Improving Admissions into Technical and Vocational Institutions Through a Statistical Classification Technique

Abstract

Polytechnic education at the Tertiary level is part of the the crowning achievement or the capstone of the traditional educational structure. It is the hub of human development worldwide. Therefore the system of admissions into these institutions anywhere is an area of research interest.  Many factors are considered before students are admitted into programmes at the tertiary level of education and this differs from country to country. However, a growing conflict is the entry requirements into the traditional universities on one hand and the polytechnic and the vocational institutions on the other hand. According to Stubbs (Stubbs, 1998), at least half of the differences in students’ academic performance could be attributed to the students’ social background and prior attainment rather than the school they attended. Using Ghana’s example as a case study, the multivariate classification tool of linear Discriminant Analysis was conducted on students’ academic performance with entry grades as the predictors and the classes they obtain at the end of their study at the Polytechnic level as the response variable.  Exploratory data analytical tools of normality, multicolinearity, homocedasticity amongst others were employed on the data. Further analysis on the transformed data using linear discriminant analysis revealed that the elective subjects as predictors have far stronger discriminating powers than the core subjects. This suggests that the current entry requirement policy into the polytechnics that focuses on the core subjects must be looked at again. The current entry regime which is more like that of the universities tends to turn away many otherwise very good technical and vocational education materials because of fails in one or two core courses.

Keywords: Discriminant Analysis, Multicollinearity, Homocedasticity, Centroids, Eigen value, Normality