Derivation of Canonical Discriminant Models for the Academic Performance of Junior Engineering Students of the University of Baguio

Carisa C. Blancas, MSME
Joseph F. Codiaman, BSCE, BSSE
Marie Aurora S. Dolipas, MACT
Victor V. Hafalla Jr., MAAS
Genelou P. Manangan, MPA
Herminigilda L. Orille, MMIR
Conrado C. Rotor Jr., Ph.D.
Aida S. Valerio, MACT


The study sought to derive discriminant models which discriminate between poor and moderate-high achievers among junior engineering students of the University of Baguio using data from 57 respondents representing 78% of the sampled population. Item analysis of the Math diagnostic exam suggested exclusion of some of the questionnaires. Results of the factor extraction using orthogonal rotation revealed four underlying factor constructs, namely: Student’s Initial Academic Achievement (X1), Analysis and Student’s Current Academic Achievement (X2), Mathematical Aptitude (X3) and Non-numerical Mental Ability (X4). The four extracted factors accounted for 71.734% of the total variability in the data. Regression analysis on the student’s academic performance during their junior year using the different factor constructs as variates revealed the four factor constructs as significant predictors of the former. Results of the agglomerative hierarchical clustering algorithm using Ward’s method suggested two groups be adopted (moderate-high achievers and poor achievers). Results of the derivation of the discriminant model produced one canonical discriminant function which is statistically significant to effect group discrimination. The model’s hit ratio of 98.2% deemed it to be a better model than a chance model. Results of the study confirmed the multi-dimensionality of tertiary academic performance and also confirmed findings from previous studies on predictors of tertiary academic performance. Further validation of the discriminant functions could be done using cross-validation. Also, it will be interesting to check the veracity of the derived factor constructs and discriminant function with the exclusion of the diagnostic exam results.

Source: UB Research Journal, Vol. XXXIII, No. 1, January – June 2009