Enhancing Academic Performance of Engineering Students: A Multivariate Statistical Analysis of Grades in a Mathematics

This study highlights the pivotal role of multivariate statistics in educational research and teaching innovation. By integrating mathematical rigour and interpretative analyses, the research refines the evaluation instruments and recalibrates weight distributions in the grade calculation formulas. Techniques such as multiple linear regression, multivariate normal distribution, principal component analysis, and factor analysis address unsatisfactory results in prior courses. The findings advocate for a shift from traditional evaluations to unit-based learning outcomes, incorporating punctual class attendance as a key metric to improve pass rates. A Chernoff face analysis validated the methodology, demonstrating its efficacy in optimising academic performance. Applied within a university-level engineering education context, this approach yielded highly positive results, offering a robust framework to enhance learning and assessment practices. This replicable model provides valuable information for educators looking for effective strategies to enhance student success in STEM disciplines.