Academic performance analysis to support proactive student advising for an electrical engineering program

Abstract

Using correlation, regression and hierarchical clustering methods, the authors examined three consecutive graduating cohorts of students in an electrical and computer engineering undergraduate program to determine which courses (or groups of courses) were the best predictors of graduation GPA. The aim was to develop predictive models that support a consistent proactive advising experience. The main impact of this study is the methodology which can be applied to other programs with similar weighted GPA schemes and with limited data sources. Other impacts were: the model identified which types of courses impacted GPA performance most, bringing clarity as to where cohort-wide intervention may be required; and the model can help us identify earlier 'at-risk' and 'exceptional' students.