Using Moodle data for early warning of dropping out
Sami Suhonen, Hanna Kinnari-Korpela · Theseus (Ammattikorkeakoulujen) · 2019
Dropout rates are high in engineering education throughout Europe. According to interviews, the main reasons were life situation, learning or studying difficulties, wrong field of studies and health problems. The interviewees would have hoped more support and guidance during their studies and 70 % of the dropouts were very interested to continue their engineering studies. Sometimes support at the right time to right students might solve the problem and studying would continue normally. How to find those students that are at risk of dropping out? Single teachers might notice absences, but they don't have the whole picture. Study counsellors need to rely on the course completion data on the transcript of records, which is always delayed weeks or months in comparison to the studying actions and possible problems in it. Therefore, not even the study counsellors have a real time view to the student progress. Many higher education institutions use digital learning management systems (LMS) like Moodle to deliver their online, blended and face-to-face courses. Students leave digital footprints on the platform and this tells about studying habits. Therefore, LMS data has the potential to show decrease in learning activity well before it becomes visible elsewhere. In this study, two engineering student groups are followed during one academic year covering both simultaneous and consecutive courses. With the data, a simulation is run to raise an alarm if a student is at risk of dropping out. These alarms are then compared with the real dropping out information of the groups.