Application of Ordinal Logistic Regression in the Study of Students’ Achievement in External Testing

Abstract

The logistic regression describes the relationship between a binary (dichotomous) response variable and explanatory variables. In this study, we have implemented logistic regression to evaluating the probability of passing the external Testing to students in Economics' Secondary School ''8 Shtatori'' in Tetovo, Macedonia. Drawing on SEC (State Examination Center) for external testing database, Gender, Ethnicity, School year and Subjects of the Course are selected as predictive variables. The results verify that students Macedonian has 7.3 times more chance to gain a passing grade in External Testing compared to a student Albanian, students of second-year has 12.2% less chance to gain a passing grade in External Testing compared to a students of first-year, also students of third-year have 12.2% less chance to gain a passing grade in External Testing compared to a students of second-year.