Journal of Abnormal Child Psychology, Vol. 33, No. 3, June 2005, pp. 283–292 ( C  2005) DOI: 10.1007/s10802-005-3565-5 A Model of Contagion Through Competition in the Aggressive Behaviors of Elementary School Students Keith Warren, 1,5 Susan Schoppelrey, 2 D. Paul Moberg, 3 and Marilyn McDonald 4 Received January 13, 2004; revision received July 15, 2004; accepted August 25, 2004 This article extends the work of Kellam, Ling, Merisca, Brown and Ialongo (1998) by applying a mathematical model of competition between children to peer contagion in the aggressive behaviors of elementary school students. Nonlinearity in the relationship between group aggression and individual aggression at 2-year follow-up is present. Consistent with the findings of Kellam et al. (1998), hierarchical linear modeling indicates that the relationship is statistically significant for those students whose initial parental ratings of aggressive behavior were above the sample median. In the context of competition between students, the behavior of initially aggressive students may be negatively reinforced. Lowering aggression in the school environment may therefore be the most effective way to lower the level of these students’ aggressive behavior. KEY WORDS: aggressive behavior; ecological psychology; elementary school; students; peers; red queen model. The last decade has seen an increased interest in peer contagion as a factor in the aggressive and delin- quent behaviors of children in elementary and middle school (Anderson, 1999; Dishion, McCord, & Poulin, 1999; Dishion, Poulin, & Burraston, 2001; Ialongo, Poduska, Werthamer, & Kellam, 2001; Ialongo et al., 1999; Kellam, Ling, Merisca, Brown, & Ialongo, 1998; Patterson, Dishion, & Yoerger, 2000; Snyder, Horsch, & Childs, 1997). Peer contagion is of both theoretical and pragmatic interest, since a better understanding of the dy- namics of the spread of aggressive behavior could guide the design of programs that aim to attenuate the effects of an aggressive environment (Ialongo, Poduska, Werthamer, & Kellam, 2001). A persistent finding in the peer contagion literature is that aggregation with peers who display aggressive 1 College of Social Work, The Ohio State University, Columbus, Ohio. 2 School of Social Work, University of Illinois at Urbana-Champaign, Urbana, Illinois. 3 Department of Population Health Sciences, University of Wisconsin- Madison, Madison, Wisconsin. 4 Wisconsin Center for Educational Research, University of Wisconsin- Madison. 5 Address all correspondence to Keith Warren, The Ohio State University, College of Social Work, Stillman Hall 325Q, 1947 College Road, Columbus, Ohio 43210; e-mail: warren.193@osu.edu. and delinquent behavior is particularly problematic for children who are themselves at risk for aggressive and delinquent behavior (Dishion et al., 1999; Kellam et al., 1999). This finding even extends to very young children. In their study of first-grade children exposed to classroom aggression, Kellam et al. (1999) found an interaction ef- fect in which exposure to aggressive classrooms lead to increased aggression in middle school in the most aggres- sive elementary school children, but had little effect on others. Snyder et al. (1997) found that preschool children who associated with aggressive peers at school were likely to show increased aggression 3 months later. Why does exposure to aggressive peers worsen the behavior of those children who already behave aggres- sively? One possibility lies within the model developed by researchers at the University of Oregon, which suggests that aggressive children aggregate into peer groups that positively reinforce aggressive and delinquent behavior (Dishion et al., 2001; Reid, Patterson, & Snyder, 2002). Kellam et al. (1999, p. 183) seem to take this position, sug- gesting that, “ ... the experience of the aggressive child in aggressive first grade classrooms sets the pattern of the child’s behavioral responses ... as well as member- ship in poor behaving peer groups and lack of attach- ment to school.” On the other hand, Snyder et al. (1997, p. 154) observe that, “Disagreeable, coercive behavior 283 0091-0627/05/0600-0283/0 C  2005 Springer Science+Business Media, Inc.