The purpose of this study was to examine the mediation and moderation effects of classroom peer norms (CPN) on the associations of teacher support with student engagement and aggression during sixth grade (first year of middle school). Current literature suggests that early adolescence is a critical developmental period with many changes occurring, including a peak in aggression, decreased academic engagement, increased peer influence, and decreased teacher support. Despite these challenging changes, teachers often serve as a powerful yet “invisible hand” (Farmer, Lines, & Hamm, 2011) that can influence student behaviors directly or indirectly via CPN. However, these two routes of teacher influence have rarely been examined simultaneously. The current study investigated these two routes by examining the partial mediation and moderation effects of CPN on the associations of teacher support with four student behaviors (overt and relational aggression, involved and disruptive behavior) among a diverse sample of 312 students from 32 classrooms across the fall and spring of sixth grade. Students self-reported teacher support (academic, emotional) and engagement (involved, disruptive behavior) and peer-nominated aggression (overt, relational). CPN was computed as a class average of a certain behavior (aggression or engagement). Multilevel modeling was used to test the mediation and moderation effects, considering the nested nature of the data (i.e., students nested within classrooms). Results indicated full mediation effect of CPN on the relations of fall teacher support with spring overt aggression, relational aggression, and disruptive behavior, and a moderation effect on spring involved behavior. Interestingly, the average classroom perception of teacher support showed a significant direct negative effect on relational aggression and disruptive behavior, and a direct positive effect on involved behavior, all of which became smaller and non-significant when CPN was taken into consideration (i.e., when the indirect or mediation effect was considered). In terms of the mediation effects, classes which had high levels of average perception of teacher support in the fall tended to have low disruptive behavior CPN, which was further associated with low levels of individual student disruptive behavior in the spring. The same mechanism applied to overt and relational aggression, except that the significance level of the relations of CPN with overt and relational aggression was marginal (i.e., p < .10). The moderation effect suggests that students who reported high levels of teacher support in the fall (relative to his/her classmates, regardless of the class average) tended to report high levels of involved behavior in the spring only if the fall involved behavior CPN was also high. These findings highlight the importance to investigate both teacher and peer influence within a classroom in order to better understand student behaviors during the first year of middle school. Theoretical and practical implications are discussed, as well as limitations and future directions.