Modelling engineering student academic performance using academic analytics

Internationally, the recruitment, management and retention of students has become a high priority for universities. The use of information technology systems and student data by institutions to understand and improve student academic performance is often referred to as ‘academic analytics’. This paper presents an academic analytics investigation into the modelling of academic performance of engineering students enrolled in a second-year class. The modelling method used was binary logistic regression, and the target predicted variable was ‘success status’—defined as those students from the total originally enrolled group that achieved a final unit grade of pass or better. This paper shows that student data stored in institutional systems can be used to predict student academic performance with reasonable accuracy, and it provides one methodology for achieving this. Importantly, significant predictor variables are identified that offer the ability to develop targeted interventions to improve student success and retention outcomes.