Predicting Performance in a Subject Using Ordinal Logistic Regression

Abstract

Subjects with important failure rates demand teachers having relevant information about their students to accomplish a differentiated teaching process. Ordinal logistic regression is used to build an equation that relates students' marks in Probabilistic Models of Processes (MPP according to its acronym in Spanish), which is a subject taught in the second year of Industrial Engineering studies, with first year results. Data from 274 students belonging to two different academic years were used to achieve as the best equation, the one that relates MPP grade with the average mark of science subjects studied in the first year. Probability estimations obtained with this model were used as a starting point to develop a method that allowed the prediction of MPP condition for those students who took this subject in the academic year 2011-2012.