Paper Type
Complete
Paper Number
PACIS2025-1975
Description
Motivational information system design, such as the use of gamification, has become a prominent dimension of IS design and research especially in user-facing systems where user engagement is crucial such as in crowdsourcing, electronic-commerce or online education. A major bottleneck currently is the inability to tailor or personalize motivational design to more accurately match users’ behavioral patterns, background traits or personality. Towards this end we investigate how user behavior logs, collected from the last 15 years comprised of 10,000 users of GitHub, can be used to predict user engagement with different types of motivational design through machine learning. The results show that behavior-related metrics are the most predictive, followed by social network-related metrics, while user profile-related metrics are less influential. These findings challenge static user classification models, emphasizing the need for adaptive, network-aware gamification strategies. This study offers theoretical insights into gamification engagement and practical recommendations for personalized gamification design.
Recommended Citation
Tang, Hairui; Xi, Nannan; and Hamari, Juho, "What Kinds of Users Engage with Different Types of Motivational Design: A Machine Learning Study in GitHub" (2025). PACIS 2025 Proceedings. 17.
https://aisel.aisnet.org/pacis2025/dig_plat/di/17
What Kinds of Users Engage with Different Types of Motivational Design: A Machine Learning Study in GitHub
Motivational information system design, such as the use of gamification, has become a prominent dimension of IS design and research especially in user-facing systems where user engagement is crucial such as in crowdsourcing, electronic-commerce or online education. A major bottleneck currently is the inability to tailor or personalize motivational design to more accurately match users’ behavioral patterns, background traits or personality. Towards this end we investigate how user behavior logs, collected from the last 15 years comprised of 10,000 users of GitHub, can be used to predict user engagement with different types of motivational design through machine learning. The results show that behavior-related metrics are the most predictive, followed by social network-related metrics, while user profile-related metrics are less influential. These findings challenge static user classification models, emphasizing the need for adaptive, network-aware gamification strategies. This study offers theoretical insights into gamification engagement and practical recommendations for personalized gamification design.
Comments
Platform