Children's classroom engagement is important for their learning and academic achievement. Extending Kindermann's (2007) study of peer influence on adolescents' engagement to an ethnically homogeneous sample, the current study examined how different aspects of diversity affect the peer influence process. Three types of diversity were considered: ethnic diversity existing at the school level, relational diversity at the peer network level, and motivational diversity at the group level. Ethnic diversity was observed in the student body as well as among school teachers and staff. Relational diversity was measured by z-scores resulting from binomial tests reflecting how closely two pair of individuals were connected in the peer network. Finally, motivational diversity was measured as the dispersion (SD) around peer group mean engagement levels, thereby reflecting the diversity of engagement within each peer group. The results indicated that adolescents in this ethnically diverse middle school were overall highly engaged; their engagement patterns were comparable to previous findings from homogeneous samples consisting largely of European American adolescents. Also consistent with prior findings, the mean engagement levels of students' peer group members were a significant predictor of changes in adolescents' own engagement, which suggests peer influence on adolescents' classroom engagement. Although previous literature suggests that individuals in diverse settings tend to be less well connected to one another, the adolescents in this ethnically diverse school were well connected with their peers. Unexpectedly, almost all students' peer groups were ethnically diverse. When the impact of relational diversity was examined to see whether strongly connected individuals exerted more influence on each other than weakly connected individuals (differential influence hypothesis), the results indicated that the strength of connections among peer group members did not appear to play a significant role in the magnitude of their influences on each other's changes in engagement. Nevertheless, the present study suggested new pathways and methods to examine differential peer influences. Finally, the impact of motivational diversity of peer groups was examined using a moderated model based on an interaction effect between peer group motivational diversity and individuals' initial engagement. The results indicated that the positive impact of peer group motivational diversity was moderated by individuals' initial engagement status, such that initially low engaged adolescents benefited from diversely engaged peer groups, whereas peer group motivational diversity had a comparatively small negative effect on initially highly engaged students.

The school environment is one of the primary contexts for children's social, emotional and cognitive development. While teachers are likely to be primarily focused on students' motivation and learning, for adolescents, one of the most enjoyable and important aspects of school life is likely to be centered around the time spent interacting with peers. It is well recognized that peers socialize one another but although many studies have examined the influence of peers on adolescents' risky behaviors far fewer have focused on the influence peers may have on individuals' positive behaviors. As a result this study focuses on academic development replicating previous research designed to examine whether peer group affiliation has an effect on student academic engagement. A cohort of 343 seventh grade students, primarily Caucasian, 52% male, was followed for a period of one school year. Teachers reported on students' academic engagement in the fall and again in spring using a 14-item scale (Wellborn, 1991), and students reported on their teachers' and parents' involvement in fall using 8- and 4-item scales respectively. Student grades were collected from school administrative records. To identify individual student's network affiliations socio-cognitive mapping procedures were used (Cairns, Perrin & Cairns, 1985), and then peer group profiles of engagement were calculated based on the average rating of engagement across each individual's affiliates. During the academic year peer group membership turnover was 49%, despite this, the quality of peer group profiles of engagement remained similar from fall to spring. Groups also tended to be and remain motivationally homogenous across the year. In general, girls' networks tended to be more highly engaged than boys' and networks that were more highly engaged tended to be more stable across the year. Structural equation modeling was used for the major analyses to assess whether peer group academic motivation in the fall could predict individual motivation in the spring. The results indicated that while controlling for individuals' earlier engagement, as well as for processes of group selection and parent and teacher influences, the quality of individuals' peer group engagement in the fall was significantly predictive of students' later engagement in the spring. It should be noted that within the major models academic performance was also strongly related to later engagement. While this study provides further evidence to underscore the importance of the peer group in the socialization of students' academic motivation, particularly when one considers the snowballing effects in motivation this influence may have across a student's entire academic career, it also illustrates the important role performance may play in academic motivation for young adolescents.

Academic engagement has been found to significantly predict students' future achievement. Among adolescents, the peer context becomes an increasingly important point of socialization and influence on beliefs and behavior, including academic engagement. Previous research suggests that those peers with whom an adolescent spends much of their time significantly predict change in engagement over time (Kindermann, 2007). Bronfenbrenner's ecological systems theory (Bronfenbrenner & Morris, 1998) postulates that exosystem effects (those influencing factors that are not directly connected to individuals) play an important role in development, and social network theorists have suggested that the position one occupies within the greater network is a key factor that determines one's power of influence (Borgatti, 2005). An individual's own position in a network emerges from his or her own connections, as well as from the structures formed by the connections of his or her affiliates (the exosystem). Utilizing an existing dataset, social networks analysis techniques were used to examine how three different forms of centrality (degree, closeness and eigenvector), which are markers for micro- and exo-system effects, relate to classroom engagement and its change over time. Results showed that although centrality in a network is positively related to academic characteristics at one point in time, students who have large numbers of immediate connections (degree centrality) tend to decrease in engagement over time. In contrast, eigenvector centrality showed a positive interaction with peer group influence on change in engagement over time. For those students who had highly interconnected peers the positive effect of peer group engagement was increased.