<![CDATA[The rapid expansion of digital-based learning has increased academic demands and psychological challenges, leading to higher levels of academic stress among university students. This study aims to examine the roles of social support, academic self-efficacy, and subjective well-being in predicting academic stress in digital learning environments. This study employed a quantitative correlational approach using Partial Least Squares Structural Equation Modeling (PLS-SEM) to analyze data collected from 888 university students in Indonesia. The instruments included the Perception of Academic Stress Scale (PAS), the Multidimensional Scale of Perceived Social Support (MSPSS), a composite Subjective Well-Being Scale (SWLS and PANAS), and an Academic Self-Efficacy scale. The findings revealed that social support significantly enhances subjective well-being and academic self-efficacy while reducing academic stress. Academic self-efficacy demonstrated a paradoxical role, as it significantly increased academic stress while also functioning as both a mediator and a moderator. In contrast, subjective well-being did not have a significant effect on academic stress. In conclusion, academic stress in digital learning is shaped by complex interactions between external and internal factors, with self-efficacy acting as a double-edged construct. This study contributes by challenging the conventional assumption that self-efficacy is purely protective and highlighting its dual role in digital learning contexts.]]>
Copyrights © 2026
