Article
Predicting Adolescent Social Withdrawal Using a Random Forest Model
| e-ISSN | 2982-6845 |
| p-ISSN | 1226-4474 |
| Metrics | Views 0 / Downloads 0 |
Abstract
The purpose of this study is to identify key predictors of social withdrawal among adolescents. By incorporating various factors into a single model, the study aims to examine the relative importance of individual variables. Data from the Korean Children and Youth Panel Survey (KCYPS) were utilized, specifically the 6th-year data from the 2010 cohort (collected in 2015) and the 2018 cohort (collected in 2023). The analysis targeted high school seniors (2015: 1,857 students; 2023: 2,206 students). Adolescents in the top 10% for social withdrawal scores were classified as the at-risk group. Random forest analysis was employed to identify significant predictors. Results indicated that smartphone- related variables were among the most important predictors, followed by peer-related and community perception variables. Compared to 2015, the importance of smartphone- related factors significantly increased in 2023, while the importance of teacher relationships decreased. This study is meaningful in that it considers multidimensional predictors of social withdrawal within a single model. The study empirically confirms the increasing influence of smartphones and changes in adolescents’ social relationships following the COVID-19 pandemic. Additionally, The study highlights the importance of community and neighborhood perception in understanding adolescent social withdrawal.