The aim of this website is to provide you with an overview of the learning analytics project “PredictEd” that is currently being conducted with 17 different first year modules within DCU. The work described in this project uses data generated from students’ online behaviour, in order to improve their learning experience and specifically, their performance in end-of-semester written examinations. Using log data from the University’s online virtual learning environment, Moodle, combined with past exam performance data we are able to build a software predictor which accurately classifies whether a student in the current cohort of students is likely to pass or fail the module.
This classifier leverages online behaviour and examination outcomes from past students, in order to inform current students as to how they are progressing. We target University students in their first semester when they are most vulnerable and often feel lost or overwhelmed by what is for most, a sudden change to University life. We use past, and present, log data to predict likely outcomes on a weekly basis and naturally the accuracy of our predictions is likely to get more accurate as the module progresses. As a form of alerting, students receive emails each week that advises them to study more, that they really need to study more, or that they seem to be doing OK, whatever is appropriate.
Automated, weekly, personalised emails to students is an excellent adjunct to support from the Lecturer, tutors, lab supervisors or other student supports, but even with these in place, it is inevitable that some students will slip through the cracks because of the sheer numbers of students on some modules