Attachment to Parents and Peers as a Parameter of Bullying and Victimization

The purpose of the present study was threefold. First, we tried to investigate whether the quality of attachment with parents and peers predicts bullying and victimization. Second, we also attempted a moderation analysis in order to examine whether the relationship between quality of attachments and bullying is moderated by the child’s gender. Finally, we explored whether there are significant differences in the quality of attachment between children identified as bullies, victims, bully/victims, and uninvolved. The participants were 303 fifth and sixth grade children with a mean age of 11.06 years that completed the Revised Bullying and Victimization Questionnaire and the Inventory of Parent and Peer Attachment.The results showed that poor quality of attachment with parents and peers predicts bullying and victimization.Moderation analysis revealed that the link between quality of attachment and bullying and victimization is significantly stronger for girls. Also, as hypothesized, bullies and bully/victims manifest the worst quality of attachment with parents and peers. The results are discussed with the framework of attachment and aggression theory, exploring the pathways that explain the association between poor attachment and externalizing problems during late childhood.


Introduction
During the past two decades, the bullying phenomenon has drawn an extensive amount of attention from researchers and practitioners alike [1][2][3].Bullying is an intentional and often repeated set of behaviours that aim at taunting, humiliating, physically hurting, and socially isolating the victim.It is a distinct form of systematic aggression which usually involves an imbalance of power, and it seriously affects the lives of a significant minority of children [3].Research shows that bullying may be associated with externalizing problems such as poor social adjustment, greater risk of substance use, and generally with psychological difficulties in later life [4][5][6].On the other hand, victimization is associated with internalizing problems such as depression and anxiety [3,7,8], poorer selfesteem, and interpersonal difficulties in adulthood [9,10].
1.1.Parental and Peer Correlates of Bullying.Empirical studies have provided an extensive body of evidence regarding the predictors and consequences of bullying and victimization.Thus, it is now widely believed that the contextual factors such as maladaptive parental characteristics and peer relationships may increase the chances of children's involvement in bullying [11][12][13].Georgiou [11] has reported that the authoritarian parental style was associated with higher levels of bullying, while the permissive parental style was associated with victimization.In contrast, the authoritative parental style was negatively linked to children's involvement in bullying.In the same line, Shetgiri et al. [12] showed that parents who talked with their children and met all or most of their friends had children with lower bullying odds.The same researchers also found that parents who felt angry with their children or felt that their children bothered them had children with higher bullying odds.They also found that suboptimal maternal mental health was associated with higher bullying involvement.Recently, Knous-Westfall et al. [13] showed that parental reports of any intimate partner violence predicted higher offspring overt peer victimization, while severe intimate partner violence predicted higher offspring relational peer bullying and overt peer victimization.An additional parental factor that seems to be a significant predictor of bullying behaviour is maternal depression.This factor is associated with children's involvement in bullying and other forms of externalizing behaviours [14,15].
Another line of empirical research focuses on the role of peers in bullying [16][17][18][19].For example, Sharp [20] argued that the social context of individual classes or year groups can influence the likelihood of bullying.Also, Salmivalli et al. [19], in a thorough investigation of peer correlates of bullying, found that peers can be classified into those who act as assistants of bullies by encouraging the perpetrators, defenders of victims, and passive bystanders who neither encourage the act nor intervene.While most of the studies focus on the negative role of peers, there are also researchers who claim that peers can play a positive role in reducing bullying when taught how to react when such incidents take place [21].
Numerous studies have shown the link between the quality of children's relationships with their parents to problem behaviors and emotional difficulties [22][23][24][25][26][27].Maladaptive parental involvement [28] and authoritarian parental style [11] have been shown to predict aggressive behavior in children and adolescents.On the other hand, low family involvement [29], permissive parental style, and maternal overprotectiveness [11,15] predict peer victimization.
Earlier studies have found that children who are insecure avoidant in attachment are more likely to demonstrate antisocial traits and callous-unemotional characteristics [30], while children with secure attachment have shown lower levels of aggression [31][32][33].Williams and Kennedy [34] found that girls were more likely to be physically aggressive when they had higher levels of attachment avoidance to their mothers and higher levels of attachment anxiety with their fathers.In the same study, it was also found that boys were more likely to engage in relational aggression when they experienced higher levels of attachment anxiety to their fathers.
Perren and Alsaker [35] examined peer relationships of children who get involved in bullying.The results suggest that bullies were less prosocial, but they also had more leadership skills than noninvolved children.They belonged to larger social groups and were frequently affiliated with other bullies or bully/victims.Victims, on the other hand, were more submissive, had fewer leadership skills, were more withdrawn, more isolated, less cooperative, less sociable, and frequently had no playmates.Finally, bully/victims were less cooperative, less sociable, and they had fewer playmates than noninvolved children.Interestingly, Hodges et al. [36] found that the lack of supportive friends and the presence of peer rejection are correlated with bullying and victimization, while friendships and support by peers provide a buffer against victimization.In sum, it seems that peer relationships play an important role in increasing or preventing the likelihood of bullying and victimization.
Several studies have focused on investigating the correlates of the distinct type of roles that children and adolescents have in relation to bullying activity.These studies do so in order to identify the prevalent characteristics and risk factors of children that are classified as bullies, victims, and bully/victims but also to explore how these children differ from children that stay away from bullying [37,38].The results show that bullies usually have relatively low levels of anxiety, high levels of antisocial and impulsive behaviour [3,39,40], and reveal a strong need to dominate others [41].Additionally, Kumpulainen and Räsänen [42] have found that bullies suffer from externalizing problems, such as aggressiveness, attention deficit, hyperactivity and conduct disorders.Furthermore, Kaltiala-Heino et al. [43] found that bullies exhibit internalizing problems such as depression and anxiety.Victims, on the other hand, tend to show high levels of depression and psychological distress [8,[43][44][45] and low levels of self-esteem [37].Other researchers found that victims tend to feel lonely and have poor peer relationships [7,46].Finally, bully/victims seem to have more serious pathogenic characteristics than bullies and victims [38,47,48].Empirical studies have repeatedly shown that bully/victims suffer from severe psychosocial problems.Specifically, Georgiou and Stavrinides [38,47] found that bully/victims are even more isolated than victims and more temperamental than bullies.Additionally, some studies suggest that bully/victims suffer from internalizing problems such as depression and anxiety [49,50], while other studies suggest that they also display externalizing problems such as aggression and hyperactivity [44,50].In general, bullies, victims, and bully/victims report a reduced sense of life satisfaction [51].
To conclude, the current research addresses questions about the relationship between quality of attachment with parents and peers and bullying/victimization in order to unfold the role of mother, father, and peers.Additionally, moderation effects such as children's gender are included in the analysis in order to bridge a relative literature gap that often leaves behind the conditions in which certain processes take place.How the quality of attachment is differentiated in boys or girls?And, finally, we address questions in order to investigate how the quality of attachment is different in bullies, victims, bully/victims, and uninvolved children.It is important to unfold these parameters in order to develop and design more effective prevention and intervention programs including parents and peers.

The Present Study.
The purpose of the present study is threefold.First, we tried to investigate whether the quality of attachment with parents and peers predicts bullying and victimization.Second, we also attempted a moderation analysis in order to examine whether the relationship between quality of attachments and bullying is moderated by the child's gender.Finally, we explored whether there are significant differences in the quality of attachment between children identified as bullies, victims, bully/victims, and uninvolved.
Based on prior research, the following hypotheses were stated and tested.
(1) The quality of attachment with parents and peers will be related to children's involvement in bullying experiences.That is, low levels of communication and trust and high levels of alienation will predict bullying and victimization.(2) The relationship between quality of attachment with parents and peers and bullying will be moderated by the child's gender.More specifically, the relationship between attachment and bullying will be stronger for girls.
(3) The quality of attachment between children identified as bullies, victims, bully/victims, and uninvolved is different.Uninvolved children will have higher scores in positive aspects of quality of attachment with parents and peers, while bullies and bully/victims will have higher scores in negative aspects of quality of attachment with parents and peers.

Instruments and Procedures.
The sampling procedure for the present study was as follows.Eleven public elementary schools were randomly selected (out of a total of about 200).The only criterion used for the selection was the school's location (urban versus rural).Since the demographic representation of Cyprus is about 70% urban and 30% rural, eight of the selected schools were located in urban areas, and three schools were located in small, rural communities.In this sample of schools, we used stratified cluster sampling where grades 5 and 6 represented the strata and each classroom represented a cluster.Specifically, strata are represented by age levels, while cluster is represented by each class.Using this procedure, we randomly generated a group of twelve classrooms that comprised a sample of 303 students.Twenty-one students returned incomplete questionnaires, and, therefore, they were excluded from the final sample.Also, eleven students were absent from school at the time of data collection.Without violating anonymity, we asked the principals of the participating schools to briefly describe the profiles of these students, pointing out any known difficulties that these students have with bullying or victimization.This investigation showed that students who were absent during the data collection of the present study were not different than the rest of the students in terms of bullying or victimization experiences at school.The participating children completed two instruments: the Revised Bullying and Victimization Questionnaire (BVQ-R) [52] and the Inventory of Parent and Peer Attachment-IPPA [53,54].The BVQ-R [52] consists of 20 statements that measure children's bullying and victimization experiences.Each subscale consists of ten items.Answers are given on a 5-point Likert-type scale where 1 means not true and 5 means absolutely true.In recent studies conducted in the same cultural context [11,47,[55][56][57], the instrument yielded two factors representing the constructs of bullying and victimization and had sufficient psychometric properties (Cronbach alphas between .70 and .80).Examples of items of the bullying subscale are the following: "other children complain that I hit them"; "I want other children to do as I say"; "other children complain that I tease them"; "other children are afraid of me." Examples of items of the victimization scale are the following: "I was threatened or forced by other children"; "other children have said lies or bad things about me"; "I was excluded/ignored by other children"; "I am being ridiculed by other children at school." The Inventory of Parent and Peer Attachment-IPPA [53,54] is a 75-item self-report designed to measure the quality of children's attachment to their mother, father, and peers.Each of the three parts of the instrument consists of the 25 Likert-type statements that assess how children evaluate their relationship with their mother, their father, and their peers on the following subscales: mutual trust, quality of communication, and the extent of felt alienation.Examples of items of the communication subscale are the following: "My mother/father helps me talk about my difficulties"; "My mother/father understands me"; "I tell my mother/father about my problems and troubles." Examples of items of the trust subscale are the following: "My mother/father respects my feelings"; "My mother/father accepts me as I am"; "My mother/father trusts my judgment." Finally, examples of items from the alienation subscale are the following: "talking over my feelings with my mother/father makes me feel ashamed or foolish"; "I don't get much attention from my mother/father"; "I get upset easily around my mother/father".As far as the peer relationship is part of the instrument, examples of items for communication subscale are the following: "I like to get my friends' point of view on things I am concerned about" and "when we discuss things, my friends care about my point of view." Further, examples of items for trust subscale are the following: "My friends understand me" and "my friends accept me as I am." In addition, examples of items for alienation subscale are the following: "My friends don't understand what I am going through these days" and "I feel angry with my friends."

Results
Using the raw data, we computed a composite variable for each construct, which represents the total score for each case on the items that compose each factor.Table 1 shows the means and standard deviations for each construct.Cronbach's alpha was also computed for each subscale, and the results showed satisfactory levels of internal consistency (all alphas > .60,Table 1).In line with earlier studies conducted in Cyprus [11,38,56], the mean scores for the bullying and victimization subscales are close to 15.00.
The next step was to compute a hierarchical regression analysis in order to identify the variables referring to the quality of relationship with parents and peers that significantly predict bullying and victimization.With the second step of this analysis, we also aimed at investigating whether the child's gender moderates the relationship between the quality of relationship measures and bullying and victimization.
This analyses showed that bullying is significantly negatively predicted by peers trust  = −.25,  < .01,positively predicted by father alienation  = .17, < .01,and negatively predicted by mother communication  = −.12, < .05.The total amount of variance of bullying predicted by this model was 16.4%.
Moderation analysis showed that there is a significant negative interaction between children's gender and peer trust in predicting bullying  = −.29, < .01 and a significant positive interaction between children's gender and father alienation in predicting bullying  = .18, < .01.For girls, the relationship between peer trust and bullying was stronger ( = −.46, < .01 for girls and  = −.23, < .01 for boys).The same was also true for the relationship between father alienation and bullying ( = .32, < .01 for girls and  = .26, < .01 for boys).
The same significant negative interaction was also found between children's gender and peer trust in predicting victimization  = −.42, < .01 and a positive interaction between children's gender and father alienation in predicting victimization  = .20, < .01.Again, the relationship between peer trust and victimization is stronger for girls ( = −.37, < .01 for girls and  = −.58, < .01 for boys), while the same was true for the relationship between father alienation and victimization ( = .35, < .01 for girls and  = .23, < .01 for boys).A significant interaction also emerged between gender and peer alienation in predicting victimization  = −.48, < .01.In this interaction, however, the prediction was stronger for boys with peer alienation rather than for girls ( = .35, < .01 for boys and  = .17, < .01 for girls) (Table 2).
Finally, we attempted to investigate whether children who classify in any of the four categories of bullying involvement roles (i.e., bullies, victims, bully/victims, and uninvolved) would show significantly different scores on the measures of the quality of relationship with parents and peers.
In order to classify the participants into one of the four mutually exclusive categories (bullies, victims, bully/victims, and uninvolved), we computed a new categorical variable with the following criteria.Participants with a score above the mean on the bullying subscale and below the mean on the victimization subscale were classified as bullies ( = 39).Participants with a score above the mean on the victimization subscale and below the mean on the bullying subscale were classified as victims ( = 41).Participants with a score above the mean on both the bullying and the victimization subscale were classified as bully/victims ( = 71).Finally, participants with a score below the mean on both the bullying and the victimization subscale were classified as uninvolved ( = 152).The same procedure of classification of the children was also used by earlier studies [38,47].
Based on this classification, a multivariate analysis of variance (MANOVA) was computed in order to identify whether the experience of bullying is related with different levels of quality of relationship with parents and peers.The results showed that there are significant differences of bullying experience on all parent and peers quality of relationship measures.Post hoc analysis showed that uninvolved children score significantly higher on all the positive aspects of parents and peers quality of relationship (i.e., trust and communication), while bully/victims score significantly higher on the negative aspects of parents and peers quality of relationship (i.e., alienation) (Table 3).

Discussion
The present study aimed at investigating the relationship between bullying and the quality of attachment with parents and peers.The main hypothesis of the present study was fully supported by the results: the quality of attachment with parents and peers is related to children's involvement in bullying incidents.It appears that all three agents of socialization (i.e., mothers, fathers, and peers) play a significant role in children's involvement in bullying.This is because bullying is significantly negatively predicted by peer trust, positively predicted by father alienation, and negatively predicted by mother communication.Accordingly, victimization is significantly negatively predicted by peer trust and positively predicted by father and peers alienation.Interestingly, the role of the father in victimization seems to be stronger than that of the role of the mother.However, if we view this result in the general parenting framework, our finding is in line with earlier studies regarding the relationship of the quality of attachment with parents and children's adjustment problems [34,58].Further evidence that support our findings is found in both earlier and recent studies.Bowlby [59,60] argued that children who fail to bond or connect with their caregivers with secure attachment tend to mistrust other people and have low levels of empathy and concern for others.As a result, these children may display a broad range of adjustment difficulties later in life.More recent empirical studies provided further support to Bowlby's claims showing that early insecure attachment relationships are associated with the lack of social skills, externalizing behaviour problems, severe antisocial behaviour, childhood aggression, conduct disorder, and delinquency [61][62][63][64][65][66][67].In sum, our results seem to be in line with these studies since we show that poor quality of attachment with all three significant socialization figures is correlated with bullying and victimization.
The second hypothesis is partially supported by the current findings.Gender differences may be important in shaping the attachment-bullying linkage.Although many studies show that parent attachment is stronger in girls, girls may also be more likely than boys to draw support from peers (e.g., [68]).This claim is based on the fact that girls may be more active in the pursuit of relatedness in the context of their peer relations than boys, who are emphasizing independence [68].In this line of research, other studies show that girls are more influenced by peer relations than boys [69,70].The results of the present study showed that girls with higher levels of peer trust were less likely to be involved in bullying and victimization.These results seem to be related with earlier claims about the significance of peer relationships in girls' socialization development [68].
In regards to the mother or father attachment, many authors argue that attachment to mothers is higher throughout childhood and adolescence [71,72].However, father attachment may show differential importance in various contexts.For instance, a study found that father attachment was more significantly related to social competence in various social situations [73].Interestingly, the current findings showed that girls involved in bullying and victimization were experiencing more alienation to their fathers.These findings indicate that the relationship with peers and fathers may function as a protective factor more for girls and less for boys.For boys, however, victimization was stronger in relation to peer alienation, which supports the idea of the importance of peer acceptance and group identification in boy's preadolescent and adolescent stages [74].
Finally, the third hypothesis is fully supported from our findings.Our study showed that children who are involved in bullying either as bullies or as bully/victims showed higher scores on negative aspects of quality of attachment with their parents and their peers, while uninvolved children showed higher scores on positive aspects of quality of attachment.Specifically, we found that regardless of gender, children who are involved in bullying experiences have worse communication and trust with their parents and peers, while there are more alienated by both parents and peers.In general, the findings of the present study are in line with earlier studies that showed bullying and victimization to be related with negative aspects of their relationship with their parents and peers (e.g., [11,34]).
The present study provides empirical evidence that supports the idea that the quality of attachment between children and their parents and peers is important in the development of adjustment difficulties such as bullying and victimization.It is also important to note that the child's gender may act as moderator in the relationship between attachment and bullying: while girls are affected more by paternal alienation and peer trust, boys are more influenced by the alienation with their peers.The current findings may be useful in considering interventions for bullying behaviour that are related with gender.The finding that girls are affected more by paternal alienation and peer trust contributes to the bullying literature by unfolding the conditions in which quality of attachment with parents and peers is more likely to correlate with children's involvement in bullying.Prevention and intervention programs could be designed and implemented distinctly for boys and girls.For instance, intervention programs for girls should focus on father and peer quality of relationship while for boys should focus on the role of peer attachment.

Limitations and Contribution of the Present Study.
The present study, in line with standard methodology in this area of research, utilized a number of self-report measures in order to assess the quality of the relationship between the child and their parents and peers and also to assess if each child was involved in bullying incidents.Even though such methodology is widely used, we recognize the limitations of selfreport measures.Future studies should employ multisource reporting that includes reports from parents and peers.
Despite these limitations, however, the present study contributes to the relevant literature in three aspects.First, we showed that the quality of attachment with parents and peers is significantly correlated with bullying and victimization.The current finding provides new insight about the relationship between bullying, victimization, and the quality of attachment with parents and peers.Communication, trust and alienation constructs were significantly related to bullying and victimization establishing their important role in children's bullying involvement.Therefore, earlier studies (e.g., [58]) using IPPA questionnaire have focused on the quality of attachment with parents and peers in terms on externalizing and internalizing problems.
Second, we also showed that gender moderates this relationship between bullying/victimization and the quality of the parents/peers attachment.Further, the moderation effects of gender provide new knowledge about the conditions that the quality of attachment with parents and peers related stronger to bullying involvement, when the child is a girl or a boy.The moderation role of gender was included in the analysis in order to bridge a relative literature gap that often leaves behind the conditions in which certain processes take place.
And, thirdly, we have provided evidence showing that bullies, victims, and bully/victims have significantly more difficulties in their relationships with their parents and peers when compared to uninvolved children.This finding contributes to the bullying literature by adding specific characteristics on the profiles of bullies, victims, and bully/victims.These characteristics include low levels of communication and trust with parents and peers and high levels of alienation.
In conclusion, our study provides further evidence that shows the link between bullying and the quality of attachment with parents and peers.Both bullying and victimization are associated with low levels of parents/peers quality.Perhaps, this is an important area of prevention and intervention that teachers, parents, and school psychologists can work on; programs that will aim to develop a healthy quality of attachment with parents and peers.
The participants of this study were 303 fifth and sixth grade children with a mean age of 11.06 years (SD = 1.03) during the 2011-2012 academic year.The two genders were equally represented in the sample, which included 146 males (48.2%) and 157 females (51.8%).The participants were randomly selected from eleven different schools in urban and rural areas of four educational districts in Cyprus (Nicosia, Larnaca, Limassol, and Famagusta).Each school has about 200 students. 2.1.Participants.

Table 2 :
Hierarchical regression analysis predicting bullying and victimization from parents' and peers' quality of relationship.

Table 3 :
Multiple comparisons (LSD post hoc analysis) with means of each category on the dependent variables.