Poster #224 - Bullying, Victimization and Prosocial Behaviors: Latent Class Analyses From Kindergarten to Third Grade

• In Event: 2-055 - Poster Session 06
  In Poster Session: PS 06 Section - Social Relationships

Integrative Statement

Though we know that bullying exists in preschool and early school years and that it can have devastating consequences, the bulk of the research still mainly focuses on older children and adolescents (Vlachou et al., 2013). However, this developmental period is crucial to understand how bullying begins and who is at risk. Existing studies usually find that children can be classified in four bullying-related groups: victim, aggressor, aggressor-victim and non-involved. These conclusions often come from either sociometric nominations or questionnaires with arbitrary cut-off scores. Using these methods, children experiencing moderate peer difficulties could fail to meet the arbitrary cut-off and avoid detection and needed help. Latent class analyses (LCA) overcome these shortcomings but has never been used with a preschool and early elementary sample, to our knowledge. Furthermore, while different types of aggression (i.e. indirect, physical, etc.) are often measured, results of LCA with older samples are contradictory. Some find no difference in the type of aggression used while others find different aggressor groups based on aggression type (Williford & Zinn, 2018). Including multiple types would allow us to see if these differences exist at a younger age and eventually, to adapt intervention program accordingly. Adding a prosocial behavior variable could help separate non-involved children from prosocial ones who could be included in intervention as defenders. Hence, this study aims to identify different groups related to bullying from kindergarten to third grade using teacher ratings, as they are the most reliable to detect bullying and victimization at this age (Vlachou et al., 2013).
Sample include 430 children from the Quebec Longitudinal Study of Child Development for whom teachers accepted to participate at least once in kindergarten, first or third grade. This populational sample is urban, non-clinical and was recruited randomly. Teacher answered questions about children’s proactive, indirect (relational) and physical aggression as well as victimization and prosocial behaviors. Separate LCAs for each grade were done using Mplus 7.3.
LCA results revealed that the three-category model provided the best fit in kindergarten (see table 1) and included prosocial (lowest aggression and victimization, highest prosocial behaviors, see figure 1a), average (mean aggression, victimization and prosocial behaviors) and aggressor-victim (highest aggression and victimization, lowest prosocial behaviors) categories (see figure 1a). The four-category model best fit the first and third grade data (see table 1). A victim category was added to those existing in kindergarten, characterized by low prosocial behaviors and aggression and high victimization (see figure 1b-c). As aggressive behaviors are more frequent in early childhood and diminish with age, the apparition of this category in first grade could be explained by the natural decrease of aggression in certain children. However, victims don’t seem to develop prosocial behaviors and thus remain or become target of still-aggressive children. We did not find an aggressor-only category or different categories based on aggression type. The most aggressive children of the sample were also the most victimized, suggesting an important bidirectionality. Over-time transition between classes and association with teacher-student relationship and student engagement will be done.

Authors

• Sophie Chaput-Langlois, University of Montreal
  Presenting Author

• Michelle Pinsonneault, University of Montreal
  Non-Presenting Author

• Sophie Parent, Université de Montréal
  Non-Presenting Author

• Natalie Castellanos Ryan, Université de Montréal
  Non-Presenting Author

• Richard Ernest Tremblay, University of Montreal
  Non-Presenting Author

• Jean R. Séguin, Université de Montréal
  Non-Presenting Author