Avaliação da nota de redação do Enem no estado do Ceará via modelo de regressão beta inflacionado

• Ano de defesa: 2023
• Autor(a) principal: Lima, José Wenes Pereira
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Informática; Programa de Pós-Graduação em Modelagem Matemática e computacional; UFPB
• Palavras-chave em Português: Matemática computacional - Modelagem; Enem - Exame Nacional do Ensino Médio; Modelo de regressão beta inflacionado; Computational mathematics - Modeling; Inflated beta regression model; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/29828

The National High School Exam (ENEM) is currently the most important instrument for assessing basic secondary education in Brazil. Furthermore, this assessment tool heavily relies on the scores obtained by participants as a predominant criterion for accessing public and private universities in the country. The exam objectively evaluates participants in four areas: humanities and their technologies, natural sciences and their technologies, mathematics and their technologies, and languages and codes and their technologies. In addition to these, the exam also includes a writing test, which is the only non-objective part of the exam. The score on the writing test is crucial for a candidate’s approval in various university courses, as it goes beyond mere writing skills and assesses the student’s ability to analyze a topic, construct sound arguments, provide evidence for their ideas, and propose solutions. These skills are relevant in an academic environment where conducting research, debating ideas, and participating in discussions are necessary. This dissertation aims to identify the factors that influence the scores on the writing test for participants who took the ENEM in 2019 in the state of Ceará, Brazil. To achieve this, the inflated beta regression model was used, given that the response variable exhibits asymmetry and takes values in the interval [0, 1]. The data was obtained through the INEP portal, specifically the page related to the ENEM. The number of observed participants was 74,943 from public schools and 5,279 from private schools. Through the analysis of the proposed models, it was found that both in public and private schools, the scores on the objective tests influence the scores on the writing test for students. Additionally, the scores on humanities, languages and codes, and mathematics and their technologies strongly influence the probability of participants achieving a score of 1000 score on the writing test.