The higher education system in the 21st century has moved towards internationalisation where students can study online remotely. Universities’ virtual learning environments (VLEs) have changed the relationship that student has with their course of study, which significantly supports student learning and course engagement. VLEs not only provide a platform to transfer content, but also offer high-tech tools to support learning in a flexible environment. VLEs can provide analytics on student’s course engagement, i.e., total activity time, accessed page counts, last login date, etc., to monitor student’s progression. This work investigates the usefulness of engagement assessment tools (EAT) available through Canvas VLE to predict students’ final exam scores using predictive modelling. A multiple regression model is introduced to develop an equation that predicts exam scores based on EAT and coursework performance data. The equation was developed using data from 55 students in a course, and assessed for strength of relationship, presenting a statistically significant correlation with R2 = 0.611, and p-value < 0.0005. The results highlight that the data obtained from EATs are of minor value compared to coursework performance. furthermore, an optimization routine was developed to increase the accuracy of the equation through a decision variable X for each student that represents other factors (i.e., personal interests and studying hours). Variable X has proved to have a significant contribution to the equation and predicting student performance more accurately. The broader purpose of this is to see whether EATs can be used to identify disengaged students so that early interventions can be made to build a sense of self-efficacy in any at-risk group.
|Publication status||Published - 2021|
|Event||The 9th European Conference on Education - UCL, London, United Kingdom|
Duration: 15 Jul 2021 → 18 Jul 2021
|Conference||The 9th European Conference on Education|
|Period||15/07/2021 → 18/07/2021|