Forecasting of academic performance in exact sciences for a public university admissions using binary logistic regression

Abstract

Because of the deficient secondary level education that the applicant receives before entering the university and the academic rigor of the entrance exam as well as the admission course in the Escuela Superior Politécnica del Litoral (ESPOL), there is an adverse scenario for many applicants who delay the admission to the institution or are unable to do so. Identifying and taking advantage of the benefits of using an innovate active learning methodology as an alternative to traditional learning methodology can help to improve this situation. The objective of this article is to predict the academic performance of the ESPOL applicants in exact sciences, emphasizing the differences between those who attend the course with the active learning  modality and those who do it with the traditional modality. The data analysis considers the 558 applicants registered in the intensive course in February 2020. A binary logistic regression technique is applied, with a  dichotomous dependent variable called academic performance and dependent variables of academic and demographic nature. A relevant of this research indicates that one of the main predictors of academic performance is the modality with which course is attended. Applying this innovative and technological methodology allows a process with a higher admission rate and academic performance compared to the traditional methodology.