Classroom Size and the Prevalence of Bullying and Victimization: Smaller classrooms sees more bullying

Classroom Size and the Prevalence of Bullying and Victimization: Testing Three Explanations for the Negative Association. Claire F. Garandeau, Takuya Yanagida, Marjolijn M. Vermande, Dagmar Strohmeier and Christina Salmivalli. Front. Psychol., September 20 2019. https://doi.org/10.3389/fpsyg.2019.02125

Abstract: Classroom size - i.e., the number of students in the class - is a feature of the classroom environment often found to be negatively related to bullying or victimization. This study examines three possible explanations for this negative association: (a) it is due to measurement effects and therefore only found for peer-reports (Hypothesis 1), (b) bullying perpetrators are more popular and have more friends in smaller classrooms (Hypothesis 2), (c) targets of bullying are more popular and have more friends in larger classrooms (Hypothesis 3). Multilevel regression analyses were conducted on a sample from Austria (1,451 students; Mage = 12.31; 77 classes) and a sample from the Netherlands (1,460 students; Mage = 11.06; 59 classes). Results showed that classroom size was negatively associated with peer-reported bullying and victimization in both samples, and with self-reported bullying and victimization in the Dutch sample only, suggesting partial support for Hypothesis 1. Students high in bullying were found to be more popular in smaller than in larger classrooms in the Austrian sample. The negative link between victimization and popularity was found to be stronger in smaller classrooms than in larger classrooms in the Dutch sample. However, classroom size was not found to moderate links between bullying or victimization and friendship in either sample. Hypotheses 2 and 3 were supported, but only for popularity and in a single sample. Further research is needed to better understand the higher prevalence of bullying found in smaller classrooms in many studies.

Introduction
The prevalence of bullying and victimization in classrooms is not merely the result of individual characteristics of the bullying perpetrators and their targets but is influenced by features of the classroom environment (Saarento et al., 2015). These contextual characteristics include the anti-bullying attitudes and behaviors of peer bystanders (Salmivalli et al., 2011) and of teachers (Veenstra et al., 2014; Oldenburg et al., 2015), as well as aspects of the peer social network, such as the degree of status hierarchy in the classroom (Garandeau et al., 2014). Classroom size - i.e., the number of students in the class - is a structural feature that has often been investigated in relation to academic achievement (see Finn et al., 2003), with smaller classrooms often found to be beneficial for academic performance (Hoxby, 2000; Shin and Raudenbush, 2011) and even earnings later in life (Fredriksson et al., 2013). Intuitively, we would expect the same advantageous effects of small classrooms on bullying. Smaller classrooms should logically protect against bullying thanks to higher adult/child ratios, allowing a more effective monitoring of children’s negative behaviors by school personnel.
Classroom size has been investigated in many studies on victimization and bullying, often as a control variable rather than a main predictor of interest. Surprisingly, very few studies found evidence of a protective effect of smaller classroom networks on bullying or victimization (Whitney and Smith, 1993; Khoury-Kassabri et al., 2004). The large majority of studies examining the link between classroom size and bullying or victimization found them to be either negatively associated (e.g., Vervoort et al., 2010) or unrelated (e.g., Thornberg et al., 2017). However, the reason why bullying and victimization would be more prevalent in smaller classrooms remains unclear.
The present study aims to test for three possible explanations for this negative association: First, the negative association may not reflect an actual social phenomenon but result from a measurement effect, related to the way peer-reported scores are computed. In this case, the prevalence-size link should be negative for peer-reported, but not for self-reported bullying and victimization (Hypothesis 1). Second, it is possible that bullying perpetrators enjoy higher status and are more socially connected in smaller classrooms, which in turn facilitates their bullying behavior. Engaging in bullying may be associated with higher perceived popularity and more friendships in smaller than in larger classrooms (Hypothesis 2). Third, victims may have less social support and fewer opportunities for friendships in smaller classrooms, which in turn could contribute to the maintenance of their victimization. Being victimized may be associated with lower perceived popularity and fewer friendships in smaller than in larger classrooms (Hypothesis 3). These hypotheses will be tested with large samples from two countries, using both self-reports and peer-reports of bullying and victimization.