Abstract
Researchers have increasingly started to pay attention to how contextual factors, such as the classroom peer context and the quality of student–teacher interactions, influence children’s aggressive behavior. This longitudinal study was designed to examine the degree to which benefits and costs of different teaching practices (child-centered and child-dominated) would be dependent on the initial peer-group composition (aggregate levels of aggression and victimization at the beginning of first grade). Teachers provided ratings of aggression and victimization (N = 523 first-grade students; M age at the beginning of first grade = 7.49 years, SD = 0.52). Information about different teaching practices was obtained via observations. Our results show that whereas child-centered practices are beneficial in high-victimization classrooms, child-dominated practices inhibit the development of aggression in low-victimization classroom contexts. Our findings highlight the importance of moving beyond main-effect models to studying how different contextual influences interact to promote, or inhibit, the development of aggression.
ACKNOWLEDGMENTS
We thank Ernest V. E. Hodges for his feedback on earlier versions of this article. Our special gratitude goes to all the participants who made this study possible.
FUNDING
This study was supported by the Estonian Research Council Grant IUT03-03.
Notes
1 As classroom means are somewhat ambiguous indices, we correlated classroom means for aggression and victimization with number of aggressors and victims (number of children in each classroom who scored above the 75th percentile; for aggression, the score was greater than 2.3, and for victimization, the score was greater than 2). Percentage of aggressors varied from 0 to 67% and percentage of victims varied from 0 to 56% across classrooms. Number of aggressors and mean levels of aggression correlated at .95, and number of victims and mean levels of victimization correlated at .90.
2 We also tested whether class-average aggression, as well as interactions with teaching practices, would become significant if class-average victimization was not part of the models. In the main-effect model, class-average aggression did indeed promote the development of aggression from fall to spring (unstandardized estimate = 0.327, SE = 0.147, p = .026), 95% CI [0.038, 0.616]. However, neither of the two interactions with teaching practices reached statistical significance.