Introduction: This study investigated the differential trajectories and relevant determinants of depressive symptoms in adolescents by following cohorts that included junior, senior, and vocational high school adolescents, over a 3-year period in Taiwan. Methods: Longitudinal data were obtained from 575 adolescents who participated in the Taiwan Adolescent to Adult Longitudinal Study. Data analysis included latent class growth with time-varying covariate, univariate, and multivariate analysis. Results: A three-class (“low but increasing trajectory,” “moderate and stable trajectory,” and “high but decreasing trajectory”) model fit the data of the cohort. Our findings indicated that 29%, 38%, and 33% of the adolescents were in the low but increasing, moderate and stable, and high but decreasing trajectories, respectively. After confounders were controlled for, bullying experiences were identified as a risk factor for depressive symptoms. The protective factors against depressive symptoms included resilience and peer and social support. Conclusions: The transitions between different educational stages critically influence the depressive symptoms of adolescents, and the adolescents follow different depressive trajectories, that have different etiology. Therefore, identifying adolescents at high risk for depression and designing student-centered intervention programs through individualized and multidimensional assessment of depressive symptoms are crucial for adolescents.

Original languageEnglish
Pages (from-to)879-892
Number of pages14
JournalJournal of Adolescence
Issue number5
Publication statusAccepted/In press - 2023


  • adolescent
  • depression
  • longitudinal studies
  • multivariate analysis

ASJC Scopus subject areas

  • Pediatrics, Perinatology, and Child Health
  • Social Psychology
  • Developmental and Educational Psychology
  • Psychiatry and Mental health


Dive into the research topics of 'Identifying differential trajectories and predictors for depressive symptoms in adolescents using latent class growth analysis: A population-based cohort study'. Together they form a unique fingerprint.

Cite this