Modelling of Students' Performance with Logistic Regression.

      Аннотация: The problem of incessant decline in academic performance of Nigeria students in recent years cannot be over emphasized. Despite importance attached to academic performance, researchers have shown that students’ performance is declining. Ordinal Logistic Regression (Proportional Odds Model) is used to model the data and the results reveal that only sex of students is not a determinant factor of final grade that students may attained at graduation. This research also finds that there is equal chance for both male and female students to graduate from a university with First Class and also that younger students perform better than the older ones.Cochran-Mantel-Haenszel (CMH) statistics also reveal that odds of a student graduating with High Grade (Second Class Upper and above) is about half of graduating with Low Grade (Second Class Lower and below). Breslow-Day and Tarone’s statistics show that the odds ratio across the ten departments (relating to age at entry & final grade and sex & final grade) are all equal.

Ключевые слова: Binary, Categorical Data, Contingency table, grades, modelling, Odds ratio, Probability, Performance Logistic Regression