When Anger Strikes: Using AI Modelling to Understand How Negative Emotions Impact Performance in Digital Math Games

Area of Science:

• Educational Technology
• Cognitive Psychology
• Learning Sciences

Background:

• Digital game-based learning environments (DGBLEs) are common in education.
• Little is known about how student emotions affect digital gameplay performance.

Purpose of the Study:

• Examine the influence of emotions and arousal on performance in DGBLEs.
• Investigate differences in performance and emotional impact between math and non-math games.

Main Methods:

• Thirty-two 5th-grade students played math and non-math digital games.
• Real-time heart rate and affective data were collected.
• An interpretable machine learning approach (XGBoost) analyzed the data.

Main Results:

• Students performed better on non-math games than math games.
• Real-time anger correlated with lower performance, especially in math games.
• Other emotions and physiological measures did not significantly predict performance.

Conclusions:

• Discrete emotions, particularly anger, significantly impact math game performance.
• Emotional regulation is vital for optimizing digital math learning.
• Findings underscore the importance of affective dynamics in DGBLEs.