The Main Factors of Enem: A Literature and Microdata Perspective

This is a Research Full Paper. Enem is a test applied annually since 1998 in Brazil with the purpose of evaluating high school students. The data from the tests are made available in Comma-separated values (CSV) files by Inep, enabling analysis to be carried out on more than 102 million records, combining factors and levels with a wealth of detail from a social and educational point of view. This is one of the most consistent databases in the world and yet there is no article that aims to establish, from the literature, the most frequent factors, relating the factors evidenced with data mining to interactions. From this discovery, it was possible to create unique visualizations of the last 25 years, exploring interactions between factors and levels, which enabled the discovery of new knowledge. To guide this study, a bibliographic search was carried out of all the articles returned in the Google Scholar search, making it possible to identify the most referenced factors. With this information, data was pre-processed using the Pareto principle (80/20), calculating averages and establishing grade levels and income levels in order to normalize the data, since the tests have varied in difficulty over the years and there have been significant changes in the economy since the first version of the exam. As for student performance, the financial investments made at primary and secondary level are enough to provide an indication to establish assumptions and, for this reason, family income is strongly related to educational opportunities and all the main factors. The best primary and secondary education institutions are federal and private schools. Skin color exerts some influence, but does not interfere significantly when one has advantageous financial conditions. The results highlight the importance of thinking about Brazil's poorest population, because as well as being socially vulnerable and disadvantaged, their performance is below what can be achieved.