Entre privilégios e barreiras: modelagem linear para estimar desempenho de participantes no Enem

Abstract

A paradoxical situation permeates sapiens, on the one hand, individuals who have almost the same characteristics of body composition can be seen, and on the other, the particularities that individualize each person are evident. In fact, human diversity is a reality, which scholars point out as a condition that can contribute to the improvement of experiences in society. The consideration of these aspects is no exception for education, especially for evaluation. Since some people already start at a disadvantage because they are born into groups with sociohistorically marginalized attributes, which is evident in performance evaluations, especially the Enem, which, when crossing socioeconomic history with exam performance, suggests that the antecedent context arises as an extra barrier for the candidate to overcome, a consideration that the test formulation should address. In the meantime, the objective was to develop a model that estimates the performance of the individual evaluated in the 2019 Enem Human Sciences test. For this purpose, descriptive statistics were applied, it was analyzed whether there is a difference in performance between the groups of characteristics of those examined by ANOVA, as well as the characteristic curve of items was investigated and a linear model was developed to estimate the participant's performance, considering their socioeconomic characteristics. The findings suggest that some exam questions have a greater ability to discriminate between students who are maintaining a consistency of correct answers and those who are not. The profile of those who obtained the lowest average results was also identified, called ‘origin’ or 'basic' profile. And attributes of candidates that influence performance on the test were identified, especially for the factors family income and type of high school.