Estatística descritiva, análise multivariada e modelo linear para analisar as notas da redação do ENEM 2017

Abstract

The main dissertation developed here, has in its main analysis in the method of evaluation of the Essays of ENEM. This is evaluated in five competencies; each can have a grade ranging from zero to two hundred. These competences are independently assessed by two evaluators, and the final grade is the average of the two grades given by them. Writing is part of school life from elementary school and lasts until the end of life. Through writing, the subject that produces it has in his writing a way of exposing his speech and therefore this is still a speech, a speech through writing, and with that this subject is made to exist for the other, who is the reader. Through the analysis of descriptive, multivariate statistics and finally a linear modeling, it is noticed that the scores of the five competencies are correlated with each other, where some are more correlated than others. The competence that is in charge of perceiving the cohesion of the text is the one that has the greatest influence on the scores of the other competencies, as well as in the writing. The competence that proposes a solution to the problem proposed in the main theme, this is the one that most distances itself from the others, it is practically as if it were another member almost independent of the text. With the linear modeling of competences, analyzed through the variables: Gender, Age, Ethnicity, Father's education, Maternal education, Family income, Internet access and Type of high school attended by the participant; it can be seen that there are factors that influence the increase in the scores of these skills, such as family income, internet access and having attended private school in high school. Therefore, the income factor has a relationship with the final grades of these competencies, where an increase in income already perceives an increase in the grades of the competences, consequently of the wording.