Latent class/profile analysis in research on school bullying behavior patterns: a systematic review.

BACKGROUND

School bullying is a major public health concern affecting adolescents worldwide. Traditional cutoff point and outlier approaches have inherent limitations in identifying bullying behavior patterns and fail to capture the heterogeneity of bullying roles and the co-occurrence of bullying types. Latent class/profile analysis (LCA/LPA), as a person-centered analytical approach, offers a novel pathway to deepen understanding of complex bullying behavior patterns.

METHODS

A systematic search of PubMed, Web of Science, CNKI, and Wanfang databases identified 1,071 records. Original empirical studies using LCA/LPA to investigate school bullying patterns were included. Information on study design, sample characteristics, model fit indices, class labels, and associated variables was extracted.

RESULTS

A total of 113 studies were included. LCA/LPA identified three core typological features in school bullying patterns: heterogeneous patterns of victims, heterogeneous patterns of bullies, and complex subtypes of bully-victims. Research has also extended to the co-occurrence of bullying with other behaviors (substance use, discrimination experiences, adverse childhood, etc.), as well as the differential associations of risk factors (gender, age, social support, family factors, etc.) and negative health outcomes with distinct bullying patterns. Bully-victims were further differentiated into multiple subtypes and exhibited the most severe internalizing and externalizing problems.

CONCLUSION

LCA/LPA effectively reveals the heterogeneous patterns of school bullying and provides empirical evidence for the development of individualized intervention strategies. Future work should unify measurement tools, cover more bullying subtypes, and adopt longitudinal designs to track pattern changes over time.