Decision Tree Applied to the Study of SAEB Performance in Mathematics of Students from Amapá

Abstract

SAEB is a national assessment that measures student performance in Portuguese and Mathematics in Basic Education. SAEB results provide data for monitoring and improving the quality of education in the country. The Amapá Mathematics Education Research Group – GEMAP, linked to the Federal University of Amapá, began a series of studies on SAEB regarding student performance in Mathematics and Portuguese with a focus on Amapá and the Norte region of Brazil. The objective of this study is to estimate profiles of students from Amapá below the national average who participated in the SAEB test. As a methodology, the classification tree algorithm, CHAID method, was used, with a sample base of 9462 students from Amapá, year 2019, the database was downloaded from the INEP website. It was observed that 68.5% of students are below the national average (277,3 points in Mathematics in 2019), regarding profiles, the classification tree indicated that the profiles with the highest probability of being below average are students from “public schools that live in an area rural” with 83,7%. The study helped to show which profiles are most likely to have below average grades in mathematics and thus contribute to the formulation of public educational policies.