Abstract
BACKGROUND: Loneliness in children has been a major topic of interest in both clinical and developmental psychology. Further studies to investigate predictors of loneliness are needed for educational practices. METHODS: A total of 1088 late elementary school-aged children (48.81% boys, M(age) = 10.35) participated in a 1-year longitudinal survey. We used hierarchical linear modeling and mover-stayer latent transition analysis. DISCUSSION: Findings from the variable- and person- centered approaches suggested that less positive peer relations, higher victimization, and higher relational aggression are predictors of higher future loneliness. Meanwhile, both approaches did not reach an agreement concerning predictors to reduce loneliness. This result highlighted a utility of a combined approach and sounded an alarm for overreliance on the variable-centered approach dominating child research. CONCLUSION: To protect young children from loneliness, it will be more beneficial to prevent the development of loneliness rather than to apply interventions to reduce loneliness once established. Preventive practices need to be implemented to protect children from loneliness.