Applying Learning Analytics to Detect Sequences of Actions and Common Errors in a Geometry Game.

Area of Science:

• Educational Technology
• Learning Analytics
• Game-Based Learning

Background:

• Games offer learning and assessment benefits but face limited classroom adoption.
• Barriers include lack of actionable data and unclear student interaction/learning outcomes.
• Teachers need better insights into gameplay for effective educational integration.

Purpose of the Study:

• To provide teachers with interpretable and actionable sequence and process mining metrics.
• To enhance understanding of student interactions within the geometry game Shadowspect.
• To support teachers in making informed assessments and personalized interventions.

Main Methods:

• Utilized data from the implementation of Shadowspect in U.S. schools.
• Developed and integrated two sequence and process mining metrics.
• Implemented these metrics within an interactive dashboard for teachers.

Main Results:

• The developed metrics offer insights into student action sequences and common errors.
• Teachers can interpret gameplay patterns to understand student learning processes.
• The dashboard facilitates a clearer view of how students engage with the game.

Conclusions:

• Sequence and process mining metrics can bridge the gap in educational game assessment.
• Actionable data empowers teachers to provide targeted support.
• Integrating such metrics enhances the effectiveness of game-based learning in schools.