"Não sabe", "não respondeu" na inscrição do ENEM: o potencial interpretativo das respostas imprecisas e ausentes

Abstract

This work aims to identify the analytical potential of imprecise data (“Do not know”, “Do not answer” types) and missing data from the ENEM 2022 Examination registration questionnaire, in order to raise hypothesis related to the choice of this type of declaration. Thus, it also aims to contribute to decision-making processes with data with such limitations. Methodologically, we opted for descriptive and inferential analyses, using linear regression models, to look for similarities and statistically significant distinctions between the candidates who declared inaccurate answers and the other categories present in the questions. The results indicate that two situations summarize the trends observed in the data: the difficulty in understanding a question in the questionnaire, which is associated with consistently lower grades in the exam; the presence of intervening factors or hidden categories in the data. Furthermore, this work presents interpretations for each question with imprecise or missing data, with emphasis on color/race, marital status, school administrative dependency and location, and parents' education and occupation.