The social networks in which children participate in are strongly associated with their involvement in bullying and defending (Juvonen & Graham, 2014; Salmivalli, 2010). It is likely that peer effects – referring to selection and influence processes – explain this association. Children seek out friends who are similar to them in terms of their level of behavior and these friends then serve to reinforce their behavior over time. Research suggests that popular individuals might be more influential than others in shaping their peers’ bullying and defending at school. Yet, few studies have empirically tested this assumption. Moreover, few studies have provided insight into the extent to which classroom characteristics affect selection and influence processes in bullying and defending. To this end, the aims of this study was to investigate 1) whether selection and influence explain bullying and defending behavior, 2) whether popular individuals are more influential than others, and 3) whether these effects differ depending on the way classrooms are structured. Students typically interact with each other in classrooms where they share the same grade level, so-called single-grade classrooms. Increasingly, however, students interact with others from different grades in so-called multi-grade classrooms (Veenman, 1995). From a power imbalance perspective (Rodkin et al., 2015), it can be argued that, compared to single-grade classrooms, multi-grade classrooms produce hierarchical group structures based on larger age differences between the children in such classrooms. Therefore, in this classroom context, we expected stronger peer norms favouring bullying, and accordingly stronger selection and influence effects regarding bullying and weaker selection and influence effects regarding defending. Alternatively, from an evolutionary perspective (Ellis et al., 2012) it can be argued that mixed-age settings rather than age-segregated settings are the natural environment for child development, and, as such, single-grade classrooms rather than multi-grade classrooms dysregulate the natural enviroment, promoting bullying and demoting defending in children via selection and influence processes in this classroom context instead. Data were derived from KiVa, a school program aimed at enhancing positive group processes among children from grades 3-6 in elementary education (8-12 years; 50% boys) in the Netherlands. For the current study, only data from the control schools were used, collected over two waves within one school year. 60 single-grade (1,481 students) and 34 multi-grade (820 students) classroom friendship networks (Who is your best friend?) are available. Analysis were conducted separately for single-grade and multi-grade classrooms with longitudinal multi-level Siena analysis (Koskinen & Snijders, 2015). Table 1 shows that similarity in bullying and defending between friends was explained by selection and influence in both single-grade and multi-grade classrooms. However, in single-grade classrooms, not multi-grade classrooms, friends’ popularity significantly moderated peer influence in bullying and defending. The preliminary findings indicate that single-grade and multi-grade classrooms are similar in the extent to which bullying and defending can be explained by selection and influence processes, and that social standing among peers plays a stronger role in single-grade classrooms than in multi-grade classrooms. This suggests that individuals who occupy central positions should be targeted differently in these two classroom types.