Modeling the sustainability perspectives on personalized digital games for digital citizenship education: A PLS-SEM approach

Abstract

As digital citizenship becomes an essential educational priority in the digital age, there is a growing need for sustainable and engaging instructional designs that foster students' ethical and responsible use of technology. Addressing this gap, this study modeled the sustainability perspectives underlying personalized digital game-based learning through a partial least squares structural equation modeling (PLS-SEM) approach. A longitudinal repeated-measures design was conducted with 372 lower secondary students in Thailand, using fuzzy logic and decision tree algorithms to personalize ethical digital scenarios. The proposed model examined how pedagogical design, content quality, usability, behavioral decisions, and motivation shape students' perceptions of sustainability. Results indicated that sustained motivation at later learning stages was the strongest predictor of perceived sustainability, while pedagogical and experiential factors exerted significant indirect effects through motivational engagement. The analysis also confirmed the longitudinal influence of early motivational experiences on later engagement, emphasizing the importance of adaptive feedback and reflective learning processes. These findings advance understanding of how AI-driven personalization can promote sustainable digital citizenship learning by integrating adaptive pathways, culturally relevant content, and motivational scaffolds to support long-term behavioral change. Implications for educational design, pedagogy, and policy are discussed to guide the development of scalable AI-supported learning environments.