In adolescence, bullying victimization is typically represented in terms of a three-fold factor structure reflecting three components of verbal, physical, and social victimization. Recent studies have suggested the usefulness of alternativte models including both general and component-specific factors. In this study, we assessed the empirical and theoretical validity of an instrument assessing verbal, physical and social victimization using a set of alternative models of victimization: a unidimensional model, a three-factor model, and a bifactor model. Association between emerging factors and student variables were explored to establish theoretical fit of the models. Sample consisted of upper primary and lower secondary school students [N = 1311; 53% Male; Mean age (SD) = 10.73 (1.45)] and their teachers. The three factor and bifactor models showed good fit. In spite of acceptable fit, the unidimensional model showed lower empirical support when compared with the other models. The dimensions of the three-factor model showed similar associations with most student variables, while the bifactor showed more heterogeneous, and theoretically coherent associations. General victimization decreased with age and was positively related with externalizing and internalizing symptoms, student-teacher conflict and negative expectations. Verbal victimization showed increased prevalence among girls and older students. Physical victimization showed increased prevalence among boys and younger students, and positive associations with externalizing symptoms and student-teacher conflict. Social victimization was more frequent among girls, and positively related with internalizing symptoms and negative expectations toward teachers. These findings highlight the usefulness of modeling victimization using both general and form-specific dimensions for both assessment and theory-building purposes.
Keywords: bi-factor analysis; bullying; dimensional structure; factor analysis; victimization.