Factors associated to academic performance in university students

Abstract

This study present a methodology to determine factors associated to academic performance (AP) in a 2014 cohort of students from different faculties at a northwest private university in Colombia. A canonical discriminant and a logistic regression are used to identify the association between pedagogical, institutional, and sociodemographic variables with low AP by using only information obtained from the UCO’s administrative systems. Discrimant analysis succesfully classified 100% of students with low AP trough institutional and sociodemographic variables. Pedagogical variables did not show adequate classification, except for faculty of law. Logistic regressions showed significant associations between student degree path, having a scholarship, the repetition indicator, and the indicator of canceled subjects per semester. It is concluded that discriminat analysis and logistic regression allow to classify and establish the factors that explain the low AP.