A Discriminant Approach for Allocating Newly Admitted Candidates into the Appropriate Course of Study in Federal Polytechnic Nekede, Owerri

Abstract

This paper demonstrates the use of linear discriminant analysis (LDA) in the admission processes of a higher education in Nigeria. Joint Admission and Matriculation Board (JAMB) and Post Unified Tertiary Matriculation Examination (PUTME) scores of candidates applying for admissions into Science and Laboratory Technology (SLT) and Statistics (STA) departments, since the two departments has the same admission requirements, in 2013/2014 from Federal Polytechnic Nekede, Nigeria was used to derive the linear classification rule after assumption justifications corroborate the case of multivariate normality and equal covariance matrices. Based on this LDA rule, an Apparent Error Rate (APER) of 32% was obtained. This suggests a relatively high rate of misclassification in the admission procedure as could be depicted from a partition plot and a predicted classification scores plot. A possible reason for the high APER could be that JAMB and PUTME may not be the only factors needed to discriminate and classify a student into SLT or STA departments. Other factors like the student's choice may be valuable. Meanwhile, the set rules were used to classify a sample of five new entrants into their appropriate course of study.