Exploring the Evolution of Educational Serious Games Research: A Topic Modeling Perspective

Abstract

This study aims to reveal the dominant research interests and models in serious games research using topic modeling analysis. The dataset of this study covers a comprehensive collection of 2676 articles from the past to the end of 2022, indexed in the Scopus database. The study begins by presenting descriptive attributes of the articles, including their publication years, subject areas, and the journals in which they are published. Subsequently, employing topic modeling analysis, a form of unsupervised machine learning, the study identifies concealed themes, research interests, and tendencies within the literature. The findings indicate a notable surge in publications in this domain, particularly post-2009 and 2019. Furthermore, the study identifies eleven primary topics dominating the literature, with notable emphasis on 'Training of STEM-related fields,' 'Programming learning,' and 'Medical education'. To gauge the dynamics within these topics, the study calculates accelerations both within individual topics and in comparison to others over time. Remarkably, 'Child and adolescent health' emerges as the topic with the highest self-acceleration, while 'Medical education' stands out for its acceleration in comparison to other topics. In sum, the outcomes of this study, which provides a comprehensive overview of the serious games field, are anticipated to yield valuable insights for understanding the current landscape, guiding future research endeavors, and shaping the trajectory of this field.