Associations between Social-Media Usage Patterns and Semester  CGPA: A Cross-Sectional Study of University of Dhaka  Undergraduates

The widespread adoption of social media has significantly reshaped students’ daily lives, particularly in the realm of education. As digital connectivity becomes more embedded in academic routines, understanding its actual impact on student performance has become crucial. While global research has documented both positive and negative associations between social media usage and academic outcomes, evidence from Bangladesh remains scarce. This study addresses that gap by examining how usage patterns-categorized by purpose (education, entertainment, communication) and duration-relate to semester CGPA among undergraduates at the University of Dhaka. A cross-sectional quantitative survey was conducted using online data collected from 100 students. Statistical analyses, including correlation and multiple regression, were applied to explore associations. The results revealed a strong negative correlation between total weekly usage time and CGPA, indicating that longer time spent on social media generally predicts lower CGPA. However, when social media is used for educational purposes, it shows a moderate and statistically significant positive relationship with semester CGPA. Entertainment-driven usage exhibited a strong negative effect, while communication-related usage showed a weak but positive correlation. Multiple regression confirmed that time spent, educational use, and entertainment use are significant predictors of CGPA, explaining 49,3% of the variation in academic performance. Educational purpose emerged as the only positive predictor among the three, while communication use was not significant in the multivariate model. These findings suggest that social media's impact is not inherently negative or positive but depends on how and why it is used. Students, educators, and policymakers should focus on promoting purposeful and educational use while minimizing distractions linked to entertainment overuse. As the data were drawn from a single institution using convenience sampling, generalizability is limited. Nonetheless, this research provides valuable insights into optimizing digital behavior to support academic success within the Bangladeshi context.