ORIGINAL ARTICLE Relations between close friendships and adolescent aggression: Structural and behavioral friendship features* Naomi C. Z. Andrews 1 | Laura D. Hanish 2 | Dawn DeLay 2 | Carol Lynn Martin 2 | Kimberly A. Updegraff 2 1 York University/Mothercraft 2 Arizona State University Correspondence Naomi Andrews, Early Intervention Department, Mothercraft, 860 Richmond St W, Unit 100, Toronto, ON, Canada, M6J 1C9. Email: naomi.andrews@mothercraft.org Funding information T. Denny Sanford Foundation; T. Denny Sanford School of Social and Family Dynamics; Social Sciences and Humanities Research Council of Canada Abstract Friendships have the potential to perpetuate or mitigate youth’s aggressive behavior. We investigated concurrent and longitudinal effects of friendships on aggression by examining both structural (size and interconnectedness of the local friendship network) and behav- ioral (friends’ aggression) friendship features. Participants were 868 sixth to eighth grade middle-school students (M 5 12.10 years; 49.9% girls; 44% Latina/o) who completed questionnaires at two time points. Participants nominated their friends; reciprocal friendship nominations were used to calculate structural friendship group features (size and interconnectedness). Peer nominations were also used to measure youth’s and their friends’ aggression. Having more reciprocal friends was associated with more aggression concurrently (particularly for youth whose friends were highly aggressive), and having an intercon- nected friendship group was associated with decreased aggression over time. Given that findings were different for group size and inter- connectedness, we discuss the unique importance of each of these structural friendship features. Practical implications regarding the potential to decrease aggressive behavior based on specific friendship features are also discussed. KEYWORDS aggression, friendship, adolescence 1 | INTRODUCTION Aggression is used widely during adolescence, despite the harm associated with aggressive behavior (Mishna, Wiener, & Pepler, 2008; Pepler, Jiang, Craig, & Connolly, 2008). Though rates of aggression are relatively high during *This article is based on a dissertation submitted in partial fulfillment of the requirements for Naomi Andrews’ doctoral degree at Arizona State University. Social Development. 2017;1–15. wileyonlinelibrary.com/journal/sode V C 2017 John Wiley & Sons Ltd | 1 Received: 23 January 2017 | Revised: 28 November 2017 | Accepted: 30 November 2017 DOI: 10.1111/sode.12277