Using Multinomial Logistic Regression Model to Predict the Effect of Social Media on Academic Performance of College Students

Abstract

Social media networking has become an integral part of communication today, with widespread usage across various demographics. This study was conducted to investigate the impact of social media on student academic performance, recognizing its prevalence and influence in educational settings. A random sample of 1,692 students was selected to participate in the study. A multinomial logit model was developed to predict student performance based on significant predictors, including age, marital status, monthly budget for social networks, monthly stipend, and daily private study time on social media. The results showed that age, marital status, monthly social network subscription budget, monthly stipend, and private study time on social media were statistically significant. The likelihood of achieving a 2.40–3.49 CGPA was highly dependent on age, marital status, monthly budget for social media subscription, and private study time with p-values of 0.018, 0.000, 0.000, and 0.000 respectively. Those students who studied less than 1 hour and those who spent 1-2 hours daily on social media were more likely to attain a 2.40-3.49 CGPA. Additionally, a 1.50-2.39 CGPA was influenced by monthly stipend, marital status, and daily private study time on social media with p-values of 0.017, 0.000, and 0.000 respectively.