Individual- and School-Level Predictors of Latent Profiles of Bullying Victimization : Comparison across Korea, Taiwan, and USA

Background/purpose:
Significant experiences of bullying victimization in childhood most likely hinders a normative transition into adulthood. Most existing studies on bullying victimization have examined the relationship between bullying victimization and negative developmental outcomes. There is increasing demand for research that guide direction for targeted intervention by identifying distinct latent groups of victimization experience using a person-centered approach and by exploring uniqueness/commonalities in victimization experience across countries through international comparative research. Additionally, it is important to examine which individual- and school-level variables predict such latent profiles of victimization experience.

Methods:
Data was obtained from the 2015 Trends in International Mathematics and Science Study. The analytic sample included students from Korea (n = 4,586; n = 5,306), Taiwan (n = 3,846; n = 5,702), and USA (n = 8,278; n = 8,580) in grades 4 and 8, respectively.
Latent profile analysis (LPA) was conducted based on 8 items (for grade 4) or 9 items (for grade 8) of the school bullying questionnaire. Descriptive differences across latent profiles by age and country were examined.
Multilevel logistic regression was conducted using LPA results as dependent variables. Independent variables include both individual-level (gender, material goods, parent’s education, absence, interest in math/science) and school-level (neighborhood affluence, school resources, delinquency, educational interest) factors.

Results:
Four to five bullying victimization latent profiles were identified for each country and each grade.
Common across different age groups and countries was that the low risk group comprises the majority.
Developmental differences were also observed: In grade 4 only, most individuals experienced all types of victimization to a similar extent, with the exception of a few individuals who experienced a spike in only 1-2 types. Also ‘making fun of’ victimization type was most common for grade 4, but relational victimization types such as sharing and posting embarrassing information seemed more prevalent in grade 8.
Additionally, there were differences across countries: Compared to other countries, grade 8 students in Korea experienced a concentration of selected victimization types. Additionally, grade 4 in Korea and Taiwan reported generally low- to moderate-levels of victimization across eight victimization types, while grade 4 students in the US contained a group with high-levels for all types.
Multilevel logistic regression indicated that being female, less delinquency at the individual- and school-level, and greater academic interest at the individual- and school-level increases the probability of being in the low risk group, compared to other risk groups.

Conclusions:
The current study identified that there were both similarities and differences of bullying victimization types in age groups and cultural background. Such findings support continued and expanded efforts to create developmentally and culturally sensitive intervention for victims of school bullying. Results suggesting that individual- and school-level predictors are similar across all risk groups, compared to the low risk group, address the need for social work professions to pay attention to all children who experienced any type of and any extent of bullying victimization.