Maternal and paternal parenting styles associated with relational aggression in children and adolescents: A conceptual analysis and meta-analytic review Yoshito Kawabata a , Lenneke R.A. Alink b,⇑ , Wan-Ling Tseng a , Marinus H. van IJzendoorn b,c , Nicki R. Crick a aUniversity of Minnesota, Twin Cities Campus, United States b Leiden University, Centre for Child and Family Studies, The Netherlands c Erasmus University Rotterdam, Erasmus Center for Moral Socialization Studies, The Netherlands article info Article history: Received 10 March 2011 Revised 25 August 2011 Available online 29 September 2011 Keywords: Relational aggression Parenting Meta-analysis abstract How does parenting affect relational aggression in children? The goal of the present series of meta-analyses based on 48 studies (28,097 children) was to analyze and integrate the findings on the associations between various types of parenting behaviors and relational aggression, and to identify potential substantive and methodological factors that may moderate these associations. To distinguish between different parenting strategies, experts sorted the parenting measures used in the studies into 10 groups. Results of a multiple correspondence analysis revealed four separate clusters: Positive parenting, psychologically controlling parenting, negative/harsh parenting, and uninvolved parenting. The meta-analyses demonstrated that more positive parenting was associated with less relational aggression (combined effect sizes r = .06, p < .05, for mothers, r = .08, p < .01, for fathers). More harsh parenting (combined effect sizes r = .11, p < .01, for mothers, r = .12, p < .01, for fathers) and more uninvolved parenting (combined effect sizes r = .07, p < .01, for mothers, absent for fathers) were associated with increased relational aggression. Paternal psychologically controlling parenting was positively related to relational aggression (r = .05, p < .01), whereas maternal psychologically controlling parenting was not (combined effect sizes r = .04, p = .09). The effect of several moderators is discussed. The findings of this study suggest that dimensions of positive and neg0273-2297/$ - see front matter 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.dr.2011.08.001 ⇑ Corresponding author. Address: Center for Child and Family Studies, Leiden University, P.O. Box 9500, 2300 RB, Leiden, The Netherlands. E-mail address: [e-mail] (L.R.A. Alink). Developmental Review 31 (2011) 240–278 Contents lists available at SciVerse ScienceDirect Developmental Review journal homepage: www.elsevier.com/locate/dr ative parenting behaviors of mothers and fathers are associated with children’s relational aggression and that these associations are – in case of fathers – contingent upon a number of sampling and procedural characteristics. 2011 Elsevier Inc. All rights reserved. Introduction During the last couple of decades, researchers in the field of aggression have increasingly devoted their attention to covert and indirect forms of aggression such as relational, social, or indirect aggression (Bjorkqvist, 1994; Côté, Vaillancourt, Barker, Nagin, & Tremblay, 2007; Crick & Grotpeter, 1995; Galen & Underwood, 1997; Murray-Close, Ostrov, & Crick, 2007; Underwood, Beron, & Rosen, 2009). For these types of aggression as opposed to physical and verbal aggression, we use here the term ‘‘relational aggression’’ to be consistent and avoid confusion. Relational aggression is conceptually and qualitatively different from physical aggression in that it focuses on behaviors that harm others through damage to relationships (Crick & Grotpeter, 1995). Examples of relational aggression include various hurtful behaviors such as intentionally withdrawing friendships, spreading rumors to hurt others, and excluding others from the peer group (Crick et al., 1999). Recent studies in this area have also demonstrated that relational aggression is one of the favorite forms of aggression through the internet and digital text messages (e.g., Berger, 2007; Raskauskas & Stoltz, 2007). A substantial body of research has reported consistent, negative outcomes for both the perpetrators and the victims of relational aggression, including peer rejection, isolation from peer groups, externalizing and internalizing adjustment problems (e.g., Card, Stucky, Sawalani, & Little, 2008; Crick, Ostrov, & Kawabata, 2007; Crick et al., 1999). In addition to the study on the outcomes and correlates of relational aggression, many efforts have been made to identify the antecedents of relational aggression, and one of the major areas of research has been parenting (Crick et al., 1999). Studies on the associations between parenting and relational aggression have produced both positive and non-significant results with varying effect sizes (e.g., Reed, Goldstein, Morris, & Keyes, 2008). Because relational aggression may have detrimental consequences for the perpetrators as well as for their victims (e.g., Crick et al., 1999, 2001; Crick & Zahn-Waxler, 2003; Hawker & Boulton, 2000), it is crucial to determine whether and how strongly different aspects of parenting are associated with the perpetration of relational aggression. This knowledge can be used to inform family-based interventions aimed at preventing the perpetration of relational aggression. Therefore, the goal of the present meta-analytic study was to analyze and synthesize the existing research on the associations between parenting and relational aggression, and to identify potential substantive and methodological factors that may influence these associations. Because numerous terms have been used to describe various types of parenting, we conducted a quantitative conceptual analysis in order to create some conceptual order and structure in this burgeoning but also somewhat chaotic literature (for an example of a quantitative conceptual analysis, see De Wolff & Van IJzendoorn, 1997). Relational, social, and indirect aggression The concepts of relational, social, and indirect aggression have been interchangeably used to describe forms of aggression that are behaviorally and conceptually different from direct physical and verbal aggression. Although these constructs have a lot in common, each of them captures a somewhat different dimension of aggression. Relational aggression is usually defined as a form of aggression in which relationships serve as an agent of harm, as distinct from a direct, physical form of aggression (for a review, see Crick et al., 1999). By definition, relational aggression includes both covert and overt behaviors that are used to hurt peers through the damage and manipulation of relationships. For example, spreading mean rumors and other forms of subtle social exclusion may be more covert, Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 241 whereas ignoring a member of the group and threatening to end friendships (e.g., ‘‘I won’t be your friend unless you do what I say’’) are more overt in nature. Social aggression includes, in addition to manipulative behaviors used for relational aggression, a broader spectrum of behaviors that damage peers’ self-esteem, social status, and reputations (e.g., facial expressions of disdain, gossiping; Galen & Underwood, 1997). Indirect aggression includes ‘‘indirect’’ forms of aggression such as ignoring and social exclusion, which are similar to relational and social aggression, but it also includes behaviors that are not directly related to manipulation of relationships (e.g., criticizing the other one’s hair or clothing; Bjorkqvist, 1994). Although these constructs are similar to one another by definition of content, relational aggression includes those specific behaviors that reflect intent to harm others through the damage and manipulation of relationships (Crick et al., 1999). In this meta-analytic synthesis, we used the term ‘‘relational aggression’’ to be consistent, but we did include studies on social and indirect aggression as well. The association between parenting and relational aggression: theoretical views A long tradition of research in child development is the study of the role of parent–child relationships and parenting behaviors in the development of aggression. As such, a substantial body of literature has documented that aggression, including relational aggression, is associated with positive or negative dimensions of parenting in childhood and adolescence (for a review, see Nelson & Crick, 2002). In this section, we discuss the main theoretical mechanisms underlying the association between parenting and relational aggression in addition to the possibility of child effects and genetically mediated influences. In the next section, we will review empirical studies on the link between relational aggression and parenting styles and propose specific hypotheses based on our literature review. Social learning theory In line with social learning theory (Bandura, 1973), the association between parenting and relational aggression can be explained by modeling and (vicarious) reinforcement. Children may observe and model their parents’ own use of relational aggression. For example, children may copy their mother’s use of socially excluding her own friends or spreading rumors about them. Or children may observe their mother’s use of relational aggression as a parenting strategy directed at the children themselves. However, it is important to note that modeling is not just simple mimicry of acts (Bandura, 1989). For example, children may observe their parents’ punitive and hostile parenting behaviors. But the manifestation of children’s aggressive acts as a result of modeling may not be identical to what they have observed. That is, children may learn from modeling their parents’ behaviors that being hostile and punitive is effective in getting one’s own way; hence, these children may adopt aggressive strategies, including physical and relational aggression, in their interactions with peers. These children may choose different forms of aggression, depending on specific situations and contexts they encounter. Children may also be reinforced or observe others being reinforced (vicarious reinforcement) for their use of relational aggression. This reinforcement can be positive (praise) or negative (absence of punishment). For example, permissive parenting may be related to children’s use of relational aggression (Casas et al., 2006) because it is a means of negative reinforcement. Parents who are permissive provide negative reinforcement for children’s undesired or disruptive behaviors (e.g., aggression) via their use of inconsistent discipline or laxness (Nelson & Crick, 2002). These parents usually fail to provide sufficient behavioral control to counteract the development and maintenance of their children’s aggressive or disruptive behaviors (Nelson & Crick, 2002). Attachment theory According to attachment theory (Bowlby, 1969/1982), children develop internal working models of the (social) world based on their early experiences with their primary caregiver(s). When a caregiver is sensitive and responsive early in the child’s life, the child is likely to develop a secure attachment relationship with this caregiver. However, growing up with an insensitive and unresponsive caregiver may put the child at risk for developing an insecure attachment relationship (Ainsworth, Blehar, Waters, & Wall, 1978). Children who have experienced insensitivity and rejection by their caregiver may develop 242 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 a blueprint of relationships that is characterized by insecurity and uncertainty. As a result, insecurely attached children are likely to develop a lack of confidence in themselves and in others, which puts them at risk for problems in peer relationships (Sroufe, Egeland, Carlson, & Collins, 2005). For example, Sroufe et al. (2005) showed that insecurely attached children exhibited more negative affect in their initiation of contact with others and reaction to others’ initiations, were more aggressive and less popular according to their peers and teachers, and showed less prosocial behavior as compared to securely attached children. This lack of social skills may be an explanation for these children’s use of relational aggression and their being the targets of such aggression. Parental sensitivity plays a fundamental role in the development of a secure attachment relationship, but other parenting strategies, such as harsh or psychologically controlling parenting, may also be of importance. For example, Soenens and colleagues posit that the association between psychologically controlling parenting and relational aggression can be explained by attachment theory (Soenens, Vansteenkiste, Goossens, Duriez, & Niemiec, 2008). They note that psychologically controlling parenting can result in insecurity and a thwarted sense of belongingness in the child, which in turn may lead to an insecure model of relationships with others. To compensate for these feelings of insecurity, children may engage in relationally aggressive behavior (Soenens et al., 2008). A recent meta-analysis has shown that attachment insecurity is a risk factor for the development of externalizing problems (Fearon, Bakermans-Kranenburg, Van IJzendoorn, Lapsley, & Roisman, 2010). Unfortunately, research on the association between attachment and relational aggression is scarce. However, there are some indications that insecurely attached children are indeed more likely to show relational aggression in the interactions with their peers than securely attached children (for a review see Michiels, Grietens, Onghena, & Kuppens, 2008). Emotion regulation theory Related to attachment theory, emotion regulation theory focuses on factors influencing the early development of emotion regulation and the role of emotion regulation in later social and emotional development. In the first year of life, the caregiver plays an important role in the child’s development of emotion regulation skills. Whereas emotion regulation is a dyadic process in this first year, it gradually becomes an individual process during early childhood (Sroufe, 1995). Sensitive caregivers help the child to regulate his or her emotions so the child does not get overly distressed and thus can gradually learn how to regulate his or her own emotions. However, when the parent does not provide enough guidance and support (e.g., by exhibiting insensitive or harsh caregiving) the development of effective emotion regulation skills may be compromised (e.g., Crockenberg, Leerkes, & Lekka, 2007). Several studies have shown that emotion dysregulation may hamper social and emotional development (e.g., Alink, Cicchetti, Kim, & Rogosch, 2009). More specifically, the association between emotion regulation and relational aggression has been investigated by Crick (1995). She found that relationally aggressive children reacted with heightened anger and distress to instrumental provocations (e.g., a peer breaks the child’s toy). This suggests that when provoked, relationally aggressive children may not be able to regulate their emotions effectively, and may use relational aggression as a means of dealing with their heightened distress. Social information processing theory Implied in the theories mentioned above is the idea that aggressive behavior may be triggered by a hostile interpretation of social cues (Crick & Dodge, 1994). An important aspect of social communication is the interpretation of another person’s intent. Dodge (1980) has shown that physically aggressive children often (negatively) misinterpret ambiguous cues, and as a result, feel provoked by others and react aggressively. Research has shown that this bias does not only explain physical aggression but also relational aggression. For example, Crick, Grotpeter, and Bigbee (2002) showed that relationally aggressive children exhibited hostile attributional biases when confronted with ambiguous relational information. Parents may directly or indirectly ‘‘teach’’ their children to interpret ambiguous social information in a hostile, relationally aggressive manner. For example, parents can reinforce the child’s own skepticism of other people’s intentions. Or a mother may explicitly express that she is not inviting her friend over for dinner because this friend had celebrated his/her birthday with his/her family and Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 243 therefore had not invited the mother. Indeed, research has shown that there is an association between maternal (instrumental and relational) intent attributions and those of her child (Nelson, Mitchell, & Yang, 2008) and a link between paternal aversive parenting and his child’s hostile attribution biases (Nelson & Coyne, 2009). Another explanation for deficits in social information processing that possibly leads to relational aggression can be found in attachment theory (Bowlby, 1969/1982). As described above, insecurely attached children may have developed an inner working model representing a hostile world in which the child is not worthy of love, support and respect from others, including peers. As a result, insecure children may interpret ambiguous relational cues as more negative than securely attached children. Relatedly, emotion regulation may play an important role in the development of deficits in social information processing. Izard, Stark, Trentacosta, and Schultz (2008) argue that emotion regulation and emotion information processing mutually influence each other. Children who are not good at regulating and recognizing their own emotions may have difficulties accurately interpreting others’ (ambiguous) emotions. And in turn, when children are not able to correctly process external emotional cues, this leads to inaccurate emotion knowledge, which hampers their abilities to regulate and express (appropriate) emotions themselves. Genetic explanations and child effects Often the association between parenting and child behavior is interpreted as a unidirectional effect with parenting influencing child behavior. However, other explanations for this relationship are also possible. For example, the association between parenting and relational aggression may be explained by genetic effects (cf. Rowe, 1994). In their behavioral genetic study on 6–18-year old twins, Tackett and colleagues (2009) have shown that 49–63% of the variance in (self- and mother-reported) relational aggression can be explained by genetic differences. The same genes may account for (dysfunctional) parenting as well as child relational aggression. Another possibility is that child behavior may evoke certain parenting strategies. Ge et al. (1996) investigated adopted children who were at genetic risk for antisocial behavior, based on their birth parents’ psychopathology. They showed that these at risk children were more likely to evoke harsh and inconsistent parenting than other adopted children, suggesting that the children’s behavior was (at least partially) influencing the type of parenting they received. On a similar note, Jaffee et al. (2004) have shown that the genetic factors that influenced corporal punishment and child antisocial behavior were largely similar, suggesting that child antisocial behavior partially evokes parental physical punishment (although this was not true for physical maltreatment). Similar effects may be involved in the association between parenting and relational aggression although to date the empirical evidence for this is scarce. One of the few studies that did investigate the direction of effects in the association between parenting and relational aggression revealed that in mother–child dyads the association between psychological control and child relational aggression represented a reciprocal process (Kuppens, Grietens, Onghena, & Michiels, 2009a). Relational aggression and parenting Relational aggression has been examined at various developmental stages – in preschool (e.g., Crick, Casas, & Mosher, 1997), middle childhood (e.g., Crick & Grotpeter, 1995), and adolescence (e.g., Werner & Crick, 1999). Whereas numerous studies have demonstrated consistent findings regarding negative consequences of relational aggression, the findings of existing research examining relational aggression and parenting vary with effect sizes ranging from small to moderate at best (Nelson & Crick, 2002). It has long been theorized that what children learn, how they react to the events and to others around them, and what they expect from themselves and others is affected by their relationships with parents and the behaviors of parents (Collins & Laursen, 1999; Hartup & Rubin, 1986; Maccoby & Martin, 1983). Parent–child interactions constitute an important developmental context in which children acquire social and behavioral expectations and learn social behaviors that could carry over into other relationship contexts such as peer relationships (e.g., Maccoby & Martin, 1983). Indeed, a substantial body of literature has demonstrated that dimensions of positive or negative parenting (i.e., authoritative, authoritarian, permissive, and neglecting) are differentially 244 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 associated with a wide variety of children’s social behavior such as aggression and prosocial behavior toward peers (Alink, Mesman, et al., 2009; Romano, Tremblay, Boulerice, & Swisher, 2005; Steinberg, Lamborn, Darling, Mounts, & Dornbusch, 1994). However, one of the limitations of this research is the fact that the majority of the samples include only one child per family. With such research designs, it is impossible to disentangle parenting and genetic influences. Indeed, behavioral genetic studies have shown that a large part of the variance in relational aggression can be explained by genetic differences (49–63%; Tackett, Waldman, & Lahey, 2009). Nevertheless, a substantial amount of the variance is explained by shared environmental factors (37–51%; Tackett et al., 2009), indicating that parenting, in addition to other shared environmental factors such as siblings, friends, school environment, do play an important role in the development of relational aggression. Taken together, it is crucial to fully understand whether and how dimensions of positive or negative parenting such as Baumrind’s three parenting styles are associated with children’s use of relational aggression in the peer context. Baumrind describes three prototypical parenting styles which are authoritative, authoritarian, and permissive parenting (Baumrind, 1971). Each parenting style is thought to differ in terms of two dimensions (i.e., responsiveness/warmth and control). Authoritative parenting is described as parenting behavior that is high in control and high in responsiveness, whereas authoritarian parenting is characterized as high in control and low in responsiveness. Permissive parenting is low in control and high in responsiveness. These parenting styles have been shown to be differentially linked with child behaviors in the short and long run. For example, whereas authoritative parenting is predictive of children’s adaptive social skills and coping mechanisms such as high levels of social competence, prosocial behavior, and emotion regulation, authoritarian parenting typically results in maladaptive behaviors such as aggression, delinquency, and social withdrawal (Baumrind, 1971). In this section, we discuss whether each type of parenting style is associated with specific child outcomes, including aggression, especially relational aggression. In addition, we discuss the association between psychologically controlling parenting and relational aggression, because this parenting strategy has been studied quite extensively in relation to relational aggression. Authoritative parenting Early studies on childrearing strived to identify parenting styles that would promote children’s competent behaviors such as self-reliance, self-control, and friendliness (Baumrind, 1967, 1971). This research yielded consistent results; that is, a combination of warmth, firm control, and clear standards of conduct formed the important characteristics of positive parenting. The term authoritative parenting, which meant that parents are high on both warmth and demand, was later used to describe positive parenting. Authoritative parent generally sets clear limits for the child and encourages compliance, but is also warm and uses reasoning as a means of explaining the rules. It has been widely documented that positive parenting is associated with better peer relations and with fewer aggressive behaviors (Baumrind, 1973). More recent studies have demonstrated that positive parenting, characterized as mutually contingent, warm, sensitive, and responsive interactions with the child, is associated with more social competence, including prosocial behavior and positive interactions with peers, and with less aggression and delinquency (e.g., Chen, Dong, & Zhou, 1997; Collins & Steinberg, 2006; Harrist, Pettit, Dodge, & Bates, 1994; Mize & Pettit, 1997; Pettit, Bates, & Dodge, 1997; Steinberg et al., 1994). Overall, parenting behaviors that are warm, responsive, and synchronous in the interactions with children may lead to higher levels of children’s social competence and lower levels of aggression toward peers. Researchers who are interested in a more covert or indirect form of aggression such as relational aggression have contended that positive dimensions of parenting may also influence the development of relational aggression. Whereas some studies have generated weak and null findings in this area (Casas et al., 2006; Russell, Hart, Robinson, & Olsen, 2003; Underwood, Beron, Gentsch, Galperin, & Risser, 2008), other studies have found that positive parenting is associated with lower levels of relational aggression (Crick et al., 1999). Similarly, Brown and colleagues showed that positive maternal affect (features of authoritative/positive parenting) is associated with less relational aggression in preschoolers (Brown, Arnold, Dobbs, & Doctoroff, 2007). Taken together, it is hypothesized that positive parenting would lead to less use of relational aggression toward peers. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 245 Authoritarian parenting Authoritarian parenting, in contrast to authoritative parenting, is characterized by high control and low responsiveness (Baumrind, 1967, 1971). Authoritarian parents value obedience and use strict rules. They usually use punitive discipline tactics, have an absolute standard, restrict the autonomy of the child, are low in warmth, and do not use induction as a means of explaining their demands. This type of parenting has been shown to be associated with children’s negative social adjustment such as being moody, hostile toward others, and low in self-esteem (Baumrind, 1973; Coie & Dodge, 1998). For example, a large body of literature has demonstrated that authoritarian parenting styles (e.g., punitive or coercive discipline) predict children’s use of physical aggression and hostile behaviors toward peers (e.g., Chen et al., 1997; Coie & Dodge, 1998; Hart, DeWolf, Wozniak, & Burts, 1992; Ladd & Pettit, 2002). Whereas findings with regard to the link between authoritarian parenting and children’s physical aggression are generally consistent, evidence for relational aggression again is somewhat varying with some weak and null findings. For example, mothers’ and fathers’ authoritarian parenting (Casas et al., 2006), maternal and paternal coercion (Hart, Nelson, Robinson, Olsen, & McNeilly-Choque, 1998), and mothers’ negative affect (Brown et al., 2007) have been shown to be concurrently related to increased relational aggression in preschoolers. Moreover, hostility in mother–child dyads and harsh/ineffective parenting were associated with children’s use of relational aggression toward peers (e.g., Vaillancourt, Miller, Fagbemi, Côté, & Tremblay, 2007). Other studies, however, did not find significant associations between authoritarian parenting and children’s relational aggression reported by teachers (Russell et al., 2003; Underwood et al., 2008). Overall, the majority of previous studies have shown that negative/harsh parenting may lead to children’s relational aggression; however, it is unclear how large the effect size across pertinent studies would be and which moderators might explain the divergent outcomes. Permissive parenting Another parenting style that was described by Baumrind (1967, 1971) is permissive parenting. This parenting style is characterized by low control and high responsiveness. Permissive parents are responsive to their children’s needs but do not set clear limits and boundaries. They make few demands, allow children to regulate their own actions as much as possible, and react in an accepting, nonpunitive way to their children’s impulses and wishes. This type of parenting or parent–child relationship that lacks firm rules and consistent discipline may create a socializing context for children’s negative behaviors such as aggression. For example, children with permissive parents may not fully develop an ability to control their negative emotions such as anger presumably because their parents typically do not monitor their children’s behaviors and do not make them reflect on this behavior. Consequently, these children are unable to regulate their aggressive impulses when dealing with problems, conflicts, and disagreements with peers and as a result display excessive levels of aggressive behaviors when they are angry. In support of this view, akin to negative/harsh parenting, the literature in this area has suggested that permissive or inconsistent parenting or lack of monitoring are generally related to higher rates of physical aggression in children (Rubin, Stewart, & Chen, 1995). Similarly, research has shown that permissive parenting, including lack of parental consistent monitoring, is associated with relational aggression at various developmental ages such as preschool age (e.g., Casas et al., 2006), school age (e.g., Sandstrom, 2007; Vaillancourt et al., 2007), and adolescence (e.g., Stocker, 2000). Psychologically controlling behavior A parenting style that has not been specifically distinguished by Baumrind but has been investigated in association with relational aggression is psychologically controlling parenting. Psychological control, in contrast to behavioral control in which children’s behaviors are the loci of control, is a negative form of parental control that interferes with children’s psychological autonomy (Barber, Stolz, & Olsen, 2005). Psychological control often hinders children’s development of social competence because it limits their opportunities to develop self-regulatory and coping skills to deal with social situations (Nelson, Hart, Wu, et al., 2006). Psychologically controlling parenting behaviors include tactics such as love withdrawal, guilt induction, negative affect-laden expressions (e.g., the expression of 246 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 disappointment and shame to the child), and excessive possessiveness or protectiveness, which are very similar to relationally aggressive strategies (i.e., manipulation of parent–child relationships or control of child’s behavior; Nelson & Crick, 2002). Hence, parental use of psychological control tactics may serve as a primer that makes children more likely to adopt relationally aggressive strategies in their interactions with peers (Casas et al., 2006). Supporting this view, the majority of studies in this area have demonstrated that parental psychological control is related to increased relational aggression at varying ages and in different cultures, including preschool girls in the United States (Casas et al., 2006) and in China (Nelson, Hart, Yang, Olsen, & Jin, 2006) as well as school-aged children (girls) in the United States (Nelson & Crick, 2002) and in Belgium (Kuppens et al., 2009a). Hence, it was hypothesized that psychologically controlling behavior would be related to increased relational aggression. To investigate whether the parenting constructs in the studies that were included in the current meta-analysis could be grouped according to Baumrind’s parenting styles (including psychologically controlling behavior), we asked experts in research on parenting to sort the parenting constructs and conducted a multiple correspondence analyses to cluster the different constructs. Separate meta-analyses were performed for each of these parenting constructs. Moderators Sex of child and parent An important research question regarding the development of aggression is whether the link between parenting and the development of relational aggression differs for males and females (Casas et al., 2006). Early work examining gender differences (main effects) in relational forms of aggression suggested that whereas males were more likely to engage in physical forms of aggression, females were more likely to exhibit relational forms of aggression (see Crick et al., 1999, for a review; Ostrov & Crick, 2007). This is especially the case during the late elementary school years as females become increasingly focused on close relationships with peers and their friendships become increasingly close, intimate, and exclusive (Kistner et al., 2010). This proposal is consistent with findings from meta-analytic reviews documenting that, relative to boys, girls place more emphasis on close, dyadic relationships, are more likely to adopt relational orientations, and are more frequently exposed to relational conflicts (Rose & Rudolph, 2006). Given these theoretical views and findings regarding sex differences favoring girls in relational aggression and forming close relationships, it is conceivable that parenting behaviors may be differentially linked with relational aggression, depending on the gender of parent and the child. Indeed, several studies in this area have demonstrated that paternal and maternal parenting behaviors are differentially associated with the development of relational aggression for boys and girls. For example, mothers’ use of psychological control (i.e., love withdrawal and guilt induction; Casas et al., 2006), physical coercion (Nelson, Hart, Wu, et al., 2006), and permissive parenting (Casas et al., 2006) and fathers’ use of psychological control (Casas et al., 2006; Nelson & Crick, 2002; Nelson, Hart, Wu, et al., 2006) and authoritarian parenting (Casas et al., 2006) were predictive of relational aggression for girls. Moreover, mothers’ use of coercive control (i.e., verbal hostility and corporal punishment; Nelson & Crick, 2002) and fathers’ use of love withdrawal, one dimension of psychological control that closely resembles relational aggression, were associated with increased relational aggression for boys (Casas et al., 2006). Another study revealed that physical coercion was positively related to forms of aggression (relational and physical) for boys, and psychological control was largely predictive of forms of aggression for girls (Nelson, Hart, Wu, et al., 2006). These sex-specific associations may be congruent with the view that relational aggression, which co-vary with high levels of intimacy and closeness, is considered to be relatively normative for females (i.e., mothers and daughters) as compared to males (i.e., fathers and sons; Crick, 1997; Crick, Bigbee, & Howes, 1996; Crick & Grotpeter, 1995). We expect the associations between parenting styles such as psychological control (parental relational aggression) and relational aggression would be stronger for girls than for boys. Accordingly, the present meta-analysis examined the moderating effect of children’s sex on the association between maternal or paternal parenting and relational aggression. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 247 Other sample characteristics Besides sex, we also considered moderators concerning sample characteristics such as the children’s age and their cultural background. The varying effect sizes regarding the associations between parenting and relational aggression (i.e., ranging from moderate to null) may be due to the age differences in the samples. Given that there are developmental trends and differences with regard to the development and maintenance of relational aggression (Crick et al., 1999) and the significance of parenting behaviors (Darling & Steinberg, 1993), differential associations between parenting and relational aggression may exist for children at different developmental ages. For example, given the literature that has suggested that peer relationships and friendships become more salient (Hartup & Stevens, 1997) and that relationships with parents are less central for older children and adolescents, parenting may have less influence on children’s behavior with increasing age. Further, children’s cultural backgrounds may be a moderator of the associations between parenting styles and relational aggression given the weak and null findings in the literature with different samples from different cultures. Although studies that directly examined cross-cultural differences in the association between parenting and relational aggression are scarce, they seem to indicate that relational aggression is more salient for children in Asian cultures than for children in Western cultures. For example, the literature suggests that for cultures in which individuals place more emphasis on relationships (e.g., Japan, China), as compared to Western cultures, relational aggression, which is covert and interpersonally aversive, is more commonly used than physical aggression, which is direct, confrontational in nature; relational aggression also seems to be more detrimental to children’s social–psychological adjustment problems such as depressive symptoms (Crick, Ostrov, & Kawabata, 2007; Kawabata, Crick, & Hamaguchi, 2010a, b). These findings suggest that relative to children in Western cultures, relational aggression, which is interpersonally aversive in nature, exerts a greater impact on child outcomes for children in non-Western cultures. Combined with the view that parent–child relationships are closer, more intimate, and more interdependent for Asian cultures such as Japan, it is possible that the impact of parenting on child behaviors may be stronger for children in such (Eastern) relationship-oriented cultures and, thus, place these children at higher risk for developing negative behaviors such as relational aggression. For example, these children who experience psychologically controlling parenting (i.e., being ignored, manipulated, and controlled by parents) within relationships that are more interdependent in nature may be more likely to learn relationally aggressive behaviors through modeling of parents’ behaviors than children who form more independent relationships with parents in Western cultures. Procedural moderators The varying magnitudes of the associations between parenting and relational aggression may also be due to methodological issues (i.e., differences in the assessment methods or in the sources of informants used in the studies). Therefore, we evaluated the effects of the methods of assessment (questionnaires, peer nominations, observations), sources of relational aggression (child, peer, parent, or teacher) and parenting (parent, child, or observer) on the strength of the association between parenting and relational aggression. Next, we tested whether the type of aggression that was assessed (relational, social, or indirect) was a significant moderator. In addition, it has been well documented that information about both the predictor and the outcome measure obtained from a single informant may be contaminated with informant-specific error of measurement (Kuppens, Grietens, Onghena, & Michiels, 2009b), which in turn may inflate true relations between parenting and relational aggression. For example, maternal psychological control was not associated with teacher-rated relational aggression (Casas et al., 2006; Hart et al., 1998), whereas parental psychological control was indeed associated with relational aggression when both constructs were reported by mothers (Casas et al., 2006). Therefore, we tested whether effect sizes are larger when measures were obtained from the same informant (on both relational aggression and parenting constructs) versus different informants. Other potentially important moderators include research design (i.e., concurrent versus longitudinal), whether physical aggression was used as a control variable in the analyses, and the inclusion of other covariates (e.g., age, sex). We hypothesized that concurrent designs would provide higher correlations between parenting and relational aggression than longitudinal ones because assessments of these constructs at the same time period may inflate the strength of the targeted association. Given 248 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 that relational aggression has been shown to be highly correlated with physical aggression (Card et al., 2008) and that both forms of aggression are associated with parenting (Crick et al., 1999), controlling for the contribution of physical aggression may allow us to examine the unique effect of parenting on relational aggression. Hence, we expected that the association between parenting and relational aggression would be stronger for studies, which did not take into account the effect of physical aggression statistically. In the same vein, given that sex, age, and other covariates are supposedly involved in the effects of parenting on relational aggression, we hypothesized that the association between parenting and relational aggression would be attenuated by controlling for the effects of these variables. Finally, in order to evaluate potential publication bias, we evaluated the effect of publication status of the studies (i.e., published versus unpublished/dissertation) on the association between parenting and relational aggression. It is possible that larger effect sizes are reported in published studies relative to unpublished studies or dissertations given that significant findings are more likely to be published than non-significant findings (Rosenthal, 1979). In the present study, we conducted separate meta-analyses for fathers’ and mothers’ parenting behaviors given that the association between parenting and relational aggression may vary by the sex of the parent (Nelson & Crick, 2002; Underwood et al., 2008). Indeed, research has generated results with varying degrees of statistical significance concerning associations between maternal and paternal parenting (i.e., psychologically controlling behavior) and relational aggression for boys and girls (Casas et al., 2006; Hart et al., 1998; Nelson & Crick, 2002). To date, the majority of studies have examined the association between maternal parenting behaviors and relational aggression, and no systematic comparison of the effects of maternal and paternal parenting on relational aggression has been done. Hence, the present meta-analytic study was conducted to examine the effects of paternal and maternal parenting on relational aggression separately. The present study The present study includes a conceptual analysis of the parenting constructs used in research on parenting and relational aggression, and it presents a set of quantitative meta-analyses on this association in each of the parenting domains uncovered in the conceptual analysis. Meta-analyses run the risk of combining too heterogeneous constructs, and thus comparing apples and oranges (Lipsey & Wilson, 2001). The conceptual approach paves the way to a better focused quantitative analysis of dimensions of parenting associated with relational aggression. Comparisons of effect sizes across these domains may provide some insight into what the most important components of parenting are to target in preventive interventions. Our overall hypothesis is that parenting style of mothers and fathers is substantially associated with relational aggression, for boys as well as for girls, but that different dimensions of parenting might show varying effect sizes depending on the sex of the parent and of the child. To the best of our knowledge, the present study is the first meta-analytic study that summarizes research findings on the associations between relational aggression and parenting and investigates multiple potential moderators of these associations across studies. Study 1: clustering the parenting constructs Method Literature search We obtained relevant articles and dissertations through the electronic databases EconLit, ERIC, PAIS, PsychInfo, Science Citation Index Expanded, Social Sciences Citation Index, and Art and Humanities Citation Index, using the keywords for parenting behaviors and parent–child relationships (parent⁄ – the asterisk indicates that the search includes the word or word fragment; for example, the keywords were parenting, parental, parents etc. – attachment, discipline, control, warmth, involvement, abuse, maltreatment, monitoring or sensitivity) and the keywords for relational aggression (social aggression, indirect aggression, relational aggression, victimization, or bullying). We also searched relevant studies using the reference lists of the collected papers and dissertations. To minimize the risk of publication bias, Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 249 we included unpublished data, which were available for our meta-analysis. To be able to include PhD theses, we also searched ‘‘Dissertation Abstracts International’’. We included PhD theses as Rosenthal (1991) showed how excluding PhD dissertations leads to inflated effect sizes (see also Van IJzendoorn, Bakermans-Kranenburg, & Alink, 2012). The last search was conducted at the end of January 2011. We selected studies that met the following criteria: (1) statistics were presented on the association between relational aggression (including social and indirect aggression) and parenting behaviors that directly involved parents and their children (i.e., constructs that assessed family environment, marital relationships, and sibling relationships were excluded); and (2) the sample consisted of children who were 18 years or younger (i.e., adult samples were excluded). We started with the number of 4218 articles and selected the final number of those fulfilling criteria, after reading the abstract, and sometimes the paper itself. Several studies on similar parenting constructs presented data on (partly) overlapping samples (e.g., Côté et al., 2007; Vaillancourt et al., 2007; and Loukas et al., 2005; Paulos, 2007). Because in a meta-analysis studies should be independent and participants should only included once in the same meta-analysis, studies using the same sample of participants cannot be included in the same meta-analysis more than once. The papers that reported on the largest groups of participants (Vaillancourt et al., 2007) and/or had a longitudinal design (Paulos, 2007) were included in our meta-analysis. In total, 48 independent studies, including 28,097 children, were included in our meta-analyses. Conceptual analysis: the sorting task Overall, 142 different parenting constructs were identified in the studies that were selected based on our literature search. We designed a separate card for each construct, including the definition that was given in the paper and examples for the specific parenting construct (e.g., items that were used in a questionnaire). To avoid that parenting constructs would be grouped together because they were assessed with the same method, the specific method that was used to assess parenting was not mentioned on the cards. In addition, any information about the source of the construct was absent. In order to cluster the constructs we asked experts, who were blind to the hypotheses of our study, to sort them. Because a number of these 142 constructs were almost identical, we decided to ask two experts to group the constructs that were obviously similar. The two experts agreed on 88% of the constructs (i.e., they both placed them in the same group or kept them as a separate construct), and reached consensus on grouping the other constructs. Of the 142 constructs, 42 were used separately and 100 were grouped into 34 different groups. This resulted in a set of 76 different constructs. Next, we asked other experts to sort these parenting constructs into 10 groups. This number of groups was chosen to facilitate the sorting task and to guarantee the statistical power of the subsequent analyses (Verkes, Van der Kloot, & Van der Meij, 1989). The instructions for the coders were as follows: ‘‘These cards contain different parenting constructs, all referring to various aspects of maternal and paternal behavior toward the child. On each card, the name and definition and/or examples of the specific construct are written. We would like to ask you to sort these constructs into 10 subgroups based on your idea of conceptual similarity between the constructs. A ‘‘group’’ of one card is also possible. The number of constructs in each group does not have to be equal.’’ Experts were defined as persons who had been actively involved in research on parenting for several years and who were at least participating in a graduate program in the child and family studies. A total of 17 experts were asked and 15 of them actually completed the sorting task. All of these coders had had extensive training in observing parent–child interactions. Almost half of the coders (N = 7) had a doctoral degree; the others were advanced graduate students. Data analysis We created a 76 (parenting constructs) 15 (expert coders) data matrix. The same numerical code (1–10) was given to parenting constructs that were sorted in the same group by a specific coder. Constructs that were sorted into different groups received different numerical codes. Next, we performed a multiple correspondence analysis in order to specify dimensions according to which the different constructs could be defined. Multiple correspondence analysis can be considered the counterpart of principal component analysis for categorical data. It detects and represents underlying structures, 250 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 or dimensions, in a data set. For each parenting construct, object scores are generated that represent the construct’s score on a specific dimension. Finally, we conducted a two-step cluster analysis on the object scores that resulted from the multiple correspondence analysis in order to cluster the parenting constructs into different groups. Results A multiple correspondence analysis with three dimensions specified yielded results that were most clearly interpretable. Inertias for the three dimensions were .88, .79, and .72 respectively. The dimensions (which we labeled positive–negative, uninvolved–involved, and harsh–nonharsh) each explained 88%, 79%, and 72% of the variance. The three object scores on the dimensions that each parenting construct received were saved. We performed a two-step cluster analysis on these object scores. To specify the optimal number of clusters we searched for the lowest value of the Schwarz’s Bayesian Criterion (BIC) in combination with a large Ratio of BIC Changes and the Ratio of Distance Measures. For four clusters, the BIC was lowest (147.42) and the Ratio of BIC Changes and the Ratio of Distance Measures were reasonably large (0.16 and 2.63, respectively). The clusters are presented in Table 1 and Fig. 1. Cluster 1 consisted of 24 parenting constructs, used in 23 different studies (in some cases, similar parenting constructs were used in different studies and some studies used several parenting constructs that fitted in the same cluster), and was labeled ‘‘Positive parenting.’’ Cluster 2 was labeled ‘‘Psychologically controlling parenting’’ and consisted of 25 parenting constructs, used in 22 different studies. Cluster 3 consisted of 21 constructs reflecting ‘‘Negative/harsh parenting,’’ used in 24 different studies, and six constructs that were used in eight different studies were included in Table 1 Parenting clusters yielded by the two-step cluster analysis on the multiple correspondence analysis object scores. Cluster 1 Cluster 2 Cluster 3 Cluster 4 Positive parenting Psychologically controlling parenting Negative/harsh parenting Uninvolved parenting Nurturance Monitoring Emotional maltreatment Permissive (2) Attachment Psychological control (3) Parent–child conflict Poor family management Positive relationship Parental dignity Negative affect Inconsistent parenting Positive reinforcement Overprotective/intrusive Authoritarian physical Laxness Authoritative (2) Directiveness Harsh discipline Poor monitoring Responsiveness Rule violation Hostile/ineffective parenting Positive and involved parenting Love withdrawal Coercion Discussion Monitoring/rules Negative interaction Help and guidance Consistent parenting Sexual abuse Positive affect Erratic emotional behavior Physical punishment Warmth Guilt induction Neglect Authoritative/sensitivity Non-physical discipline Inconsistent/physical discipline Positive interaction (2) Power assertion Physical abuse Conflict resolution Conformity Conflict Acceptance Training parenting Authoritarian: verbal/physical Encouragement Behavioral control Authoritarian Parental responsiveness Invalidating feelings Differential treatment Autonomy granting Personal attacks Non-reasoning/punitive strategies Open communication with parents Intervention Verbal hostility Intimacy Encouragement of modesty Conflict mutual Validation Relational aggression Negativity Coaching Constraining verbal expression Psych. control/love withdrawal/guilt induction Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 251 cluster 4, which was labeled ‘‘Uninvolved parenting.’’ It is clear that the sorters did not completely reproduce the Baumrind typology of parenting. For example, the permissive style was not represented as the uninvolved style did not load on anything close to warmth or positive affect. Study 2: meta-analyses Method Coding system The studies that were included in the meta-analyses are indicated in the list of references at the end of this paper with an asterisk and presented in Table 2. The studies were coded with the coding system presented in Table 3. As mean age we took the age at the first assessment in case of a longitudinal study. If the exact numbers for boys and girls were only given for maternal parenting and not for paternal parenting but the total number of participants for whom paternal parenting measures were available was smaller than that for maternal parenting, we used the percentage of boys derived from the analysis on maternal parenting to estimate the number for paternal parenting (e.g., Hart et al., 1998). Unfortunately, the numbers of effect sizes for the subgroups of effect sizes on indirect and social aggression were too small to conduct moderator analyses. Therefore, we decided to exclude this moderator from our analyses. Intercoder reliability was assessed by two coders who indepenFig. 1. Two-dimensional representations (for each combination of dimensions) of the parenting concepts in the multiple correspondence solution. 252 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 dently coded 10 of the studies. The agreement between the coders across the moderator variables was 100%. Meta-analytic procedures If studies reported data separately on boys and girls, both effect sizes were used in the meta-analysis (e.g., Hart et al., 1998; Underwood et al., 2008). If effect sizes for the relation between relational aggression and several parenting measures that would fall into the same parenting cluster were presented (e.g., Casas et al., 2006), we first conducted a meta-analysis within the study, and included the combined effect size in the final meta-analytic dataset. Several studies presented effect sizes based on data from multiple informants (e.g., correlations between mother-, father-, teacher-, and peer-reported aggression versus child-, mother-, and father-reported parenting; Kuppens et al., 2009a, 2009b, 2009c). For these studies we choose to include the most conservative effect sizes, i.e., effect sizes for which parenting and aggression data from different informants had been used (e.g., correlation between mother-reported aggression and child- and father-reported parenting, father-reported aggression and child- and mother-reported parenting, etc.). In addition, to be able to test whether the design of the study (longitudinal or concurrent) would affect the effect size, we included the longitudinal effect size when available (instead of the concurrent effect size). Generally, if studies reported effect sizes from analyses which controlled for the effect of physical aggression, we used this effect size, with the exception of the study by Nelson et al. (2006). In this case, the multivariate effect sizes controlled for physical aggression were only available for the comFig. 1 (continued) Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 253 bined measure of parenting (including both mothers and fathers). Because the bivariate correlations were available for fathers and mothers separately, we decided to include those in our meta-analyses. In addition, because no clear statistics were available for the relation between coaching and relational aggression in the study by Updegraff et al. (2005), we only focused on parental warmth (cluster 1) and intervention (cluster 2) for this study. Finally, a few studies presented the range of N (e.g., Risser, 2007; Russell et al., 2003). In these cases, we took the lowest number as a conservative estimate. Separate meta-analyses were performed for the four parenting clusters for fathers and mothers (studies that reported on the ‘‘person most knowledgeable’’ were included in the set on mothers; e.g., Foster, 2001). The meta-analysis on the relation between maternal positive parenting and relational aggression included 42 effect sizes on 25,538 children and the meta-analysis on this parenting cluster for fathers included 18 effect sizes and 3161 children. There were 36 effect sizes and 11,691 children in the meta-analysis on maternal psychologically controlling parenting and 18 effect sizes and 3288 children in the one on fathers. The meta-analysis on negative/harsh parenting contained 41 effect sizes and 7938 children for maternal parenting and 15 effect sizes and 1889 children for paternal parenting. Finally, 15 effect sizes and 4227 children were included in the meta-analysis on maternal uninvolved parenting. Because there were only two effect sizes in this cluster for fathers, it was not possible to perform a meta-analysis on the relation between paternal uninvolved parenting and relational aggression. Five studies yielding nine effect sizes (two for positive parenting, four for psychologically controlling parenting, and three for negative/harsh parenting) presented data on the average score for fathers and mothers or asked the children about ‘‘at least one of your parents’’ (Baier, 2005; Burnette & Reppucci, 2009; Foo, 2002; Rodgers, 2001). These studies could not be grouped in either the ‘‘mother’’ or the ‘‘father’’ clusters. Fig. 1 (continued) 254 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 Table 2 Studies included in the meta-analyses. Study Parent Parenting cluster a Sample size Age range (years) Country Albrecht et al. (2007) Mother, father 2 530 12–19 b Canada Ando et al. (2005) Mother 1, 2 2301 12–15 Japan Baier (2005) Combined 2 382 13–15 b Germany Baldry (2004) Mother, father 1 661 11–15 Italy Brown et al. (2007) Mother 1, 3, 4 23 boys, 34 girls 5–8 USA Burnette and Reppucci (2009) Combined 3 121 13–19 USA Campbell (1999) Mother 1, 3, 4 73 boys, 66 girls 9–11 USA Casas et al. (2006) Mother, father 1, 2, 3, 4 52 boys, 70 girls 2–6 USA Crick (2009) Mother 1, 2, 3, 4 70 boys, 63 girls 3–6 USA Crick (1999a) Mother, father 1, 2, 3 60 boys, 57 girls 9–10 USA Crick (1999b) Mother, father 1, 2, 3 61 boys, 69 girls 10–12 Japan Cullerton-Sen et al. (2008) Mother 3 410 6–13 USA Donovan (2009) Mother 3 221 boys 200 girls 10–13 USA Doyle (2010) Mother 1 513 7–8 USA Foo (2002) Combined 3 39 boys, 43 girls 8–9 b USA Foster (2001) Mother 1, 2, 3 7183/2302/737 c 2–11 Canada Gaertner et al. (2010) Combined 2 89 9–12 USA Hart et al. (1998) Mother, father 1, 2, 3 M: 101 boys, 106 girls F: 82 boys, 85 girls 3–7 Russia Herrenkohl et al. (2007) Mother 4 1749 12–15 USA Hutchison (2003) Mother, father 1 119 boys, 180 girls 11–15 USA Jung (2005) Mother 1, 2, 3 349 10–13 Korea Kuppens et al. (2009a, 2009b, 2009c) Mother, father 2 552 8–10 b Belgium Li (2007) d Mother, father 1, 2, 3 307 boys, 364 girls 9–11 China Lindsey et al. (2009) Mother 1, 3 135 boys, 133 girls 12–14 USA McNamara et al. (2010) Mother 1, 2 119 3–6 USA Murray et al. (2010) Mother 1, 2 141 12 USA McNeill (2002) Mother 1, 3, 4 213 boys, 151 girls 5–9 USA Mrug et al. (2008) Mother 1, 2 330 10–11 USA Nelson et al. (2006) Mother, father 2, 3 100 boys, 115 girls 4–6 China Ostrov and Bishop (2008) Mother 3 47 3.6 e (SD = 0.7) USA Park et al. (2005) Mother 3 207 3–4 b USA Paulos (2007) Mother 2, 3 230 boys, 261 girls 10–14 b USA Penney (2007) Mother 1, 2 1348 11–14 Canada Perkins (2009) Mother 1, 2, 3 201 and 143 12 Canada Pernice-Duca et al. (2010) Mother, father 1 158 9–12 USA Reed et al. (2008) Mother 2 21 boys, 20 girls 5–15 USA Risser (2007) Mother 2 73 boys, 73 girls 8–9 USA Rodgers (2001) Combined 1, 2 281 boys, 276 girls 11–12 USA Russell et al. (2003) Mother 1, 3 M cl 1: 349, M cl 3: 350 F cl 1: 210, F cl 3: 211 4–6 Australia USA Sandstrom (2007) Mother 1, 3, 4 Cl 1,2: 82 Cl 4: 45 boys, 37 girls 9–11 USA Soenens et al. (2008) Mother, father 2 284 15–20 Belgium Tseng and Kawabata (2008) Mother, father 2 (M + F), 3 (M) 61 boys, 69 girls 10–11 Taiwan (continued on next page) Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 255 Statistical analyses The meta-analyses were performed using the Comprehensive Meta-Analysis (CMA) program (Borenstein, Rothstein, & Cohen, 2005, Version 2). For each study, an effect size (correlation) was calculated. Combined effect sizes were computed in CMA. Significance tests and moderator analyses were performed through random-effect models, which are more conservative than fixed-effect models. In the random-effect model, the true effect could vary between studies, depending on characteristics of the specific sample. Because of these different characteristics, there may be different effect sizes underlying different studies (Borenstein, Hedges, Higgins, & Rothstein, 2009). To test the homogeneity of the overall and specific sets of effect sizes, we computed Q-statistics (Borenstein et al., 2009). In addition, we computed 85% confidence intervals (CIs) around the point estimate of each set of effect sizes (cf. Van IJzendoorn, Juffer, & Klein Poelhuis, 2005). Q-statistics and p-values were also computed to assess differences between combined effect sizes for specific subsets of study effect sizes grouped by moderators. Contrasts were only tested when at least two of the subsets consisted of at least four studies each (cf. Bakermans-Kranenburg, Van IJzendoorn, & Juffer, 2003). Finally, we examined differences in (absolute values of) combined effect sizes between the four parenting constructs for mothers and fathers separately, and we compared effect sizes for the same parenting construct between mothers and fathers. Nonoverlapping 85% confidence intervals indicated significant differences (Goldstein & Healy, 1995; Van IJzendoorn, Juffer, & Klein Poelhuis, 2005). Funnel plots for each subset were examined in order to detect possible publication bias. A funnel plot is a plot of each study’s effect size against its standard error (usually plotted as 1/SE, or precision). It is expected that this plot has the shape of a funnel, because studies with smaller sample sizes (larger standard errors) have increasingly large variation in estimates of their effect size as random variation becomes increasingly influential, representing the broad side of the funnel, whereas studies with larger sample sizes have smaller variation in effect sizes which represents the narrow end of the funnel (Duval & Tweedie, 2000b; Sutton, Duval, Tweedie, Abrams, & Jones, 2000). However, smaller studies with non-significant results or with effect sizes in the non-hypothesized direction are less likely to be published, whereas for large studies, publication of small or non-significant effect sizes or effect sizes in the non-hypothesized direction is more likely because large studies are generally deemed more trustworthy. Therefore, a funnel plot may be asymmetrical around its base (i.e., for small studies no effect sizes for non-significant results or results in the non-hypothesized direction). The degree of asymmetry in the funnel plot was examined by estimating the number of studies which have no symmetric counterpart on the other side of the funnel. The ‘‘trim and fill’’ method was used to test the influence of possible adjustments of the sets of studies for publication bias (Duval & Tweedie, 2000a, 2000b). Table 2 (continued) Study Parent Parenting cluster a Sample size Age range (years) Country Underwood et al. (2008) Mother, father 1, 3 128 boys, 128 girls 9 b USA Updegraff et al. (2005) f Mother, father 1, 2 370 13–16 USA Vaillancourt et al. (2007) Mother 1, 3, 4 736 boys, 665 girls 2 b Canada Wang et al., 2009 Mother 1 7182 12–16 USA Werner et al. (2006) Mother 1, 2 87 3–5 USA Yu and Gamble (2008) Mother 2, 3 433 14.3 e (SD = 2.1) USA a 1 = Positive parenting; 2 = psychologically controlling parenting; 3 = negative/harsh parenting; 4 = uninvolved parenting. b Longitudinal studies; Age at first assessment is given. c Foster (2001): the number of participants was Winsorized separately for the different parenting clusters and was adjusted to,7183 for cluster 1, 2302 for clusters 2, and 737 for cluster 3. d Li (2007): we used only peer measures for aggression, because these clearly reflect relational aggression. e Age range not clear, mean and SD is given. f Updegraff et al. (2005): the effect sizes on the parenting measure ‘‘coaching’’ were unclear so we did not use this in the analyses. 256 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 We checked for outlying effect sizes and sample sizes separately for the different subsets of studies. Z-values below 3.29 or greater than 3.29 were considered outliers (Tabachnick & Fidell, 2001). No outliers were detected for effect size, but there was one study with outlying sample sizes (Foster, 2001). We winsorized (highest non-outlying number plus 1) this number of participants separately for the different parenting clusters, which resulted in an N of 7183 for cluster 1, 2302 for clusters 2, and 737 for cluster 3. Results Maternal positive parenting The combined effect size for the relation between relational aggression and maternal positive parenting was small but significant (r = .06, 85% CI = .09, .03, p < .05; Fig. 1) in a heterogeneous set of studies (Q = 201.34, p < .01; Table 4). Overall, more experience with positive parenting by mothers was related to less relational aggression in children. None of the moderators were significant. Table 3 Coding system for meta-analyses. Variable Coding Sample Age (at first assessment) 1 = 0–5 2 = 6–10 3 = 11–20 4 = broad Sex of children 1 = boys 2 = mixed 3 = girls Country 1 = east 2 = west Design Aggression 1 = relational 2 = social 3 = indirect Informant aggression 1 = child 2 = peers 3 = parent 4 = teacher 5 = observer 6 = other Type of assessment aggression 1 = questionnaire 2 = peer nomination 3 = observation 4 = other Informant parenting 1 = parent 2 = child 3 = observer 4 = other Type of assessment parenting 1 = questionnaire 2 = observation 3 = other Same informant aggression and parenting? 0 = no 1 = yes Longitudinal design 0 = no 1 = yes (parenting assessed > 3 mo before aggression) Controlled for physical aggression? 0 = no 1 = yes Other covariates in the analysis? 0 = no 1 = yes Published 0 = no (including dissertation) 1 = yes Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 257 Maternal psychologically controlling parenting The meta-analytic relation between maternal psychologically controlling parenting was not significant (r = .04, 85% CI = .01, .07, p = .09; Fig. 1) in a heterogeneous set of studies (Q = 121.94, p < .00; Table 5). Again, none of the moderators that we tested were significant. Table 4 Meta-analytic results of studies relating maternal positive parenting and relational aggression. Characteristics kn r 85% CI Qa p Total set 42 25,538 .06⁄⁄ .09, .03 201.34 .00 Sample Age 0.27 .88 0–5 years 9 1017 .05 .12, .02 6–10 years 15 2357 .04 ..10, 01 17.51 11–19 years 15 13,580 .07⁄ .11, .02 139.54⁄⁄ Broad 3 8584 .08 .16, .01 16.27⁄⁄ Sex 0.14 .93 Boys 13 2078 .05 .10, .01 9.24 Mixed 14 20,531 .07⁄ .11, .03 167.14⁄⁄ Girls 15 2929 .06 .11, .01 12.32 Culture/country 0.26 .61 West 36 22,078 .06⁄⁄ .09, .03 190.53⁄⁄ East 6 3451 .04 .10, .03 8.09 Study design Informant aggression 0.17 .92 Child 8 12,322 .07 .13, .01 135.56⁄⁄ Peers 3 753 Parent 3 8584 Teacher 16 2359 .05 .10, .00 14.19 Other 0 Multiple 12 1520 .06 .10, .00 12.84 Type of assessment aggression 0.01 .92 Questionnaire 35 24,535 .06⁄⁄ .09, .03 194.73⁄⁄ Peer nomination 3 753 Observation 0 Multiple 4 250 .05 .17, .06 2.17 Informant parenting 0.92 .82 Parent 21 11,142 .05 .09, .01 45.96⁄⁄ Child 9 12,420 .09⁄ .14, .03 133.88⁄⁄ Observer 6 464 .07 .16, .03 5.51 Other 0 Multiple 6 1512 .03 .11, .04 8.86 Type of assessment parenting 0.02 .89 Questionnaire 36 25,074 .06⁄⁄ .09, .03 195.60⁄⁄ Observation 6 464 .07 .16, .02 5.51 Multiple 0 Same informant aggr-parenting? 1.56 .21 No 31 4463 .04 .07, .01 22.72 Yes 11 21,075 .09⁄⁄ .13, .05 159.92⁄⁄ Partially Longitudinal 0.68 .41 Concurrent 33 22,758 .07⁄⁄ .10, .04 182.06⁄⁄ Parenting – aggression 9 2780 .03 .09, .02 6.83 Controlled for phys aggr? No 40 24,137 .06⁄⁄ .09, .03 195.08⁄⁄ Yes 2 1401 Other covariates? 0.82 .36 No 33 15,667 .05⁄ .08, .02 178.26⁄⁄ Yes 9 9871 .09⁄ .14, .03 17.71 Publication 1.80 .18 Published paper 23 14,091 .08⁄⁄ .12, .04 39.44 Dissertation/unpublished data 19 11,447 .03 .07, .01 156.69⁄⁄ Note. Subsets with fewer than four effect sizes were not included in the moderator tests. 258 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 Maternal negative/harsh parenting The combined effect size for the relation between maternal negative/harsh parenting and relational aggression was significant (r = .11, 85% CI = .08, .13, p < .01; Fig. 1) in a heterogeneous set of studies Table 5 Meta-analytic results of studies relating maternal psychologically controlling parenting and relational aggression. Characteristics k n r 85% CI Q a p Total set 36 11,691 .04 .01, .07 121.94⁄⁄ .00 Sample Age 0.19 .91 0–5 years 10 883 .05 .01, .11 4.76 6–10 years 9 1616 .03 .02, .09 3.32 11–19 years 14 6850 .05⁄ .01, .09 65.55⁄⁄ Broad 3 2342 Sex 0.84 .66 Boys 11 1136 .06 .01, .12 6.16 Mixed 13 8958 .02 .02, .07 20.71⁄ Girls 12 1597 .05 .00, .11 3.25 Culture/country 0.25 .62 West 26 7895 .03 .01, .07 83.01⁄⁄ East 10 3796 .05 .01, .11 36.72⁄⁄ Study design Informant aggression 6.25 .10 Child 8 5284 .00 .03, .04 37.79⁄⁄ Peers 4 886 .06 .00, .13 0.84 Parent 1 2301 Teacher 8 600 .00 .07, .06 4.74 Other a 1 433 Multiple 14 2187 .09⁄⁄ .05, .13 6.33 Type of assessment aggression 0.80 .67 Questionnaire 24 9589 .02 .01, .07 102.62⁄⁄ Peer nomination 4 886 .07 .02, .15 0.84 Observation 0 Multiple 8 1216 .06 .01, .13 2.62 Informant parenting 2.47 .29 Parent 22 4416 .02 .02, .06 51.21⁄⁄ Child 10 5720 .02 .03, .07 40.77⁄⁄ Observer 0 Other 0 Multiple 4 1555 .11⁄ .04, .18 2.06 Type of assessment parenting Questionnaire 35 11,321 .03 .00, .06 113.34⁄⁄ Observation 0 Multiple 1 370 Same informant aggr-parenting? 0.92 .34 No 27 3879 .05⁄ .01, .08 22.15 Yes 9 7812 .01 .03, .06 69.51⁄⁄ Partially 0 Longitudinal 0.85 .36 Concurrent 31 9977 .03 .01, .06 103.61⁄⁄ Parenting – aggr 5 1714 .07 .01, .14 1.23 Controlled for phys aggr? No 35 11,258 .09 .06, 24 116.89⁄⁄ Yes 1 433 Other covariates? .00 .96 No 31 10,296 .04 .00, .07 103.22⁄⁄ Yes 5 1395 .03 .03, .11 10.44⁄ Publication 0.52 .47 Published paper 18 5732 .05 .01, .10 58.51⁄⁄ Dissertation/unpublished data 18 5959 .02 .02 .06 59.94⁄⁄ Note. Subsets with fewer than four effect sizes were not included in the moderator tests. a Yu and Gamble (2008): Sibling reports were used to assess sibling relational aggression. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 259 (Q = 77.87, p < .01; Table 6). More maternal negative/harsh parenting was associated with more relational aggression in the child. One of the studies in this subset presented data on child maltreatment (Cullerton-Sen et al., 2008) which may reflect a parenting strategy that is much more severe than the other strategies in this cluster. However, excluding this study did not make a difference. Studies in which no covariates were used in the analyses generated larger effect sizes (r = .12, p < .01, k = 35, n = 5635) as compared to studies in which covariates were included (r = .05, p = .16, k = 6, n = 2303; Q = 3.92, p < .05). None of the other moderators were significant. Maternal uninvolved parenting Maternal uninvolved parenting was significantly positively related to relational aggression (r = .07, 85% CI = .05, .09, p < .01; Table 7, Fig. 1). The set of effect sizes was homogeneous (Q = 13.93, p = .46). None of the moderators were significant. Paternal positive parenting The combined effect size for the relation between paternal positive parenting and relational aggression was also significant (r = .08, 85% CI = .13, .04, p < .01; Table 8, Fig. 1) in a heterogeneous set of studies (Q = 37.75, p < .01). More paternal positive parenting was related to lower levels of child relational aggression. The gender of the children was a significant moderator (Q = 5.86, p = .05). The combined effect sizes for the mixed gender group (r = .17, p < .01, k = 4, n = 1241) was significantly different from that only girls (r = .02, p = .59, k = 7, n = 953; Q = 5.90, p < .05) but not from the group of boys (r = .06, p = .23; k = 7, n = 809; Q = 2.44, p = .12) The difference between the effect size that result from published papers (r = .14, p < .01; k = 10, n = 1944) versus those from nonpublished data and dissertations (r = .00, p = .94; k = 8, n = 1217) was also significant (Q = 10.03, p < .01). None of the other moderators were significant. Paternal psychologically controlling parenting Results of the meta-analysis on paternal psychologically controlling parenting indicated that this construct was significantly positively related with relational aggression (r = .05, 85% CI = .03, .08, p < .01; Fig. 1) in a homogeneous set of studies (Q = 11.64, p = .82; Table 9). More controlling parenting by fathers was associated with more relational aggression in their children, in particular in girls. When the combined effect size for boys was compared to that of girls, sex appeared to be a significant moderator. The relation between paternal psychologically controlling parenting and relational aggression was significantly larger for girls (r = .10, p < .01, k = 7, n = 829) than for boys (r = .01, p = .87, k = 7, n = 723; Q = 3.90, p < .05). None of the other moderators were significant. Paternal negative/harsh parenting The meta-analytic relation between paternal negative/harsh parenting and relational aggression was also significant (r = .12, 85% CI = .08, .16, p < .01; Fig. 1) in a homogeneous set of studies (Q = 21.58, p = .09; Table 10). This indicates that more paternal negative/harsh parenting was related to more relational aggression in children. None of the moderators was significant. Comparisons between effects of parenting strategies To test whether the effect sizes for the different maternal and paternal parenting strategies were significantly different, we compared the 85% confidence intervals (of the absolute values of the effect sizes). Nonoverlapping confidence intervals indicated a significant difference (Goldstein & Healy, 1995; Van IJzendoorn et al., 2005). None of the effect sizes for the three parenting clusters (positive parenting, psychologically controlling parenting, and negative/harsh parenting) differed between mothers and fathers (the fourth cluster, uninvolved parenting, could not be compared because the number of effect sizes for fathers was too small). However, for both maternal and paternal parenting, the effect sizes for negative/harsh parenting (r = .11, 85% CI = .08, .13, p < .01 for mothers and r = .12, 85% CI = .08, .16, p < .01 for fathers) were significantly larger than those for psychologically controlling parenting (r = .04, 85% CI = .01, .07, p = .09 for mothers and r = .05, 85% CI = .03, .08, p < .01 for fathers; Fig. 2). 260 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 Table 6 Meta-analytic results of studies relating maternal negative/harsh parenting and relational aggression. Characteristics kn r 85% CI Q a p Total set 41 7938 .11⁄⁄ .08, .13 77.87 .00 Sample Age 0.42 .94 0–5 years 10 1074 .11⁄⁄ .05, .17 10.21 6–10 years 16 2226 .10⁄⁄ .05, .14 17.05 11–19 years 11 2293 .10⁄⁄ .05, .15 12.87 Broad 4 2345 .13⁄⁄ .06, .19 36.47⁄⁄ Sex 2.04 .36 Boys 16 2571 .11⁄⁄ .06, .15 11.61 Mixed 9 2816 .14⁄⁄ .09, .18 37.59⁄⁄ Girls 16 2551 .08⁄⁄ .04, .20 20.16 Culture/country 0.02 .88 West 32 6443 .10⁄⁄ .08, .13 70.09⁄⁄ East 9 1495 .11⁄⁄ .06, .17 7.48 Study design Informant aggression 0.30 .86 Child 3 692 Peers 7 1398 .11⁄⁄ .05, .17 6.49 Parent 3 2138 Teacher 13 1641 .10⁄⁄ .05, .15 11.47 Other a 1 433 Multiple 14 1645 .11⁄⁄ .06, .16 15.39 Type of assessment aggression 0.58 .75 Questionnaire 26 5712 .11⁄⁄ .08, .14 58.83⁄⁄ Peer nomination 7 1398 .11⁄⁄ .06, .17 6.49 Observation 0 Multiple 8 837 .07 .00, .14 12.17 Informant parenting 0.90 .83 Parent 25 5041 .10⁄⁄ .07, .13 60.78⁄⁄ Child 4 912 .11⁄⁄ .03, .18 2.36 Observer 5 532 .16⁄⁄ .07, .24 5.15 Other b 1 410 Multiple 6 1043 .09⁄ .02, .16 6.79 Type of assessment parenting 0.90 .34 Questionnaire 33 6857 .10⁄⁄ .07, .13 70.61⁄⁄ Observation 5 532 .16⁄⁄ .07, .24 5.15 Other b 1 410 Multiple 2 139 Same informant aggr-parenting? 2.50 .29 No 35 4876 .11⁄⁄ .08, .13 34.07 Yes 5 2629 .13⁄⁄ .07, .18 36.11⁄⁄ Partially 1 433 Longitudinal 0.84 .36 Concurrent 29 4693 .12⁄⁄ .09, .15 55.87⁄⁄ Parenting – aggr 12 3245 .09⁄⁄ .05, .13 16.34⁄ Controlled for phys aggr? No 38 6104 .12⁄⁄ .10, .15 55.46⁄ Yes 3 1834 Other covariates? 3.92 .05 No 35 5635 .12⁄⁄ .10, .15 52.62⁄ Yes 6 2303 .05 .00, .10 9.68 Publication 0.31 .58 Published paper 20 4055 .10⁄⁄ .06, .13 32.70⁄ Dissertation/unpublished data 21 3883 .11⁄⁄ .08, .15 36.71⁄ Note. Subsets with fewer than four effect sizes were not included in the moderator tests. a Yu and Gamble (2008): Sibling reports were used to assess sibling relational aggression. b Cullerton-Sen et al. (2008): CPS reports were used to assess parenting. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 261 Publication bias Using the trim and fill method (Duval & Tweedie, 2000a, 2000b), asymmetry (missing studies in the non-hypothesized direction) was only found for the meta-analysis on maternal psychologically conTable 7 Meta-analytic results of studies relating maternal uninvolved parenting and relational aggression. Characteristics kn r 85% CI Q a p Total set 15 4047 .07⁄⁄ .05, .09 13.93 .46 Sample Age 0.07 .79 0–5 years 4 255 .06 .04, .15 2.17 6–10 years 8 642 .08 .02, .14 9.42 11–19 years 1 1749 Broad 2 1401 Sex 2.30 .13 Boys 7 1212 .02 .02, .06 4.07 Mixed 1 1749 Girls 7 1086 .09⁄⁄ .04, .13 5.81 Culture/country West 15 4047 .07⁄⁄ .05, .09 13.93 East 0 Study design Informant aggr .10 .76 Child 1 1749 Peers 2 82 Parent 2 1401 Teacher 6 560 .08 .02, .15 7.97 Other 0 Multiple 4 255 .06 .04, .15 2.17 Type of assessment aggr Questionnaire 11 3832 .07⁄⁄ .04, .10 11.14 Peer nomination 2 82 Observation 0 Multiple 2 133 Informant parenting 0.47 .49 Parent 8 1980 .04 .01, .08 3.58 Child 1 1749 Observer 4 196 .08 .01, .20 6.88 Other 0 Multiple 2 122 Type of assessment parenting 0.16 .69 Questionnaire 11 3851 .07⁄⁄ .04, .09 6.91 Observation 4 196 .10 .01, .20 6.88 Other 0 Multiple 0 Same informant aggr-parenting? No 12 897 .07⁄ .02, .12 11.67 Yes 3 3150 Partially 0 Longitudinal Concurrent 13 2646 .09⁄⁄ .06, .11 11.98 Parenting – aggr 2 1401 Controlled for phys aggr? 2.04 .15 No 11 2564 .09⁄⁄ .06, .12 10.48 Yes 4 1483 .04 .00, .08 1.41 Other covariates? No 13 2646 .09⁄⁄ .06, .11 11.98 Yes 2 1401 Publication 0.03 .86 Published paper 9 33,411 .07⁄⁄ .04, .10 10.87 Dissertation/unpublished data 6 636 .06 .00, .12 3.01 Note. Subsets with fewer than four effect sizes were not included in the moderator tests. 262 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 trolling. Two effect sizes were imputed, resulting in an effect size that was also not significant (r = .03, 85% CI = .01, .07). No asymmetry was found in the funnel plots for the other meta-analyses. Therefore, no evidence for publication bias was found in these meta-analyses. Table 8 Meta-analytic results of studies relating paternal positive parenting and relational aggression. Characteristics kn r 85% CI Q a p Total set 18 3161 .08⁄⁄ .13, .04 37.75 .00 Sample Age 0.75 .69 0–5 years 5 499 .13⁄ .22, .04 6.74 6–10 years 7 1202 .06 .13, .01 22.24⁄⁄ 11–19 years 6 1460 .08 .15, .01 5.64 Broad 0 Sex 5.86 .05 Boys 7 809 .06 .12, .01 10.14 Mixed 4 1241 .17⁄⁄ .23, .11 10.29⁄ Girls 7 953 .02 .09, .04 5.03 Culture/country 2.50 .11 West 14 2360 .11⁄⁄ .15, .06 26.29⁄ East 4 801 .01 .09, .07 2.75 Study design Informant aggr 1.13 .29 Child 2 819 Peers 2 671 Parent 0 Teacher 5 633 .13⁄ .20, .05 4.73 Other 0 Multiple 9 1038 .05 .12, .01 9.46 Type of assessment aggr Questionnaire 14 2373 .12⁄⁄ .16, .08 24.36⁄ Peer nomination 2 671 Observation 0 Multiple 2 117 Informant parenting 1.28 .53 Parent 9 880 .10⁄ .17, .03 7.24 Child 4 1118 .12⁄ .20, .03 15.15⁄⁄ Observer 0 12.06⁄ Other 0 Multiple 5 1163 .03 .11, .05 Type of assessment parenting Questionnaire 18 3161 .08⁄⁄ .13, .04 37.75⁄⁄ Observation 0 Other 0 Multiple 0 Same informant aggr-parenting? No 15 1972 .05 .09, .00 16.81 Yes 3 1189 Partially 0 Longitudinal Concurrent 16 2905 .09⁄⁄ .13, .04 37.54⁄⁄ Parenting – aggr 2 256 Controlled for phys aggr? No 18 3161 .08⁄⁄ .13, .04 37.75⁄⁄ Yes 0 Other covariates? No 17 2791 .08⁄ .12, .03 34.38⁄⁄ Yes 1 370 Publication 10.03 .00 Published paper 10 1944 .14⁄⁄ .18, .10 18.57⁄ Dissertation/unpublished data 8 1217 .00 .05, .05 3.29 Note. Subsets with fewer than four effect sizes were not included in the moderator tests. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 263 Table 9 Meta-analytic results of studies relating paternal psychologically controlling parenting and relational aggression. Characteristics kn r 85% CI Qa p Total set 18 3288 .05⁄⁄ .03, .08 11.64 .82 Sample Age 0.52 .77 0–5 years 6 504 .02 .04, .09 1.38 6–10 years 7 1470 .06⁄ .02, .10 5.88 11–19 years 5 1314 .05⁄ .01, .09 3.87 Broad 0 Sex 3.96 .14 Boys 7 723 .01 .06, .05 1.94 Mixed 4 1736 .06⁄ .02, .09 2.80 Girls 7 829 .10⁄⁄ .05, .15 2.94 Culture/country 0.04 .84 West 10 2142 .06⁄ .02, .09 3.98 East 8 1146 .05 .01, .09 7.62 Study design Informant aggr 0.00 .98 Child 1 530 Peers 4 886 .06 .02, .10 6.20 Parent 0 Teacher 2 167 Other 0 Multiple 11 1705 .06⁄ .02, .09 5.21 Type of assessment aggr 1.52 .47 Questionnaire 8 1319 .03 .01, .07 1.13 Peer nomination 4 886 .05 .01, .11 6.20 Observation 0 Multiple 6 1083 .08⁄ .04, .12 2.78 Informant parenting Parent 14 1552 .05 .01, .09 8.78 Child 1 530 Observer 0 Other 0 Multiple 3 1206 Type of assessment parenting Questionnaire 17 2918 .06⁄⁄ .03, .08 11.51 Observation 0 Other 0 Multiple 1 370 Same informant aggr-parenting? No 16 2388 .06⁄⁄ .03, .09 11.36 Yes 2 900 Partially 0 Longitudinal Concurrent 16 2206 .06⁄⁄ .03, .09 11.39 Parenting – aggr 2 1082 Controlled for phys aggr? No 18 3288 .05⁄⁄ .03, .08 11.64 Yes 0 Other covariates? No 17 2918 .06⁄⁄ .03, .08 11.51 Yes 1 370 Publication 0.07 .79 Published paper 10 2240 .05⁄ .02, .08 4.59 Dissertation/unpublished data 8 1048 .06 .02, .10 6.97 Note: Subsets with fewer than four effect sizes were not included in the moderator tests. 264 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 Table 10 Meta-analytic results of studies relating paternal negative/harsh parenting and relational aggression. Characteristics kn r 85% CI Qa p Total set 15 1889 .12⁄⁄ .08, .16 21.58 .09 Sample Age 1.62 .45 0–5 years 7 715 .11⁄ .04, .17 3.95 6–10 years 6 1044 .15⁄⁄ .09, .21 13.79⁄ 11–19 years 2 130 Broad 0 Sex 1.78 .18 Boys 7 790 .17⁄⁄ .10, .23 9.34 Mixed 1 211 Girls 7 888 .08 .02, .15 7.95 Culture/country 0.11 .74 West 9 873 .11⁄ .05, .17 6.61 East 6 1016 .13⁄⁄ .06, .19 13.99⁄ Study design Informant aggr 2.48 .29 Child 0 Peers 4 886 .16⁄⁄ .09, .23 11.27⁄ Parent 0 Teacher 5 634 .13⁄⁄ .06, .20 2.53 Other 0 Multiple 6 369 .04 .04, .13 3.87 Type of assessment aggr 0.52 .47 Questionnaire 9 886 .12⁄⁄ .06, .17 4.50 Peer nomination 4 886 .16⁄⁄ .10, .23 11.27⁄ Observation 0 Multiple 2 117 Informant parenting Parent 13 1767 .12⁄⁄ .07, .16 21.02⁄ Child 0 Observer 0 Other 0 Multiple 2 122 Type of assessment parenting Questionnaire 15 1889 .12⁄⁄ .08, .16 21.58 Observation 0 Other 0 Multiple 0 Same informant aggr-parenting? No 15 1889 .12⁄⁄ .08, .16 21.58 Yes 0 Partially 0 Longitudinal Concurrent 13 1633 .12⁄⁄ .07, .17 19.68 Parenting – aggr 2 256 Controlled for phys aggr? No 15 1889 .12⁄⁄ .08, .16 21.58 Yes 0 Other covariates? No 15 1889 .12⁄⁄ .08, .16 21.58 Yes 0 Publication 0.09 .76 Published paper 9 971 .11⁄⁄ .05, .17 5.92 Dissertation/unpublished data 6 918 .13⁄⁄ .06, .20 14.26⁄ Note: Subsets with fewer than four effect sizes were not included in the moderator tests. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 265 Discussion Four significantly different parenting clusters (i.e., positive, psychologically controlling, negative/ harsh, and uninvolved parenting) emerged from the conceptual analysis that we performed to differentiate various types of parenting constructs used in previous studies and to cluster these constructs into homogeneous groups. As expected, three of these clusters (i.e., positive parenting, negative/harsh parenting, uninvolved parenting) matched Baumrind’s parenting styles rather well. The expert sorters did not reproduce the permissive parenting style as a combination of warmth and lack of monitoring and limit setting. The second cluster included psychologically controlling parenting styles. Psychologically controlling parenting is different from uninvolved parenting by nature. Whereas uninvolved parenting represents lack of care for, awareness of, and sensitivity to children, psychologically controlling parenting includes manipulation of parent–child relationships. For example, uninvolved parents may ignore children as part of neglect, whereas psychologically controlling parents may use the same strategy because they intend to control the way that children act toward their parents. The second cluster (psychologically controlling parenting) is particularly relevant to the present meta-analysis, given that a large body of literature has suggested that this type of parenting is a basis for modeling relational aggression (e.g., Nelson & Crick, 2002). As for the association between these four types of parenting and relational aggression, results demonstrated, as hypothesized, that more positive parenting, including parental warmth and emotional sensitivity, was related to lower levels of relational aggression, and that more harsh parenting, psychologically controlling parenting (i.e., paternal), and uninvolved parenting were associated with increased relational aggression. Harsh as well as uninvolved parenting styles of mothers appeared equally effective in elevating the level of relational aggression in their children. Both parenting styles of mothers and of fathers were associated in similar ways with relational aggression. Further, several moderators of the association between paternal parenting (not maternal parenting) and relational Uninvolved parenting Positive parenting Negative/harsh parenting Psychologically controlling parenting Mother Father .15 Effect size (correlation) * * k = 42 N = 25,538 k = 18 N = 3,161 k = 18 N = 3,288 k = 15 N = 1,889 k = 36 N = 11,691 k = 41 N = 7,938 k = 15 N = 4,227 .10 .05 .00 Note. We plotted the absolute values of the effect sizes; the direction of effects for positive parenting was negative. Fig. 2. Meta-analytic relation between relational aggression and parenting (absolute values including 85% CIs). 266 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 aggression were significant. Children’s gender was an important moderator in that the association between paternal psychologically controlling parenting and increased relational aggression was stronger for girls than for boys. In addition, published studies generated larger effect sizes for the relation between parental positive parenting and relational aggression as compared to unpublished datasets and dissertations. The present study demonstrated that higher levels of both maternal and paternal positive parenting were associated with lower levels of relational aggression. Positive parenting may provide a safe, comfortable environment in which children can learn positive social behaviors such as prosocial behavior and empathy and at the same time inhibit negative social behaviors (Baumrind, 1973, 1978). Research has shown that children with sensitive parents are more likely to be securely attached and as a result, develop a positive cognitive representation of relationships. They are more self-confident and have more confidence in others, which enhances their social skills (Sroufe et al., 2005). Children with such positive blueprint of relationships are also less likely to negatively interpret ambiguous social information and therefore do not react with relational aggression in these ambiguous social situations. Relatedly, past studies have revealed that positive parenting offers children a socializing context that promotes their emotion regulation and social competence (Parke, Cassidy, Burks, Carson, & Boyum, 1992; Zhou et al., 2002), and that these variables in turn may be related to a lower level of aggressive behavior (Eisenberg, Cumberland, et al., 2001; Eisenberg, Gershoff, et al., 2001b). That is, emotion regulation or social skills may mediate the relation between positive parenting and children’s aggressive behavior, including both physical and relational aggression. Taken together, children who experience positive parenting are less likely to exhibit relational aggression perhaps because they develop high levels of social competence and emotion regulation in the context of positive parent–child relationships (see Fig. 3). The present meta-analytic study also showed that more paternal psychologically controlling parenting, but not maternal psychologically controlling parenting, was associated with more relational aggression. One possible explanation for this finding stems from social learning theory (Bandura, 1973) that can be used to explain children’s observational learning of the use of aggression (Patterson, 1986). That is, children may become relationally aggressive by observing their father’s relationally aggressive behaviors. This view suggests that parenting involving the use of psychological control tacPositive parenting Psychologically controlling parenting Uninvolved parenting Negative/harsh parenting Guidance of emotion regulation Observational learning of rel. aggr. /social information processing bias Negative reinforcement of relational aggression Genetics parent Genetics child Interactions with relationally aggressive peers Promoting secure attachment and positive sense of self Bias in social information processing + - + + + + + + + + + - - - - - - - - - - Relational aggression Fig. 3. Theoretical model of the relation between parenting and relational aggression. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 267 tics may serve as a socializing context in which children are more likely to use relationally aggressive strategies in peer interactions (Casas et al., 2006). The literature has also suggested that psychologically controlling parenting may lead to children’s lack of self-esteem and belongingness for which they may compensate by engaging in relationally aggressive behavior (Soenens et al., 2008). Soenens et al. (2008) contended that psychologically controlling parenting may lead to an insecure and negative representation of the self and the parent–child relationship, which in turn may be related to increased relational aggression. It has been documented that a history of insecure attachment is associated with internalizing and externalizing adjustment problems (Allen, Porter, & McFarland, 2007; Fearon et al., 2010) and these adjustment problems generally covary with problems with peers such as relational aggression (Card et al., 2008; Crick & Grotpeter, 1995; Murray-Close, Ostrov, & Crick, 2007). Perhaps children who experience parental psychological control may have an insecure sense of the relationship with the parent and of themselves, which may lead to the formation of conflictual, enmeshed relationships with peers in which they use more relational aggression (see Fig. 3). Relatedly, a self-regulatory system, such as emotion regulation, may serve as an important mediator in the relation between psychological control and children’s relationally aggressive behaviors. This is in line with the literature that has suggested that children who experience psychologically controlling parenting may be unable to develop a positive sense of self-efficacy, psychological autonomy, and emotion regulation (Barber et al., 2005). It may be that excessive use of parental psychological control, which often intrudes on children’s psychological and emotional development creates a negative home environment, in which children have limited opportunities to express themselves properly and learn how to regulate or inhibit negative emotions such as anger, sadness, and upset in social contact. The inability to control negative emotions, in turn, may lead to more relational aggression in the context of peer relationships (see Fig. 3). Contrary to our hypothesis, maternal psychologically controlling parenting behaviors were not associated with relational aggression for boys and girls. However, this does not reject the view that maternal psychological control plays a role in the development of relational aggression. There may be other factors (e.g., hostile attribution biases, attachment, emotion regulation) that can explain the indirect link between psychologically controlling parenting behaviors and children’s relational aggression. For example, studies in this area have demonstrated that maternal hostile attribution biases are positively related to children’s hostile attribution biases (Nelson et al., 2008), which in turn are associated with more relational aggression (e.g., Crick et al., 2002). In contrast, Nelson et al.’s study (2008) revealed that paternal hostile attribution biases were directly predictive of children’s relational aggression. Hence, it is possible that mothers’ psychologically controlling behaviors may indirectly predict children’s relational aggression via their hostile attribution biases and that fathers’ psychologically controlling behavior may serve as a more direct, important predictor of children’s relational aggression than their social-cognitive processes, including hostile attribution biases. In addition, we showed that other maternal parenting strategies were more important in explaining child relational aggression than maternal psychologically controlling parenting. This suggests that researchers studying parenting and relational aggression may benefit from shifting the focus from maternal psychological control towards these other parenting strategies. Importantly, more maternal and paternal use of negative/harsh parenting was relatively strongly related to higher levels of children’s relational aggression. This association was significantly stronger than that for psychologically controlling parenting for both fathers and mothers. This finding may be explained by the idea that socio-cognitive variables such as lack of emotion regulation, self-confidence, and social competence, which may result from experiencing harsh parenting, indirectly contribute to the development and maintenance of relational aggression. Notably, studies that have examined the etiology of physical aggression have suggested that the relation between harsh parenting and children’s use of physical aggression may be mediated by their poor emotion regulation (Chang, Schwartz, Dodge, & Mc Bride-Chang, 2003). The same logic can be applied to relational aggression. In other words, chronic harsh parenting may be the environmental context in which children fail to develop an optimal level of emotion regulation and social competence, which is crucial to cooperatively and harmoniously interact with peers (see Fig. 3). Children who lack sufficient emotion regulation and social skills may find it difficult to inhibit feelings of anger or hostility when they get mad at 268 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 peers, and as a result, display aggressive behaviors toward peers (they may show either physical aggression and/or relational aggression, depending on the provocative context that they encounter). Higher levels of uninvolved parenting were related to higher levels of relational aggression. This is in line with research that generally has demonstrated that uninvolved parenting and lack of monitoring lead to the use of aggression regardless of its form (e.g., Casas et al., 2006; Rubin, Stewart, & Chen, 1995). It is possible that parents, who are uninvolved and are not engaged in limit-setting parenting or interactions with children, provide the socializing context that promotes negative reinforcement for children’s disruptive behavior, including relational aggression (see Fig. 3). Moreover, many uninvolved parents may not have the ability to offer a sufficient level of behavioral control or supervising/monitoring to curve or stop their children’s aggressive behavior. Hence, children who experience uninvolved parenting may fail to develop sufficient inhibitory/self-regulatory skills, which are developmentally crucial to keep aggressive behavior in check (see Fig. 3). Co-occurrence of physical aggression (i.e., the overlapping correlation between relational aggression and physical aggression; Card et al., 2008) may inflate the magnitude of the association between parenting, harsh parenting in particular, and relational aggression. To minimize the confounding effect of physical aggression, the present study used partial correlations (i.e., effect sizes that were obtained after controlling for the effect of physical aggression), whenever available. In addition, we aimed to test whether controlling for the effect of physical aggression alters the strength of the association between four parenting clusters and relational aggression. However, in all sets of studies the number of effect sizes controlling for physical aggression was too small to test for moderation, except for the set on maternal uninvolved parenting. For this subset, we did not find evidence for different effect sizes depending on whether physical aggression was a control variable in the analyses. Future studies should delve into the unique association between parenting and relational aggression above and beyond the contribution of physical aggression. It is unlikely that the mechanisms that have been proposed for the association between the different parenting strategies and relational aggression are exclusively valid for relational aggression. Similar processes have been suggested for physical aggression (e.g., Chang et al., 2003). Unfortunately, there are only a few studies investigating the differential effect of parenting on relational and physical aggression. One of these (Kuppens, Grietens, Onghena, & Michiels, 2009c) found that parental psychological control was exclusively related to the child’s relational aggression whereas parental physical punishment was exclusively associated with overt aggression. However, in our meta-analysis we found that harsh parenting was (quite strongly) related to relational aggression. Since we were not able to test whether this association was unique for relational aggression, it may be mediated by the effect of harsh parenting on physical aggression. More research is needed to disentangle the unique effects of parenting on relational and physical aggression. The majority of past studies examining parenting and relational aggression have examined maternal and paternal parenting behaviors, and comparisons between mothering and fathering within empirical studies are relatively common. However, the findings of these studies concerning differences in mothering and fathering vary substantially with respect to statistical significance. The present meta-analytic study showed that both maternal and paternal parenting play a role in relational aggression, with no significant differences between maternal and paternal parenting. Moderating effect of children’s sex Whereas no moderators were significant for the association between any dimensions of maternal parenting and relational aggression, we found that the child’s sex moderated the effect of paternal psychologically controlling and positive parenting on relational aggression. Paternal psychologically controlling parenting was more strongly related to girls’ relational aggression. In fact, paternal psychologically controlling parenting was not significantly related to boys’ relational aggression. One possible explanation for this finding is that paternal psychologically controlling behavior may be non-normative or at least less common within father–son dyads. Previous studies conducted during the past decades generally have demonstrated that relational aggression is less typical for males (Crick & Grotpeter, 1995; Crick, Ostrov, & Kawabata, 2007; Crick & Zahn-Waxler, 2003), suggesting that relational aggression may not be frequently and reciprocally used by fathers and sons. Hence, it is conceivY. Kawabata et al. / Developmental Review 31 (2011) 240–278 269 able that paternal psychologically controlling behavior, which may be non-normative in father–son relationships, may be unlikely to serve as a model for developing relational aggression for boys, who are typically less engaged in such relationally aversive acts. By contrast, paternal psychologically controlling behavior may be more normative within father–daughter dyads, where fathers may use more indirect, subtle parenting than direct, physical punishment and, thus, daughters may receive parenting that is similar to relational victimization from fathers (i.e., being ignored, manipulated, and controlled by fathers). This fathering style may provide a socializing context in which girls in turn use more relational aggression through observational learning or modeling. Overall, the present metaanalytic review provides a new insight concerning paternal psychologically controlling behavior and girls’ (and boys’) relational aggression, which may not be possible to detect in a single-sample study. Limitations Despite the numerous strengths of the present meta-analytic study, some limitations also need to be addressed. First, as we attempted to minimize the confounding effect of physical aggression, it was not possible to fully control for this effect due to a very small number of pertinent studies. Hence, the unique effect of parenting on relational aggression remains unexamined. Second, whereas the present study revealed that four clusters of parenting were related to the development of relational aggression, the mechanisms mediating this association are unclear. Nevertheless, we have speculated about possible mechanisms. For example, it is possible that psychologically controlling parenting may be the context in which children can observe relational aggression in interactions with their parents and, thus, subsequently generalize relationally aggressive behavior to the context of peer relationships. Alternatively, emotion regulatory processes may mediate the relation between parenting and relational aggression. Third, the number of studies and subjects included in the current set of meta-analyses was comparable to several other influential meta-analyses in similar domains (e.g., Olatunji & Wolitzky-Taylor, 2009; Van der Bruggen et al., Stams, & Bogels, 2008). In our case, the power to detect a combined effect size of r = .06 amounted to more than .90 according to the power analysis provided by Cafri and his colleagues (2010). However, the power of the moderator analyses might have been too low to detect some relevant but small moderator effects. We used the conventional criterion of at least four effect sizes in the smallest moderator sub-set (Bakermans-Kranenburg et al., 2003), but larger numbers of studies might have revealed more significant moderators. Fourth, the directionality of the association between parenting and relational aggression is not clear, and bidirectionality might be the rule instead of the exception. This domain of research is dominated by correlational studies that preclude definite conclusions about direction of effects and causality in general. Intervention studies investigating whether changing parenting would also change children’s use of relational aggression would clarify the issue of directionality. Fifth, we found disappointingly small effect sizes for the associations between parenting dimensions and relational aggression although they may still be relevant from a theoretical and practical perspective (see below). Because the number of studies and subjects included in the meta-analyses is rather large, the effect sizes can be considered reliable and significant estimates of the true population values. Furthermore, clustering might have led to obscuring stronger effect sizes for specific components of the clusters, and with more studies these components might have emerged as separate and relevant clusters showing stronger effect sizes between the parenting component and relational aggression. Sixth, we clustered parenting variables into a small set of broader categories of parenting strategies. The clustering was required on statistical grounds, to avoid multiple testing and inflated significance levels, but it has the disadvantage of a loss of information on the association of specific parental behaviors with relational aggression. The structuring of a myriad of disparate findings into a transparent and economic framework, however, is a distinct advantage of the prior clustering approach. Lastly, the present meta-analytic study did not allow us to examine unique associations between each parenting dimension and children’s relational aggression because most studies included here did not cover all dimensions. Given that many parenting constructs are typically measured as continuous variables, some of the parenting clusters may conceptually and qualitatively overlap with one another. For example, negative/harsh and psychologically controlling parenting potentially include abusive, hostile, and/or inconsistent parenting strategies; however, these parenting dimensions are 270 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 conceptually distinct in that the former represents physical attacks and direct confrontations toward children and the latter entails psychological manipulation of parent–child relationships, which is more subtle by nature. Likewise, uninvolved and psychologically controlling parenting share seemingly similar behavioral strategies (e.g., ignoring, emotional disengagement). However, these parenting clusters are different in terms of whether parents are actively involved in the relationship with the child. While uninvolved parents are typically passive in and indifferent to their children, psychologically controlling parents are indeed involved in parenting but they often treat their children negatively through manipulating the parent–child relationship. Taken together, each cluster of parenting may seem to share some behavioral repertories on the surface; however, they are conceptually different in terms of the goals of the specific parenting strategies and the meaning of the behavior for the child. Nevertheless, given the reality of parenting, it is possible that the same parents use different strategies that may jointly and differentially influence children’s relational aggression. Hence, the findings of the present meta-analytic report need to be interpreted with the understanding of this limitation. Future implications The present meta-analytic study revealed that, as hypothesized, more positive parenting was associated with less relational aggression, and that harsh parenting, (paternal) psychologically controlling parenting, and uninvolved parenting were associated with increased relational aggression. These findings suggest that parenting plays an important role in the development of relational aggression. However, the effect sizes were unexpectedly small, considering the strong emphasis on parenting as a determinant of relational aggression in the research literature (Crick et al., 1999; Nelson & Crick, 2002). Considering Cohen’s (1988, 1992) criteria for strong (d = 0.80), moderate (d = 0.50), and small (d = 0.20) effect sizes, the combined effects we found in the current set of meta-analyses should be seen as small. Also, the magnitudes of our effect sizes seem to be a bit smaller than those found in other meta-analytic studies that examined the relations between parenting behaviors and child outcomes such as anxiety, depression, and delinquency (correlations ranged from .14 to .30; e.g., Hoeve et al., 2009; McLeod, Weisz, & Wood, 2007; McLeod, Wood, & Weisz, 2007). Although Cohen’s criteria are conventions based on previous meta-analyses in a wide range of research domains (Lipsey & Wilson, 2001), they are of course debatable (see for example McCartney & Rosenthal, 2000). In fact, these criteria have been criticized as too conservative by Rosenthal who introduced the concept of Binomial Effect Size Display (BESD; see McCartney and Rosenthal, 2000, for details) and argued that effect sizes as small as r = .03 may still be extremely important for theory development and clinical or practical applications. Nevertheless, the present meta-analytic review revealed rather small effect sizes for the predicted associations compared to other domains of inquiry and, thus, should stimulate investigators of parenting and relational aggression to either modify some of the basic tenets of their theoretical framework, for example by emphasizing the role of child or contextual factors, in explaining relational aggression, or to search for more precise and valid assessments of both parenting constructs and dimensions of relational aggression. Of course, against the background of the behavioral genetics study on the roles of genes and (shared and non-shared) environmental factors in predicting relational aggression by Tackett et al. (2009) one would have expected a stronger association between parenting and relational aggression. They demonstrated that a substantial variance in relational aggression is accounted for by genes (49–63%) but also by shared environmental factors (37–51%) which might largely be due to parenting styles that are shared by all siblings within a family and that make them more alike in terms of relational aggression. We speculate that the traditional ways of measuring parenting styles in this research domain might only tap into few aspects of parenting that seem important for the development of relational aggression, and leave out some important dimensions that still have to be uncovered. For example, direct assessment of parental modeling of relational aggression in interactions with partner or siblings might be a promising avenue for further exploring dimensions of unintentional parenting. Alternatively, widely used (self-report) assessment of parenting may fail to validly index parenting dimensions relevant for relational aggression. The relatively small effect sizes call for new approaches to observe parental interactions as experienced by the children in their daily lives and in stressful situations or during quarrels in which parents serve as models for how to solve Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 271 (marital or discipline) problems in the family. Unfortunately, the current meta-analyses cannot be used to hint at the direction that should be taken as only few moderating factors were found to be significant. In another behavioral genetic study, Brendgen and colleagues reported that in addition to genes and shared environmental factors, non-shared (unique) environmental factors such as influences of peers and friends may explain a fair amount of variance in relational aggression (Brendgen et al., 2005, 2008). Consistent with this finding, some studies have documented that relationships with peers and friends influence the development and maintenance of relational aggression. For example, a study examining the features of relationally aggressive friendships found that friendships of relationally aggressive youth were more exclusive and abusive (Grotpeter & Crick, 1996). This finding is in line with that of Werner and Crick (2004) who found that children who frequently interacted with relationally aggressive friends were more likely to be relationally aggressive within the same friendships a year later. In the same vein, Brendgen et al. (2005) revealed that friends’ relational aggression was directly related to children’s relational aggression above and beyond the genetic influences on this behavior. Hence, it is conceivable that friendships of relationally aggressive children may be closed, exclusive, and enmeshed in nature, which may provide a socializing context in which children use more relational aggression by modeling their friends’ relational aggression. Importantly, there is evidence that parents and parent–child relationships influence children’s selections of friends (as described in Collins, Maccoby, Steinberg, Hetherington, and Bornstein (2000)), so the relation between parenting and relational aggression may also be (partially) mediated by the quality of interactions with friends. Taken together, in addition to the influences of genes and parenting (as shared environmental factors), non-shared environmental factors such as peers and friends may play a crucial role in the development of relational aggression (see Fig. 3). Given the literature that has indicated the interactive effects of genes and environmental factors on aggressive behavior (Caspi et al., 2002; Dodge, 2009), it is possible that genes, shared environmental factors (e.g., parenting), and unique environmental factors (e.g., peers) jointly and in interaction influence the development of relational aggression. That is, the influences of parenting and/or peer relationships may play a larger role in the development of relational aggression for children with a certain genetic make-up. For example, Caspi and colleagues (2002) showed that (male) children who were subjected to harsh and maltreating parenting and who were also carriers of a genotype conferring low levels of MAOA, developed aggressive and antisocial behaviors in (young) adulthood whereas those children with the genotype conferring high levels of MAOA seemed to be (partially) protected against such negative consequences. Children who share half of their genes with their relationally aggressive parents and who experience harsh or controlling parenting may be more likely to display relational aggression, as compared to children without such genotype. Similarly, relative to peers, children with such genotypes and with relationally aggressive friends may be more vulnerable to interpersonal conflicts and, thereby, more prone to relational aggression. Gene by environment interactions might be more powerful in explaining relational aggression than only assessing parenting styles (see Fig. 3). The genes, environment, and gene by environment interactions may also exert an impact on the development of aggression in non-human primates. To date, the majority of studies examining aggression in primates have focused on an overt form of aggression and demonstrated that non-human primates such as monkeys frequently use physical and verbal aggression (e.g., hitting, pulling hairs of enemies, hostile vocalization) to maintain social order and maximize provisions of the resources for the higher-ranked families (Suomi, 2005). Given that some primates are stratified into a strict social hierarchy (Suomi, 2005), relational aggression such as social exclusion, dominance, and isolation (e.g., not to share food with other monkeys in the same group) may be an effective tool to maximize survival of offspring in the given social group. Hence, it is conceivable that similar to humans, non-human primates may exhibit behavioral repertoires that are overt (i.e., competition, physical and verbal aggression) and covert (i.e., social exclusion), and thus, examining the antecedents, manifestations, and consequences of non-humans’ relational aggression is warranted. To elucidate the functional value of relational aggression for survival and enhancement of inclusive fitness primate studies are indispensable. 272 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 Another direction for future studies of relational aggression pertains to internet aggression. With the advent of advanced technology, adolescents and even younger children frequently use internet relational aggression. Internet relational aggression is harmful given that it is anonymous and widespread in nature. Recent studies that examined relational aggression via the internet and digital text messages revealed that such on-line relational aggression was associated with traditional relational aggression (which typically occurs within classrooms) and was predictive of social–psychological maladjustment (e.g., anxiety, loneliness, uneasiness) and negative peer experiences (e.g., traditional peer victimization) above and beyond the contribution of traditional relational aggression (Berger, 2007; Raskauskas & Stoltz, 2007; Werner, Bumpus, & Rock, 2010). Hence, internet relational aggression is an emerging area of study within the field of aggression and examining the mechanisms that may differentially lead to internet aggression versus traditional, classroom-based aggression is an important area for future research. Further, the findings of the present meta-analytic study may help to improve clinical and counseling policies and practices. That is, various aspects of maternal and paternal parenting influence the development of relational aggression, which stresses the importance of developing parent training programs for mothers as well as for fathers that help to prevent or reduce relational aggression. It is important to note that the development of universal instead of targeted relational aggression programs might be preferred as we found rather small effect sizes for the association between parenting and relational aggression. Thus, on basis of the current meta-analysis it is not possible to screen and select specific groups at risk for relational aggression and to target interventions on the parents of these children. However, the effect sizes are sufficiently substantial to allow for universal preventive intervention reaching the population of parents at large (McCartney & Rosenthal, 2000). Although a substantial body of literature has presented a wide variety of interventions for physically aggressive children, research that focuses on preventive intervention of relationally aggression in children has just begun (Ostrov et al., 2009). In sum, four parenting styles (positive parenting, psychologically controlling parenting, harsh parenting, and uninvolved parenting) emerged from the conceptual analysis on the parenting constructs used in previous studies. More positive parenting was related to lower levels of relational aggression, and harsh parenting, (paternal) psychologically controlling parenting, and uninvolved parenting were associated with increased relational aggression. Parenting styles of both mothers and of fathers were associated in similar ways with relational aggression. However, effect sizes were generally rather small and not sufficient to cover the large shared environment components found in some twin studies. More careful assessments of parenting as experienced by the children in a variety of (conflict and stress) settings in interaction with genotypes found to be relevant for increasing children’s vulnerability to controlling, uninvolved, or harsh parenting seem badly needed. Acknowledgments The work on this meta-analysis was supported by a Rubicon Grant from the Netherlands Organization for Scientific Research to Lenneke Alink and the NWO/SPINOZA Prize from the Netherlands Organization for Scientific Research that was awarded to Marinus van IJzendoorn. We are grateful to Nazanin Mohajeri-Nelson for her contribution to the first phase of this project. Finally, we would like to thank the experts in parenting research for their assistance in Study 1. References References marked with an asterisk indicate studies included in the meta-analysis. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum. ⁄Albrecht, A. K., Galambos, N. L., & Jansson, S. (2007). Adolescents’ internalizing and aggressive behaviors and perceptions of parents’ psychological control: A panel study examining direction of effects. Journal of Youth and Adolescence, 36, 673–684. Alink, L. R. A., Cicchetti, D., Kim, J., & Rogosch, F. A. (2009). Mediating and moderating processes in the relation between maltreatment and psychopathology: Mother–child relationship quality and emotion regulation. Journal of Abnormal Child Psychology, 37, 831–843. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 273 Alink, L. R. A., Mesman, J., Van Zeijl, J., Stolk, M. N., Juffer, F., Koot, H. M., et al (2009b). Maternal sensitivity moderates the relation between negative discipline and aggression in early childhood. Social Development, 18, 99–120. Allen, J. P., Porter, M., & McFarland, C. (2007). The relation of attachment security to adolescents’ paternal and peer relationships, depression, and externalizing behavior. Child Development, 78, 1222–1239. ⁄Ando, M., Asakura, T., & Simons-Morton, B. (2005). Psychosocial influences on physical, verbal, and indirect bullying among Japanese early adolescents. Journal of Early Adolescence, 25, 268–297. ⁄Baier, D. (2005). Deviant behavior in adolescence. An empirical comparison of different theoretical explanations. Journal for Sociology of Education and Socialization, 25, 381–398. Bakermans-Kranenburg, M. J., Van IJzendoorn, M. H., & Juffer, F. (2003). Less is more: Meta-analyses of sensitivity and attachment interventions in early childhood. Psychological Bulletin, 129, 195–215. ⁄Baldry, A. C. (2004). The impact of direct and indirect bullying on the mental and physical health of Italian youngsters. Aggressive Behavior, 30, 343–355. Bandura, A. (1973). Aggression: A social learning analysis. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1989). Social cognitive theory. In R. Vasta (Ed.), Annals of child development. Six theories of child development (Vol. 6, pp. 1–60). Greenwich, CT: JAI Press. Barber, B. K., Stolz, H. E., & Olsen, J. A. (2005). Parental support, psychological control, and behavioral control: Assessing relevance across time, method, and culture. Monographs of the Society for Research in Child Development, 70(4). Baumrind, D. (1967). Child care practices anteceding three patterns of preschool behavior. Genetic Psychology Monographs, 75, 43–88. Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology, 4(1, Pt. 2), 1–103. Baumrind, D. (1973). The development of instrumental competence through socialization. In A. D. Pick (Ed.). Minnesota symposium on child psychology (Vol. 7, pp. 3–46). Minneapolis: University of Minnesota Press. Baumrind, D. (1978). Parental disciplinary patterns and social competence in children. Youth & Society, 9, 239–251. Berger, K. S. (2007). Update on bullying at school: Science forgotten? Developmental Review, 27, 90–126. Bjorkqvist, K. (1994). Sex differences in physical, verbal, and indirect aggression: A review of recent research. Sex Roles, 30, 177–188. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. (2009). Computing effect sizes for meta-analysis. Chichester, UK: John Wiley & Sons, Ltd. Borenstein, M., Rothstein, H., & Cohen, J. (2005). Comprehensive meta-analysis: A computer program for research synthesis [Computer software]. Englewood, NJ: Biostat. Bowlby, J. (1969/1982). Attachment and loss, Vol. 1: Attachment. New York: Basic Books. Brendgen, M., Boivin, M., Vitaro, F., Bukowski, W. M., Dionne, G., Tremblay, R. E., et al (2008). Linkages between children’s and their friends’ social and physical aggression: Evidence for a gene–environment interaction? Child Development, 79, 13–29. Brendgen, M., Dionne, G., Girard, A., Boivin, M., Vitaro, F., & Pe´ russe, D. (2005). Examining genetic and environmental effects on social aggression: A study of 6-year-old twins. Child Development, 76, 930–946. ⁄Brown, S. A., Arnold, D. H., Dobbs, J., & Doctoroff, G. L. (2007). Parenting predictors of relational aggression among Puerto Rican and European American school-age children. Early Childhood Research Quarterly, 22, 147–159. ⁄Burnette, M. L., & Reppucci, N. D. (2009). Childhood abuse and aggression in girls: The contribution of borderline personality disorder. Development and Psychopathology, 21, 309–317. Cafri, G., Kromrey, J. D., & Brannick, M. T. (2010). A meta–meta-analysis: Empirical review of statistical power, Type I error rates, effect sizes, and model selection of meta-analyses published in psychology. Multivariate Behavioral Research, 45, 239–270. ⁄Campbell, J. J. (1999). Familial antecedents to children’s relational and overt aggression (Unpublished doctoral dissertation). University of North Carolina, Greensboro. Card, N. A., Stucky, B. D., Sawalani, G. M., & Little, T. D. (2008). Direct and indirect aggression during childhood and adolescence: A meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Development, 79, 1185–1229. ⁄Casas, J. F., Weigel, S. M., Crick, N. R., Ostrov, J. M., Woods, K. E., Jansen Yeh, E. A., et al (2006). Early parenting and children’s relational and physical aggression in the preschool and home contexts. Applied Developmental Psychology, 27, 209–227. Caspi, A., McClay, J., Moffitt, T., Mill, J., Martin, J., Craig, I. W., et al (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851–854. Chang, L., Schwartz, D., Dodge, K., & Mc Bride-Chang, C. (2003). Harsh parenting in relation to child emotion regulation and aggression. Journal of Family Psychology, 17, 598–606. Chen, X., Dong, Q., & Zhou, H. (1997). Authoritative and authoritarian parenting practices and social and school performance in Chinese children. International Journal of Behavioral Development, 21, 855–873. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon (Ed.). In & N. Eisenberg (Eds.). Handbook of child psychology (Vol. 3, pp. 779–862). New York: Wiley. Collins, W. A., & Laursen, B. (Eds.). (1999). The Minnesota symposia on child psychology: Relationships as developmental contexts (Vol. 30). Hillsdale, NJ: Erlbaum. Collins, W. A., Maccoby, E. E., Steinberg, L., Hetherington, E. M., & Bornstein, M. H. (2000). Contemporary research on parenting: The case for nature and nurture. American Psychologist, 55, 218–232. Collins, W. A., & Steinberg, L. (2006). Adolescent development in interpersonal context. In W. Damon & N. Eisenberg (Eds.). Handbook of child psychology (Vol. 4, pp. 1003–1067). New York: Cambridge University Press. Côté, S. M., Vaillancourt, T., Barker, E. D., Nagin, D., & Tremblay, R. E. (2007). The joint development of physical and indirect aggression: Predictors of continuity and change during childhood. Development and Psychopathology, 19, 37–55. Crick, N. R. (1995). Relational aggression: The role of intent attributions, feelings of distress, and provocation type. Development and Psychopathology, 7, 313–322. Crick, N. R. (1997). Engagement in gender normative versus nonnormative forms of aggression: Links to social–psychological adjustment. Developmental Psychology, 33, 610–617. 274 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 ⁄Crick, N. R. (1999a). [Project KIDS]. Unpublished raw data. ⁄Crick, N. R. (1999b). Unpublished raw data [Japan Project]. ⁄Crick, N. R. (2009). Unpublished raw data [Pre to K Project]. Crick, N. R., Bigbee, M. A., & Howes, C. (1996). Gender differences in children’s normative beliefs about aggression: How do I hurt thee? Let me count the ways. Child Development, 67, 1003–1014. Crick, N. R., Casas, J. F., & Mosher, M. (1997). Relational and overt aggression in preschool. Developmental Psychology, 33, 579–588. Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social information processing mechanisms in children’s social adjustment. Psychological Bulletin, 115, 74–101. Crick, N. R., & Grotpeter, J. K. (1995). Relational aggression, gender, and social–psychological adjustment. Child Development, 66, 710–722. Crick, N. R., Grotpeter, J. K., & Bigbee, M. A. (2002). Relationally and physically aggressive children’s intent attributions and feelings of distress for relational and instrumental peer provocations. Child Development, 73, 1134–1142. Crick, N. R., Nelson, D. A., Morales, J. R., Cullerton-Sen, C., Casas, J. F., & Hickman, S. E. (2001). Relational victimization in childhood and adolescence: I hurt you through the grapevine. In J. Juvonen & S. Graham (Eds.), Peer harassment in school: The plight of the vulnerable and victimized (pp. 196–214). New York: Guilford Press. Crick, N. R., Ostrov, J. M., & Kawabata, Y. (2007). Relational aggression and gender: An overview. In D. J. Flannery, A. T. Vazsonyi, & I. D. Waldman (Eds.), The Cambridge handbook of violent behavior (pp. 245–259). New York: Cambridge University Press. Crick, N. R., Werner, N. E., Casas, J. F., O’Brien, K. M., Nelson, D. A., Grotpeter, J. K., et al (1999). Childhood aggression and gender: A new look at an old problem. In D. Bernstein (Ed.), Nebraska symposium on motivation (pp. 75–141). Lincoln: University of Nebraska Press. Crick, N. R., & Zahn-Waxler, C. (2003). The development of psychopathology in females and males: Current progress and future challenges. Development and Psychopathology, 15, 719–742. Crockenberg, S. C., Leerkes, E. M., & Lekka, S. K. (2007). Pathways from marital aggression to infant emotion regulation: The development of withdrawal in infancy. Infant Behavior and Development, 30, 97–113. ⁄Cullerton-Sen, C., Cassidy, A. R., Murray-Close, D., Cicchetti, D., Crick, N. R., & Rogosch, F. A. (2008). Childhood maltreatment and the development of relational and physical aggression: The importance of a gender-informed approach. Child Development, 79, 1736–1751. Darling, N., & Steinberg, L. (1993). Parenting style as context: An integrative model. Psychological Bulletin, 113, 487–496. De Wolff, M. S., & Van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development, 68, 571–591. Dodge, K. A. (1980). Social cognition and children’s aggressive behavior. Child Development, 51, 162–170. Dodge, K. A. (2009). Mechanisms of gene–environment interaction effects in the development of conduct disorder. Perspectives in Psychological Science, 4, 408–414. ⁄Donovan, K. L. (2009). Trajectories of maternal verbal aggression across the early adolescent years: What patterns are most harmful to low-income community adolescents? (Unpublished doctoral dissertation). Columbia University. ⁄Doyle, H. (2010). A longitudinal study of relational aggression among females using hierarchical linear modeling (Unpublished doctoral dissertation). Kent State University. Duval, S., & Tweedie, R. (2000a). A nonparametric ‘‘trim and fill’’ method for accounting for publication bias in meta-analysis. Journal of the American Statistical Association, 95, 89–98. Duval, S., & Tweedie, R. (2000b). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56, 455–463. Eisenberg, N., Cumberland, A., Spinrad, T. L., Fabes, R. A., Shepard, S. A., Reiser, M., et al (2001a). The relations of regulation and emotionality to children’s externalizing and internalizing problem behavior. Child Development, 72, 1112–1134. Eisenberg, N., Gershoff, E. T., Fabes, R. A., Shepard, S. A., Cumberland, A. J., Lososya, S. H., et al (2001b). Mothers’ emotional expressivity and children’s behavior problems and social competence: Mediation through children’s regulation. Developmental Psychology, 37, 475–490. Fearon, R. P., Bakermans-Kranenburg, M. J., van IJzendoorn, M. H., Lapsley, A. M., & Roisman, G. I. (2010). The significance of insecure attachment and disorganization in the development of children. Child Development, 81, 435–456. ⁄Foo, G. B. (2002). Marital conflict and childhood relational aggression: A gender-balanced approach to understanding adjustment problems in middle childhood (Unpublished doctoral dissertation). University of Minnesota. ⁄Foster, H. A. (2001). Neighborhood and family contexts of gendered aggression in childhood (Unpublished doctoral dissertation). University of Toronto. ⁄Gaertner, A. E., Rathert, J. L., Fite, P. J., Vitulano, M., Wynn, P. T., & Harber, J. (2010). Sources of parental knowledges as moderators of the relation between parental psychological control and relational and physical/verbal aggression. Journal of Child and Family Studies, 19, 607–616. Galen, B. R., & Underwood, M. K. (1997). A developmental investigation of social aggression among children. Developmental Psychology, 33, 589–600. Ge, X., Conger, R. D., Cadoret, R. J., Neiderhiser, J. M., Yates, W., Troughton, E., et al (1996). The developmental interface between nature and nurture: A mutual influence model of child antisocial behavior and parent behavior. Developmental Psychology, 32, 574–589. Goldstein, H., & Healy, M. J. R. (1995). The graphical presentation of collection of means. Journal of the Royal Statistical Society, Series A (Statistics m Society), 158, 175–177. Grotpeter, J. K., & Crick, N. R. (1996). Relational aggression, overt aggression, and friendship. Child Development, 67, 2328–2338. Harrist, A. W., Pettit, G. S., Dodge, F. A., & Bates, J. E. (1994). Dyadic synchrony in mother–child interaction: Relation with children’s subsequent kindergarten adjustment. Family Relations, 43, 417–424. Hart, C. H., DeWolf, M., Wozniak, P., & Burts, D. C. (1992). Linkages among preschoolers’ playground behavior, outcome expectations, and parental disciplinary strategies. Early Education and Development, 3, 265–283. ⁄Hart, C. H., Nelson, D. A., Robinson, C. C., Olsen, S. F., & McNeilly-Choque, M. K. (1998). Overt and relational aggression in Russian nursery-school-age children: Parenting style and marital linkages. Developmental Psychology, 34, 687–697. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 275 Hartup, W. W., & Rubin, Z. (Eds.). (1986). Relationships and development. Hillsdale, NJ: Erlbaum. Hartup, W. W., & Stevens, N. (1997). Friendships and adaptation in the life course. Psychological Bulletin, 121, 355–370. Hawker, D. S. J., & Boulton, M. (2000). Twenty years’ research on peer victimization and psychosocial adjustment: A metaanalytic review of cross-sectional studies. Journal of Child Psychology and Psychiatry and Allied Disciplines, 41, 441–455. ⁄Herrenkohl, T. I., McMorris, B. J., Catalano, R. F., Abbott, R. D., Hemphill, S. A., & Toumbourou, J. W. (2007). Risk factors for violence and relational aggression in adolescence. Journal of Interpersonal Violence, 22, 386–405. Hoeve, M., Dubas, J. S., Eichelsheim, V. I., Van der Laan, P. H., Smeenk, W., & Gerris, J. R. M. (2009). The relationship between parenting and delinquency: A meta-analysis. Journal of Abnormal Child Psychology, 37, 749–775. ⁄Hutchison, R. (2003). Quality of parent–adolescent relationship and gender influences on aggression among early adolescents (Unpublished doctoral dissertation). Miami University. Izard, C., Stark, K., Trentacosta, C., & Schultz, D. (2008). Beyond emotion regulation: Emotion utilization and adaptive functioning. Child Development Perspectives, 2, 156–163. Jaffee, S. R., Caspi, A., Moffitt, T. E., Polo-Tomas, M., Price, T. S., & Taylor, A. (2004). The limits of child effects: Evidence for genetically mediated child effects on corporal punishment but not on physical maltreatment. Developmental Psychology, 40, 1047–1058. ⁄Jung, K. (2005). The impact of family stressors, interparental conflict, and parenting behaviors on children’s overt and relational aggression: A focus on Korean families (Unpublished doctoral dissertation). Syracuse University. Kawabata, Y., Crick, N. R., & Hamaguchi, Y. (2010a). Relational aggression, physical aggression, and social–psychological adjustment in a Japanese sample: The moderating role of positive and negative friendship quality. Journal of Abnormal Child Psychology, 38, 471–484. Kawabata, Y., Crick, N. R., & Hamaguchi, Y. (2010b). The role of culture in relational aggression: Links with social–psychological adjustment problems in Japanese and US school-aged children. International Journal of Behavioral Development, 34, 354–362. Kistner, J., Counts-Allan, C., Dunke, S., Drew, C. H., David-Ferdon, C., & Lopez, C. (2010). Sex differences in relational and overt aggression in the late elementary school years. Aggressive Behavior, 36, 282–291. ⁄Kuppens, S., Grietens, H., Onghena, P., & Michiels, D. (2009a). Relations between parental psychological control and childhood relational aggression: Reciprocal in nature? Journal of Clinical Child and Adolescent Psychology, 38, 117–131. Kuppens, S., Grietens, H., Onghena, P., & Michiels, D. (2009b). Measuring parenting dimensions in middle childhood: Multitrait– multimethod analysis of child, mother, and father ratings. European Journal of Psychological Assessment, 25, 133–140. Kuppens, S., Grietens, H., Onghena, P., & Michiels, D. (2009c). Associations between parental control and children’s overt and relational aggression. British Journal of Developmental Psychology, 27, 607–623. Ladd, G. W., & Pettit, G. S. (2002). Parenting and the development of children’s peer relationships. In M. H. Borenstein (Ed.). Handbook of parenting (2nd ed.) (Vol. 5, pp. 377–409). Mahwah, NJ: Lawrence Erlbaum Associates. ⁄Li, Y. (2007). Parental influences on children’s social and overt aggression in China: Effects of parenting behavior, marital conflict, and parent–child attachment (Unpublished doctoral dissertation). Duke University. ⁄Lindsey, E. W., Chambers, J. C., Frabutt, J. M., & Mackinnon-Lewis, C. (2009). Marital conflict and adolescents’ peer aggression: The mediating and moderating role of mother–child emotional reciprocity. Family Relations: An Interdisciplinary. Journal of Applied Family Studies, 58, 593–606. Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Applied social research methods series (Vol. 49). Thousand Oaks, CA: Sage Publications. Loukas, A., Paulos, S. K., & Robinson, S. (2005). Early adolescent social and overt aggression: Examining the roles of social anxiety and maternal psychological control. Journal of Youth and Adolescence, 34, 335–345. Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the family: Parent–child interaction. In E. M. Hetherington (Ed.), Handbook of child psychology. In & P. H. Mussen (Eds.). Socialization, personality, and social development (Vol. 4, pp. 1–101). New York: Wiley. McCartney, K., & Rosenthal, R. (2000). Effect size, practical importance, and social policy for children. Child Development, 71, 173–180. McLeod, B. D., Weisz, J. R., & Wood, J. J. (2007). Examining the association between parenting and childhood depression: A metaanalysis. Clinical Psychology Review, 27, 986–1003. McLeod, B. D., Wood, J. J., & Weisz, J. R. (2007). Examining the association between parenting and childhood anxiety: A metaanalysis. Clinical Psychology Review, 27, 155–172. ⁄McNamara, K. A., Selig, J. P., & Hawley, P. H. (2010). A typological approach to the study of parenting: Associations between maternal parenting patterns and child behavior and social reception. Early Child Development and Care, 180, 1185–1202. ⁄McNeill, S. K. (2002). The relation between parental discipline practices, gender, and children’s overt and relational aggression (Unpublished doctoral dissertation). University of Minnesota. Michiels, D., Grietens, H., Onghena, P., & Kuppens, S. (2008). Parent–child interactions and relational aggression in peer relationships. Developmental Review, 28, 522–540. Mize, J., & Pettit, G. S. (1997). Mother’s social coaching, mother–child relationship style, and children’s peer competence: Is the medium the message? Child Development, 68, 312–332. ⁄Mrug, S. P., Elliott, M. P., Gilliland, M. J. P., Grunbaum, J. J. P., Tortolero, S. R. P., Cuccaro, P. P., et al (2008). Positive parenting and early puberty in girls: Protective effects against aggressive behavior. Archives of Pediatrics & Adolescent Medicine, 162, 781–786. ⁄Murray, K. W., Haynie, D. L., Howard, D. E., Cheng, T. L., & Simons-Morton, B. (2010). Perceptions of parenting practices as predictors of aggression in a low-income, urban, predominately African American middle school sample. Journal of School Violence, 9, 174–193. Murray-Close, D., Ostrov, J. M., & Crick, N. R. (2007). A short-term longitudinal study of growth of relational aggression during middle childhood: Associations with gender, friendship intimacy, and internalizing problems. Development and Psychopathology, 19, 187–203. Nelson, D. A., & Coyne, S. M. (2009). Children’s intent attributions and feelings of distress: Associations with maternal and paternal parenting practices. Journal of Abnormal Child Psychology, 37, 223–237. 276 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 Nelson, D. A., & Crick, N. R. (2002). Parental psychological control: Implications for childhood physical and relational aggression. In B. K. Barber (Ed.), Intrusive parenting: How psychological control affects children and adolescents (pp. 168–189). Washington, DC: American Psychological Association. Nelson, L. J., Hart, C. H., Wu, B., Yang, C., Roper, S. O., & Jin, S. (2006). Relations between Chinese mothers’ parenting practices and social withdrawal in early childhood. International Journal of Behavioral Development, 30, 261–271. ⁄Nelson, D. A., Hart, C. H., Yang, C., Olsen, J. A., & Jin, S. (2006). Aversive parenting in China: Associations with child physical and relational aggression. Child Development, 77, 554–572. Nelson, D. A., Mitchell, C., & Yang, C. (2008). Intent attributions and aggression: A study of children and their parents. Journal of Abnormal Child Psychology, 36, 793–806. Olatunji, B. O., & Wolitzky-Taylor, K. B. (2009). Anxiety sensitivity and the anxiety disorders: A meta-analytic review and synthesis. Psychological Bulletin, 135, 974–999. ⁄Ostrov, J. M., & Bishop, C. M. (2008). Preschoolers’ aggression and parent–child conflict: A multiinformant and multimethod study. Journal of Experimental Child Psychology, 99, 309–322. Ostrov, J. M., & Crick, N. R. (2007). Forms and functions of aggression during early childhood: A short-term longitudinal study. School Psychology Review, 36, 22–43. Ostrov, J. M., Massetti, G. M., Stauffacher, K., Godleski, S. A., Hart, K. C., Karch, K. M., et al (2009). An intervention for relational and physical aggression in early childhood: A preliminary Study. Early Childhood Research Quarterly, 24, 15–28. ⁄Park, J.-H., Essex, M. J., Zahn-Waxler, C., Armstrong, J. M., Klein, M. H., & Goldsmith, H. (2005). Relational and overt aggression in middle childhood: Early child and family risk factors. Early Education and Development, 16, 233–256. Parke, R. D., Cassidy, J., Burks, V. S., Carson, J. L., & Boyum, L. (1992). Familial contribution to peer competence among young children: The role of interactive and affective processes. In R. D. Parke & G. W. Ladd (Eds.), Family–peer relationships: Modes of linkage (pp. 107–134). Hillsdale NJ: Erlbaum. Patterson, G. R. (1986). Performance models for antisocial boys. American Psychologist, 41, 432–444. ⁄Paulos, S. K. (2007). Examining the roles of family environment and internalizing symptoms on early adolescent social aggression: A one-year longitudinal study (Unpublished doctoral dissertation). University of Texas, Austin. ⁄Penney, T. L. (2007). Relational aggression and victimization in early adolescence: Assessing the influence of parenting (Unpublished doctoral dissertation). University of New Brunswick. ⁄Perkins, J. N. (2009). Foster parenting practices as predictors of foster child outcomes (Unpublished doctoral dissertation). University of Ottawa. ⁄Pernice-Duca, F., Taiariol, J., & Yoon, J. (2010). Perceptions of school and family climates and experiences of relational aggression. Journal of School Violence, 9, 303–319. Pettit, G. S., Bates, J. E., & Dodge, K. A. (1997). Supportive parenting, ecological context, and children’s adjustment: A seven-year longitudinal study. Child Development, 68, 908–923. Raskauskas, J., & Stoltz, A. D. (2007). Involvement in traditional and electronic bullying among adolescents. Developmental Psychology, 43, 564–575. ⁄Reed, T. J., Goldstein, S. E., Morris, A. S., & Keyes, A. W. (2008). Relational aggression in mothers and children: Links with psychological control and child adjustment. Sex Roles, 59, 39–48. ⁄Risser, S. D. (2007). Interparental conflict, psychological control, and children’s social aggression (Unpublished doctoral dissertation). University of Texas, Dallas. ⁄Rodgers, C. S. (2001). Family factors and relational aggression (Unpublished doctoral dissertation). University of Oregon. Romano, E., Tremblay, R., Boulerice, B., & Swisher, R. R. (2005). Multilevel correlates of childhood physical aggression and prosocial behavior. Journal of Abnormal Child Psychology, 33, 565–578. Rose, A. J., & Rudolph, K. D. (2006). A review of sex differences in peer relationship processes: Potential trade-offs for the emotional and behavioral development of girls and boys. Psychological Bulletin, 132, 98–131. Rosenthal, R. (1979). The ‘‘file drawer problem’’ and tolerance for null results. Psychological Bulletin, 86, 638–641. Rosenthal, R. (1991). Meta-analysis: A review. Psychosomatic Medicine, 53, 247–271. Rowe, D. C. (1994). The limits of family influence. New York: Guilford. Rubin, K. H., Stewart, S. L., & Chen, X. (1995). Parents of aggressive and withdrawn children. In M. H. Bornstein (Ed.). Handbook of parenting: Children and parenting (Vol. 1, pp. 225–284). Mahwah, NJ: Erlbaum. ⁄Russell, A., Hart, C. H., Robinson, C. C., & Olsen, S. F. (2003). Children’s sociable and aggressive behavior with peers: A comparison of the US and Australian, and contributions of temperament and parenting styles. International Journal of Behavioral Development, 27, 74–86. ⁄Sandstrom, M. J. (2007). A link between mothers’ disciplinary strategies and children’s relational aggression. British Journal of Developmental Psychology, 25, 399–407. ⁄Soenens, B., Vansteenkiste, M., Goossens, L., Duriez, B., & Niemiec, C. P. (2008). The intervening role of relational aggression between psychological control and friendship quality. Social Development, 17, 661–681. Sroufe, L. A. (1995). Emotional development: The organization of emotional life in the early years. New York: Cambridge University Press. Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005). The development of the person: The Minnesota study of risk and adaptation from birth to adulthood. New York: Guilford Press. Steinberg, L. D., Lamborn, S. D., Darling, N., Mounts, N. S., & Dornbusch, S. M. (1994). Over-time changes in adjustment and competence among adolescents from authoritative, authoritarian, indulgent, and neglectful families. Child Development, 65, 754–770. Stocker, C. (2000). Adolescents’ relational and physical aggression: Links with family relationships and psychological adjustment. Poster presented at the Eighth Biennial Meeting of the Society for Research on Adolescence (SRA), March, Chicago, IL. Suomi, S. J. (2005). Genetic and environmental factors influencing the expression of impulsive aggression and serotonergic functioning in rhesus monkeys. In R. Tremblay, W. W. Hartup, & J. Archer (Eds.), Developmental origins of aggression (pp. 63–82). New York: Guilford. Y. Kawabata et al. / Developmental Review 31 (2011) 240–278 277 Sutton, A. J., Duval, S. J., Tweedie, R. L., Abrams, K. R., & Jones, D. R. (2000). Empirical assessment of effect of publication bias on meta-analyses. British Medical Journal, 320, 1574–1577. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. New York: Harper & Row. Tackett, J. L., Waldman, I. D., & Lahey, B. B. (2009). Etiology and measurement of relational aggression: A multi-informant behavior genetic investigation. Journal of Abnormal Psychology, 118, 722–733. ⁄Tseng, W.-L. & Kawabata, Y. (2008). Relational aggression in Taiwanese children. Unpublished raw data. ⁄Underwood, M. K., Beron, K. J., Gentsch, J. K., Galperin, M. B., & Risser, S. D. (2008). Family correlates of children’s social and physical aggression with peers: Negative interparental conflict strategies and parenting styles. International Journal of Behavioral Development, 32, 549–562. Underwood, M. K., Beron, K. J., & Rosen, L. H. (2009). Continuity and change in social and physical aggression from middle childhood through early adolescence. Aggressive Behavior, 35, 357–375. doi:10.1002/ab.20313. ⁄Updegraff, K. A., Thayer, S. M., Whiteman, S. D., Denning, D. J., & McHale, S. M. (2005). Relational aggression in adolescents’ sibling relationships: Links to sibling and parent–adolescent relationship quality. Family Relations, 54, 373–385. ⁄Vaillancourt, T., Miller, J. L., Fagbemi, J., Côté, S., & Tremblay, R. E. (2007). Trajectories and predictors of indirect aggression: Results from a nationally representative longitudinal study of Canadian children aged 2–10. Aggressive Behavior, 33, 314–326. Van der Bruggen, C. O., Stams, G. J. M., & Bögels, S. M. (2008). Research review: The relation between child and parent anxiety and parental control: A meta-analytic review. Journal of Child Psychology and Psychiatry, 49, 1257–1269. Van IJzendoorn, M. H., Bakermans-Kranenburg, M. J., & Alink, L. R. A. (2012). Meta-analysis in developmental science. In B. Laursen, T. Little, & N. Card (Eds.), Handbook of developmental research methods. New York: Guilford Press. Van IJzendoorn, M. H., Juffer, F., & Klein Poelhuis, C. W. (2005). Adoption and cognitive development: A meta-analytic comparison of adopted and non-adopted children’s IQ and school performance. Psychological Bulletin, 131, 301–316. Verkes, R.-J., Van der Kloot, W. A., & Van der Meij, J. (1989). The perceived structure of 176 pain descriptive words. Pain, 38, 219–229. ⁄Wang, J., Iannotti, R. J., & Nansel, T. R. (2009). School bullying among adolescents in the United States: Physical, verbal, relational, and cyber. Journal of Adolescent Health, 45, 368–375. Werner, N. E., Bumpus, M. F., & Rock, D. (2010). Involvement in internet aggression during early adolescence. Journal of Youth and Adolescence, 39, 607–619. Werner, N. E., & Crick, N. R. (1999). Relational aggression and social–psychological adjustment in a college sample. Journal of Abnormal Psychology, 108, 615–623. Werner, N. E., & Crick, N. R. (2004). Peer relationship influences on the development of relational and physical aggression during middle childhood: The roles of peer rejection and association with aggressive friends. Social Development, 13, 495–513. ⁄Werner, N. E., Senich, S., & Przepyszny, K. A. (2006). Mothers’ responses to preschoolers’ relational and physical aggression. Journal of Applied Developmental Psychology, 27, 193–208. ⁄Yu, J. J., & Gamble, W. C. (2008). Familial correlates of overt and relational aggression between young adolescent siblings. Journal of Youth and Adolescence, 37, 655–673. Zhou, Q., Eisenberg, N., Losoya, S., Fabes, R. A., Reiser, M., Guthrie, I. K., et al (2002). The relations of parental warmth and positive expressiveness to children’s empathy-related responding and social functioning: A longitudinal study. Child Development, 73, 893–915. 278 Y. Kawabata et al. / Developmental Review 31 (2011) 240–278