Classifying Latent Classes and Testing Key Predictors of the Trajectory of Emotional Problems in Adolescence

Abstract

This study was conducted to classify latent classes of emotional problems in adolescence and to explore and test key predictors of change in emotional problems. To this end, latent class growth analysis was applied to data from the first through fifth waves of the middle school students’ panel of KCYPS 2018. SEM forest was used to explore the top ten most important variables explaining change in emotional problems, and the three-step approach was applied to examine how key predictors affect the classification of latent classes. The main findings are as follows. First, three latent classes in emotional problems during adolescence were classified: low-level decrease followed by increase (20.5%), middle-level maintenance (58.8%), and high-level maintenance (20.7%). Second, SEM forest analysis revealed that six variables had been studied in previous research while four other variables were newly identified. Third, happiness, self-esteem, and grit (consistency of interest) were associated with lower levels of emotional problems, while academic helplessness (lack of active performance), smartphone dependence, and parenting attitudes (inconsistency) were associated with higher levels of emotional problems. Based on the results, implications and suggestions for reducing emotional problems among adolescents are discussed.