A meta-analysis of 41 studies with 6,017 participants examined the relationship between children's aggressive behavior and hostile attribution of intent. The analysis found a robust significant association between the two, with effect sizes varying considerably across studies. Larger effects were associated with more severe aggressive behavior, peer rejection as a selection criterion, inclusion of children aged 8–12, and absence of control for intelligence. Video and picture presentation of stimuli were associated with smaller effect sizes than audio presentation, while staging of actual social interactions was associated with the largest effects. The importance of understanding moderators of effect size for theory development is stressed. The study reviewed recent research on social cognition and aggressive behavior, focusing on Dodge's model of social information processing. It proposed that aggressive behavior results from atypical processing during social situations. Studies showed that aggressive behavior is associated with deviations in encoding, goals, response generation, response evaluation, and enactment. However, results on representation were less consistent. The study aimed to clarify these inconsistent findings concerning representation. Hostile attribution of intent, or hostile attribution bias, is a key element in theoretical accounts of the development of aggressive and antisocial behavior. It has been proposed that factors such as peer rejection and harsh parenting predispose children to attribute hostile intent. Hostile attributions are believed to cause aggressive behavior and limit nonaggressive interactions that could provide opportunities to learn prosocial behaviors. Hostile attribution of intent is also considered a target for cognitive behavioral interventions. Empirical studies on hostile attribution and aggression varied considerably, with some reporting small effects, others large effects, and some finding no significant effects. The present study used meta-analytic techniques to assess differences in effect sizes between studies on hostile attribution of intent and aggressive behavior in children, and to explore which characteristics of children and methods account for differences between effect sizes. The study examined the relation between hostile attribution of intent and aggressive behavior, and tested whether child and method characteristics influenced the strength of this relation. Child characteristics that were expected to influence effect sizes were severity and kind of behavior problems, sociometric status, intelligence, age, and gender. Methodological characteristics that were thought to influence effect sizes included personal involvement, stimulus presentation, context, number of stimuli, response format, setting, scoring, and reliability. The study included 41 studies with 6,017 participants. The results showed a significant overall effect size of r = .17, with effect sizes varying considerably between studies. Effect sizes were significantly related to severity of behavior problems, with larger effects for studies involving nonreferred extreme samples and referred to nonreferred samples. The study identified several significant moderators of effect size, including sociometric status, intelligence, age, and stimulus presentation. The strongest effect was found in studies involving staged interactions, while smaller effects were found in studies involving video or picture presentation. The study concluded that understanding these moderators is important for theory development.A meta-analysis of 41 studies with 6,017 participants examined the relationship between children's aggressive behavior and hostile attribution of intent. The analysis found a robust significant association between the two, with effect sizes varying considerably across studies. Larger effects were associated with more severe aggressive behavior, peer rejection as a selection criterion, inclusion of children aged 8–12, and absence of control for intelligence. Video and picture presentation of stimuli were associated with smaller effect sizes than audio presentation, while staging of actual social interactions was associated with the largest effects. The importance of understanding moderators of effect size for theory development is stressed. The study reviewed recent research on social cognition and aggressive behavior, focusing on Dodge's model of social information processing. It proposed that aggressive behavior results from atypical processing during social situations. Studies showed that aggressive behavior is associated with deviations in encoding, goals, response generation, response evaluation, and enactment. However, results on representation were less consistent. The study aimed to clarify these inconsistent findings concerning representation. Hostile attribution of intent, or hostile attribution bias, is a key element in theoretical accounts of the development of aggressive and antisocial behavior. It has been proposed that factors such as peer rejection and harsh parenting predispose children to attribute hostile intent. Hostile attributions are believed to cause aggressive behavior and limit nonaggressive interactions that could provide opportunities to learn prosocial behaviors. Hostile attribution of intent is also considered a target for cognitive behavioral interventions. Empirical studies on hostile attribution and aggression varied considerably, with some reporting small effects, others large effects, and some finding no significant effects. The present study used meta-analytic techniques to assess differences in effect sizes between studies on hostile attribution of intent and aggressive behavior in children, and to explore which characteristics of children and methods account for differences between effect sizes. The study examined the relation between hostile attribution of intent and aggressive behavior, and tested whether child and method characteristics influenced the strength of this relation. Child characteristics that were expected to influence effect sizes were severity and kind of behavior problems, sociometric status, intelligence, age, and gender. Methodological characteristics that were thought to influence effect sizes included personal involvement, stimulus presentation, context, number of stimuli, response format, setting, scoring, and reliability. The study included 41 studies with 6,017 participants. The results showed a significant overall effect size of r = .17, with effect sizes varying considerably between studies. Effect sizes were significantly related to severity of behavior problems, with larger effects for studies involving nonreferred extreme samples and referred to nonreferred samples. The study identified several significant moderators of effect size, including sociometric status, intelligence, age, and stimulus presentation. The strongest effect was found in studies involving staged interactions, while smaller effects were found in studies involving video or picture presentation. The study concluded that understanding these moderators is important for theory development.