com o aumento da idade e que, inversamente, a média final dos alunos dos cursos de Letras aumenta 0,009 pontos, em média, com o aumento da idade.
A Regressão Logística Multinomial mostrou que os modelos que incorporaram as variáveis idade, média final do 12º ano, notas em Matemática, Língua Portuguesa, História, Direito, Física/Química, Sociologia/Psicologia, género, residência e as escolas onde os estudantes concluíram o ensino secundário contribuem significativamente para a discriminação dos cursos afetos aos departamentos da USTP. Igualmente, mostrou que a probabilidade de inscrição em determinados cursos diminui ou aumenta significativamente com base nas variáveis analisadas.
O estudo realizado mostrou que a modelação estatística é uma ferramenta eficaz para compreender e melhorar o processo de admissão na USTP, fornecendo insights valiosos para a tomada de decisões e contribuindo para a formação de um corpo discente mais diversificado e talentoso. Além disso, os modelos revelaram que mais 60% dos estudantes matricularam-se nos cursos afetos aos departamentos onde as probabilidades de admissão eram mais altas.
The present study investigates the factors influencing student admission at the University of São Tomé and Príncipe (USTP) through the application of advanced statistical analysis techniques, providing insights that could lead to improvements in selection practices and the formulation of more inclusive policies. Given the heterogeneity of subjects that constitute secondary education courses, this analysis considered two distinct scenarios: one for Humanities students (with History as the core subject) and another for Science students (with Mathematics as the core subject). The results revealed that variables such as age, gender, grades in specific subjects (Biology, Physics/Chemistry, Portuguese Language, Mathematics, History, Law, Psychology/Sociology), and the candidates' originating schools have statistically significant effects on both the final secondary education average and the probability of admission to different university departments. The analysis indicated that more than 75% of the variation in the secondary education average can be explained by the variables included in the Multiple Linear Regression model. Based on the data from the study period, it was found that in the group of students admitted to Science courses, the final average decreases by an average of 0.211 points with increasing age, whereas, conversely, the final average of Humanities students increases by an average of 0.009 points with increasing age. The Multinomial Logistic Regression showed that models incorporating variables such as age, final average of the 12th grade, grades in Mathematics, Portuguese Language, History, Law, Physics/Chemistry, Sociology/Psychology, gender, residence, and the schools where students completed secondary education significantly contribute to distinguishing courses related to USTP departments. It also demonstrated that the likelihood of enrollment in certain courses significantly decreases or increases based on the analyzed variables. The study demonstrated that statistical modeling is an effective tool for understanding and improving the admission process at USTP, providing valuable insights for decision-making and contributing to the formation of a more diverse and talented student body. Additionally, the models revealed that more than 60% of students enrolled in courses related to departments where the probabilities of admission were highest.
The present study investigates the factors influencing student admission at the University of São Tomé and Príncipe (USTP) through the application of advanced statistical analysis techniques, providing insights that could lead to improvements in selection practices and the formulation of more inclusive policies. Given the heterogeneity of subjects that constitute secondary education courses, this analysis considered two distinct scenarios: one for Humanities students (with History as the core subject) and another for Science students (with Mathematics as the core subject). The results revealed that variables such as age, gender, grades in specific subjects (Biology, Physics/Chemistry, Portuguese Language, Mathematics, History, Law, Psychology/Sociology), and the candidates' originating schools have statistically significant effects on both the final secondary education average and the probability of admission to different university departments. The analysis indicated that more than 75% of the variation in the secondary education average can be explained by the variables included in the Multiple Linear Regression model. Based on the data from the study period, it was found that in the group of students admitted to Science courses, the final average decreases by an average of 0.211 points with increasing age, whereas, conversely, the final average of Humanities students increases by an average of 0.009 points with increasing age. The Multinomial Logistic Regression showed that models incorporating variables such as age, final average of the 12th grade, grades in Mathematics, Portuguese Language, History, Law, Physics/Chemistry, Sociology/Psychology, gender, residence, and the schools where students completed secondary education significantly contribute to distinguishing courses related to USTP departments. It also demonstrated that the likelihood of enrollment in certain courses significantly decreases or increases based on the analyzed variables. The study demonstrated that statistical modeling is an effective tool for understanding and improving the admission process at USTP, providing valuable insights for decision-making and contributing to the formation of a more diverse and talented student body. Additionally, the models revealed that more than 60% of students enrolled in courses related to departments where the probabilities of admission were highest.
Description
Keywords
Universidade de São Tomé e Príncipe Modelação estatística Regressão linear múltipla Regressão logistica multinomial Acesso ao ensino superior São Tomé e Príncipe Statistical modeling Multiple linear regression Multinomial logistic regression Higher education access University of São Tomé and Príncipe
Citation
Publisher
CC License
Without CC licence