Technical Summary: The Effects of Gamification on Students' Emotions

Problem Statement

While gamification is widely adopted in educational systems to enhance engagement and learning outcomes, its specific impact on learners' emotional states remains underexplored. Existing literature has predominantly focused on cognitive and behavioral outcomes, leaving a gap in understanding how gamified environments influence basic emotions (happiness, sadness, anger, surprise, fear, and disgust). Furthermore, traditional assessments often rely on static self-reports rather than objective, real-time measures of emotional dynamics. This study addresses the need to investigate whether gamification affects the duration, frequency, and sequential transitions of students' basic emotions during interaction with a Learning Management System (LMS).

Methodology

The study employed a controlled between-subjects experimental design involving 62 university students (average age 25, from 19 countries) recruited from Tampere University.

• Experimental Design: Participants were randomly assigned to either an experimental group (using a gamified LMS) or a control group (using a non-gamified version of the same LMS).
• System: The study utilized the Eagle-edu platform. The experimental version incorporated 10 specific gamification elements (Acknowledgment, Chance, Competition, Economy, Imposed Choice, Level, Objective, Points, Progression, Stats) designed to create a meta-game interaction around learning tasks. The control version presented identical educational tasks without these elements.
• Procedure: Participants interacted with the system for at least 30 minutes in a usability laboratory.
• Data Collection: Emotional responses were captured in real-time using automated facial-expression recognition software (Noldus FaceReader) and logged via Noldus Observer XT. Six basic emotions were analyzed.
• Data Analysis:
  • Descriptive Statistics: Calculated for emotion duration and frequency.
  • Inferential Statistics: Student's t-tests were used for normally distributed data (Neutral), and Mann-Whitney U tests were used for non-normally distributed data to compare duration and frequency between groups. Effect sizes were calculated (Cohen's d or r).
  • Transition Analysis: Emotional transitions were examined descriptively to identify sequential patterns (e.g., anger-to-happiness) due to the exploratory nature of the analysis and sample size limitations.

Key Results

1. Emotional Duration and Frequency (RQ1 & RQ2)

Statistical analysis revealed no significant differences between the gamified and non-gamified groups regarding the total duration or frequency of any of the six basic emotions.

• Dominant Emotion: Anger was the most frequently observed emotion in both conditions.
• Least Expressed: Fear was the least expressed emotion in both conditions.
• Statistical Significance: All p-values for duration and frequency comparisons were non-significant (e.g., Angry duration p = .91; Happy frequency p = .345).
• Effect Sizes: Associated effect sizes were small across all emotions (ranging from r = .03 to r = .27), suggesting a limited practical impact of the specific gamification design on basic emotional expressions in this context.

2. Emotional Transitions (RQ3)

While static measures showed no difference, descriptive analysis of emotional transitions revealed distinct patterns between the two conditions:

• Gamified Condition (Experimental): The most frequent transition was from Anger to Happiness (19 occurrences). Anger also frequently transitioned to Surprise (14 times).
• Non-Gamified Condition (Control): The most frequent transition was from Disgust to Anger (28 occurrences).
• General Patterns: Both groups showed bidirectional shifts between Anger and Happiness. However, negative-to-negative transitions (e.g., Disgust to Anger) appeared more prominent in the non-gamified condition. The emotion "Scared" showed no significant transitions in either group.

Key Contributions

1. Objective Measurement: The study provides empirical evidence on emotions in gamified learning using automated, objective facial-expression recognition rather than subjective self-reports, offering a high-resolution map of emotional states during LMS interaction.
2. Focus on Dynamics: By analyzing emotional transitions alongside static states, the study highlights the temporal dynamics of student emotions, suggesting that the sequence of emotions may differ even when the frequency or duration does not.
3. Contextual Evidence: The findings contribute to the literature by indicating that a specific gamification design (meta-game elements like points and badges) did not produce detectable changes in basic emotional expressions compared to a non-gamified interface, challenging assumptions that gamification inherently alters basic affective states.

Significance and Claims

The authors claim that their findings suggest the specific gamification design used in this study did not significantly alter basic emotional expressions (duration/frequency) but may have influenced emotional dynamics (transitions).

• Modest Claims on Null Findings: The paper explicitly states that the absence of statistically significant differences should not be interpreted as definitive proof that gamification has no emotional effects. Instead, it indicates that under these specific conditions (task type, duration, specific elements), large emotional differences were not detected. The study acknowledges it may lack the power to detect small or moderate effects.
• Interpretation of "Anger": The authors caution that the high prevalence of "anger" may not solely indicate frustration but could reflect facial configurations associated with cognitive effort, concentration, or problem-solving, a known limitation of automated facial recognition.
• Theoretical Implications: The results support the view that core learning content may be a primary driver of student emotions, with gamification elements having a comparatively small impact on basic emotional intensity. The study suggests that for gamification to be effective in influencing emotions, it may need to be "intrinsic" to the learning mechanics rather than an extrinsic reward layer.
• Practical Implications: Designers are encouraged to focus on cognitive load and task meaningfulness to foster positive emotions, as gamification elements alone may not suffice. Furthermore, the identified transition patterns (e.g., anger-to-happiness) could serve as algorithmic triggers for adaptive interventions in "emotionally-aware" smart learning systems.

The paper concludes by calling for future research with larger samples, diverse cultural contexts, and alternative emotion detection methods (e.g., self-reports, physiological sensors) to validate these findings and explore the nuances of emotional transitions in educational settings.