<![CDATA[Background. Bullying represents a multifaceted and heterogeneous social phenomenon. Purpose. Within this context, the present study seeks to identify the heterogeneity of student involvement in bullying situations through latent profile analysis (LPA). Method. Data were collected from 847 junior and senior high school students in East Java, aged 12 to 18 (M=15.01, SD=0.81). The measurement utilized five indicators of roles in bullying: perpetrator, accomplice, victim, defender, and bystander (passive observer). Latent class analysis was conducted using the Python programming language via Google Colaboratory, taking into account model fit indices such as the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), entropy values, and minimum class size. Results. As a result of this analysis, Latent class analysis successfully identified six profiles (entropy value=0.779): two profiles involving active participation in bullying (Profile 1 – aggressor-reinforcer at 11%, Profile 5 – pure bully at 12.2%); three normative profiles with low to no involvement (Profile 2 – normative low involvement at 14.5%, Profile 3 – Truly Normative at 25.7%, Profile 4 – Normative Non-Involved at 22.9%), and Profile 6 (Victim-Outsider), which has a prevalence of 13.7% and is characterized by high levels of victimization and passive responses. Conclusion. These findings indicate that the role of bullying can manifest in diverse forms. This study underscores the importance of understanding bullying as a heterogeneous phenomenon based on profiles and supports the development of more specific, student-needs-based bullying interventions.]]>
