Academic Analytics: Predicting Success in the Licensure Examination of Graduates using CART Decision Tree Algorithm

Jeffrey A. Clarin, Cherry Lyn C. Sta. Romana and Larmie S. Feliscuzo

The researchers developed a fast and reliable academic analytics model, utilizing classification and regression tree (CART); aimed to help HEIs to monitor and evaluate the academic performance of teacher education students prior to taking the licensure examination for teachers (LET). This was influenced by the fact that much money had been spent on the upgrading of faculty expertise as well as institutional facilities of teacher education institutions, but to date, the success rates in the LET remain low: 48.03% in the September 2018 exams and a dismal 29.91% in the March 2018 examinations. The choice of CART, a tree-like, orderly-arranged structure, was precipitated because it can demonstrate the ideal depiction of reality. In this study, the researchers utilized 348 instances of graduates within a five-year period, from years 2012 to 2017. This study sought to find the characteristics that affected the success in the LET in terms of the following: university exam results; GWA in major subjects; GWA in professional subjects; GWA in general subjects; LET review; and, actual LET results. In essence, this study determined the success rate in the LET when predicted utilizing CART. The test case in this study was a state-run university based in Aklan, Philippines, so chosen because data was accessible to the researchers, one of whom is a member of the faculty and management team of ASU. Hence, the datasets of teacher education graduates were pilot-tested at ASU. The results of this study established the effectiveness and functionality of the researchers-developed academic analytics model utilizing CART.

Volume 12 | 01-Special Issue

Pages: 143-151

DOI: 10.5373/JARDCS/V12SP1/20201057