Related Experiment Videos
How learning analytics can early predict under-achieving students in a blended medical education course
Mohammed Saqr1,2, Uno Fors2, Matti Tedre2
1a College of Medicine , Qassim University , Qassim , Kingdom of Saudi Arabia.
Medical Teacher
|April 20, 2017
Summary
Learning analytics can predict medical students at risk of underachieving using online activity data. This early warning system helps identify students needing support to improve learning outcomes.
Area of Science:
- Medical Education
- Learning Analytics
- Educational Technology
Background:
- Learning analytics (LA) offers untapped potential in medical education for early identification of underperforming students.
- Analyzing online student data can optimize learning processes and environments.
Purpose of the Study:
- To identify quantitative markers from student online activities correlating with final performance.
- To investigate the prediction of students at risk of failing or dropping out.
Main Methods:
- 133 students in a blended medical course had their online activity data extracted from the learning management system.
- Data included logins, views, forum participation, time spent, and assessments.
- Five engagement indicators were calculated to reflect self-regulation and engagement.
Main Results:
- Final grade prediction accuracy reached 63.5%, improving to 80.8% with a binary logistic model.
- 53.9% of at-risk students were identified.
- Student engagement and consistent resource use were key predictors.
Conclusions:
- Learning analytics techniques applied to online activities in blended medical courses can predict underachieving students.
- This provides an early warning system for timely interventions to support student success.