Mapping adolescent problematic social media use patterns across 41 countries/regions: A multilevel latent class analysis with social determinants

Abstract

Background

Problematic social media use (PSMU) represents a growing concern among adolescents globally. While existing variable-centered research has provided valuable insights into PSMU determinants, person-centered approaches can offer complementary perspectives by identifying behavioral heterogeneity within populations and examining how social determinants differentially affect distinct subgroups across diverse national contexts.

Objectives

This study aimed to (1) identify distinct PSMU behavioral classes at the individual level, (2) classify countries/regions based on PSMU prevalence patterns, and (3) examine how multilevel social determinants predict class membership.

Methods

Data were from 171,447 adolescents across 41 countries/regions. Multilevel latent class analyses were first conducted on nine dimensions of PSMU to identify distinct classes at individual and national levels. Individual-level measures incorporated social relationship quality, health behaviors, and economic deprivation. National-level indicators included economic development, education, income, inequalities, and cultural values. Multinomial regressions were performed to examine associations between social determinants and PSMU class membership.

Results

Three distinct individual-level PSMU classes emerged: Low Problematic Use Class (58.0%), Moderate Problematic Use Class (37.6%), and High Problematic Use Class (4.3%). Countries/regions are clustered into three categories: Low Prevalence Region (31.7%), Moderate Prevalence Region (39.0%), and High Prevalence Region (29.2%). At the individual level, positive social relationships and healthy lifestyles demonstrated protective effects against problematic use, while economic deprivation increased risk. At the national level, educational inequality, secular values, and gender inequality significantly increased the likelihood of countries belonging to the High Prevalence Region.

Conclusion

Adolescent PSMU manifests heterogeneously across individuals and exhibits geographic variation. This study highlights the necessity of multilevel, differentiated intervention strategies, emphasizing that countries/regions should develop policies aligned with their specific characteristics to foster supportive environments for adolescent digital well-being.