846 PAULETTL COOPER.AND PERRY tities,and the like- use the th ries depart in critical wavs ch o examine cognitive pathw to children's har ssment o the that play con and of thei vior (e.g..agg ab his or her situations they ensitive to and to react re psych evant to them theories do n meas sive thar irections of influe aple it is cono able that children' who d partic it is clea hich the nd stable personal profiles indicatin nt gende-aypica pee ers,that is,to di ect more aggre children'sggre sion in the ions.To dete severity of children' dren aggression in sc on toward their victim e nay be child" (ehave simpositivethn-child bet).but one child hild'smv (ac ations is the child's aggress sign a children diffe )re istinc address the impact of social identity only on the former. fter the ative influences of situatio and trai erences acr Characteristics of Gender-Nonconforming Children on the cog here some effort identify er-atyp with bility in ag bou t the children whose victimiz the focus of stu ion partn (e.g..pee vs. or is pred othe tud than hy 200 Zakriski et al 2005).These data support Four ing ch nsk th o m)if the a problematic form of tity (high al with who one is inter ty)th for s dfcrcentiationorfeldisconmcicdiomthcrow a central ten of cogniti ender may be particularly distresse Method ing this issue by examinine whether children's coonitive struc Participants and Procedure to aggress more toward All chilren n the oys (M =10.2 ears)and 94 girls (Ma e=10.0ye gender atypical as a situationa tities, and the like—to their social behavior. Because our research questions and paradigm were stimulated by these theories and because the theories depart in critical ways from other approaches to investigating human aggression, we describe certain features of the theories and tell how they guided the present study. Cognitive theories emphasize that many persons display con￾siderable situational specificity in a social behavior (e.g., aggres￾sion) owing to interactions between their relatively enduring pro￾cessing structures and features of the situations they encounter. That is, people’s unique cognitive structures cause them to be sensitive to, and to react in particular ways to, aspects of situations that are psychologically relevant to them. These theories do not deny that aggression also possesses the qualities of a personality trait (i.e., that some people are generally more aggressive than others), but they suggest that much can be learned about aggressive motivation by studying how individuals organize their aggression around particular eliciting circumstances. Although data are still accumulating, it is clear that many children do possess unique aggression signatures (Mischel & Shoda, 1995)— distinctive and stable personal profiles indicating how their aggression varies over different situations (relative to other children’s aggression in the same situations). To determine children’s signatures, children are observed for aggression in sev￾eral situations. For each situation, each child’s aggression is con￾verted to a z score indicating the child’s aggression relative to other children, thus removing the nomothetic influence of the situation. A child’s idiographic profile of z scores across the situations is the child’s aggression signature. The child’s average z score is a measure of trait aggression (Fournier, Moskowitz, & Zuroff, 2008; Mischel & Shoda, 1995). Many children differ markedly across situations in their aggression, and this variability cannot always be accounted for by trait aggression. In other words, after the normative influences of situations and trait aggression are removed, considerable intraindividual differences across situations (signatures) often remain to be explained. There has been some effort to identify the salient features of situations that organize within-child variability in children’s ag￾gression. Chief among these situational features are the character￾istics and behaviors of a child’s interaction partners (e.g., peer vs. adult, male peer vs. female peer, liked peer vs. disliked peer; Coie et al., 1999; Hodges, Peets, & Salmivalli, 2009; Matthys, Maassen, Cuperus, & Van Engeland, 2001; Mischel & Shoda, 1995; Peets, Hodges, & Salmivalli, 2011; Rodkin & Berger, 2008; Veenstra et al., 2007; Zakriski et al., 2005). These data support Fournier et al.’s (2008, p. 533) observation that “the most salient psycholog￾ical features of the situation are found in the behavior of the individual with whom one is interacting.” Although there has been progress in identifying some features of situations (e.g., target characteristics) that underpin within-child situational variability in aggression, a central tenet of cognitive theories—the idea that individual differences in within-child situ￾ational variability in aggression are predictable from enduring cognitive processing structures— has, to our knowledge, yet to be investigated. The present study illustrates a paradigm for research￾ing this issue, by examining whether children’s cognitive struc￾tures predict their tendency to aggress more toward gender￾atypical peers than toward other peers. Thus, we conceptualize the degree to which a child’s interaction partner is characteristically gender atypical as a situational variable that interacts with chil￾dren’s processing structures to affect aggression. Our paradigm therefore is one of Person
Target interaction. To examine cognitive pathways to children’s harassment of gender-atypical peers, we used multilevel modeling. In both the fall and spring of a school year, we collected measures of children’s cognitions, of their aggression toward each classmate, and of their gender-atypical behavior. For each child, a beta coefficient (slope) was calculated indicating how the child’s aggression toward his or her classmates changed over the school year as a function of the class￾mates’ gender-atypical behavior in the fall (within-subjects analyses). To test the hypotheses, these within-child slopes were predicted from measures of participants’ cognitions (between-subjects analyses). The longitudinal design was used to help rule out certain alternative directions of influence. For example, it is conceivable that children’s cognitions are reactive to their aggression toward gender-atypical others (e.g., boys who attack boys for playing with girls may after￾ward feel more gender typical or more dominant). It is important to note that our hypotheses address only the degree to which the cognitive variables affect children’s tendencies to dis￾criminate against gender-atypical peers, that is, to direct more aggres￾sion toward gender-atypical peers than toward gender-typical peers. Our hypotheses do not address the absolute severity of children’s aggression toward their victims. Two children may be similarly in￾clined to victimize gender-atypical peers more than gender-typical ones (i.e., have similar positive within-child betas), but one child may aggress generally (across all peer targets) at a higher level than the other and therefore show a more generally elevated pattern (i.e., a higher intercept). Sensitivity to a target feature (captured by slope) and average aggression across targets (i.e., trait or general aggression, captured by intercept) are distinct constructs, and our hypotheses address the impact of social identity only on the former. Characteristics of Gender-Nonconforming Children Although the focus of this investigation was on the cognitive determinants of children’s aggression toward gender-atypical peers, the measures we collected afforded an opportunity to learn more about the children whose victimization was the focus of study— gender-nonconforming preadolescents. We describe the self-concepts and social-behavioral qualities of these children. We examine whether gender nonconformity predicts, or is predicted by, each other study variable over time. Although this aspect of our study was largely exploratory rather than hypothesis driven, we examined whether gender-nonconforming children are at greater risk for adjustment difficulties (e.g., peer victimization, internalizing problems, low self￾esteem) if they possess a problematic form of gender identity (high between-gender identity, low within-gender identity) than if they do not. For example, gender-nonconforming children who feel strong pressure for gender differentiation or feel disconnected from their own gender may be particularly distressed. Method Participants and Procedure All children in the fourth through seventh grades of a university laboratory school were invited to participate. Participants were 101 boys (M age 10.2 years) and 94 girls (M age 10.0 years) who received written informed parental consent; they represented 75% This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 846 PAULETTI, COOPER, AND PERRY