Targeting Multiple Sources of Peer Influence in Prevention Programs: A Social Network Analytic Approach

Introduction: Most universal school-based prevention programs include a general focus on peer relationships, and social network analysis (SNA) is emerging as an innovative approach for clarifying how peer relationships impact behavior. SNA has a number of advantages when compared to the traditional approach to measuring peer relationships: (1) each child self-reports on his/her relationships and behavior, providing more accurate estimates of peer similarity; and (2) children report on multiple sources of influence (e.g., friendships, popular peers), enabling researchers to disentangle the sources of peer influence on behavior. This study uses SNA to examine the relative effects of reciprocated and unreciprocated friendships’ influence on prosocial behavior and academic achievement.

Methods: Participants were 3,471 children (48% girls, 41% minority) in 156 1^st, 3^rd, and 5^th grade classrooms participating in the Classroom Peer Ecologies Project. Student and teacher surveys were collected at three measurement time points: the first six weeks of the school year (W1), approximately 2 months later (W2), and the end of the school year (W3). Prosocial behavior (e.g., “this child is helpful to peers”) and academic achievement (e.g., “this child is good at reading”) were each measured at W1 and W3 using teacher-report. Reciprocated friends’ and unreciprocated nominated peers’ academic achievement and prosocial behavior were each measured at W1. Students nominated classmates (unlimited nominations) in response to the question “who is your friend?” If two children nominated one another, they were considered to have a reciprocated friendship, otherwise, it was classified as unreciprocated. Mean level of teacher-reported academic achievement and prosocial behavior for reciprocated friends and unreciprocated nominated peers were calculated for each student.   

Results:  Models controlled for gender, race/ethnicity, grade, and the W1 score on the outcome of interest. After controlling for covariates, reciprocated friends’ (B = 0.09, p < .05) and unreciprocated nominated peers’ (B = .10, p < .05) prosocial behavior significantly predicted W3 prosocial behavior. Additionally, reciprocated friends’ levels of academic achievement significantly predicted W3 academic achievement (B = 0.07, p < .05).  Unreciprocated nominated peers’ academic achievement was not significantly related to W3 academic achievement (p > .05). Interactions with gender and grade were tested and were not significant (all p’s > .05). 

Conclusions: This study suggests that reciprocated friendships are important sources of both prosocial and academic influence, whereas unreciprocated friendships may only influence prosocial behavior. Findings indicate the importance of targeting multiple sources of peer influence in programs that aim to promote prosocial behavior and academic achievement.