Profiling students via clustering in a flipped clinical skills course using learning analytics

Area of Science:

• Medical Education
• Learning Analytics
• Educational Technology

Background:

• Flipped classrooms are a student-centered approach in medical education.
• They promote higher-order thinking and problem-solving during in-class activities.
• Effective flipped classrooms require students to engage with pre-class materials.

Purpose of the Study:

• To investigate student profiles and interaction strategies in a flipped classroom.
• To identify patterns of pre-class engagement using learning analytics.
• To inform targeted interventions for student preparation.

Main Methods:

• Collected learning management system interaction data from 375 medical students.
• Preprocessed student interaction data.
• Applied k-means clustering algorithms to analyze engagement patterns.

Main Results:

• Identified a two-cluster structure: 'high interaction' and 'low-interaction' student profiles.
• Demonstrated the ability of clustering to differentiate engagement levels.
• Highlighted distinct interaction strategies among students.

Conclusions:

• Learning analytics and clustering can effectively identify low-engaged students in flipped classrooms.
• Student profiles reveal interaction strategies with online materials.
• Results support the design of tailored interventions to improve student preparation and feedback.