Longitudinal Academic Performance Analysis Using a Two-Step Clustering Methodology

Abstract

The present study aims to examine the academic profiles of industrial engineering undergraduate students among a sample group of military college engineering students ( N= 276) in order to determine the factors impacting academic performance; to compare student groups that were identified by course scores, and to analyse performance changes over four academic years. The study started with data collection, database creation and preparation for clustering study. Atwo-step clustering methodology was used for grouping courses based on academic performance and context similarities. The clustering methodology results are validated by discriminant analysis. Student movements among clusters over the four years are identified in the longitudinal cluster analysis part of the study. Results showed that there is saturated cluster structure among students which has been preserved over years. It was concluded that the importance of background knowledge and prior motivation are effective in the academic performance rather than the change in environment. Although this study is the final stage of an ongoing project in which more than twenty officers are involved, specific data collection process and the analyses are conducted by the authors.