Associations Among Social Media Use, Sleep Duration, and Mental Health Indicators in a Synthetic Adolescent Dataset

Abstract

This study investigated the relationship between hours of social media using and hours of sleep duration, anxiety level, and stress level in a synthetic dataset containing 1,200 adolescent records. All four primary hypotheses were tested using Pearson correlations. A multiple linear regression analysis helped to examined whether daily social media use hours and screen time before bed were related to sleep duration, while accounting for age, gender, and physical activity. A one-way ANOVA compared sleep hours across four quartile-based social media use groups. The four primary correlations were approximately zero and statistically nonsignificant (rs = −.009 to .031, all ps > .28). The regression model explained almost none of the variance in sleep duration (R² = .001, p = .895), and average sleep duration did not differ significantly across the four usage groups (p = .348, η² = .003). The additional analyses of addiction and social interaction also found no significant relationships in this dataset. However, because the records were synthetically generated and the mental health measures were not documented as validated instruments, the results cannot be generalized to real adolescents. The study suggests the importance of evaluating dataset construction and measurement validity before using the findings to evaluate psychological theories.