UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The social facilitation of bullying : a multilevel analysis Shumka, Ellen 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2012_fall_shumka_ellen.pdf [ 774.24kB ]
JSON: 24-1.0073347.json
JSON-LD: 24-1.0073347-ld.json
RDF/XML (Pretty): 24-1.0073347-rdf.xml
RDF/JSON: 24-1.0073347-rdf.json
Turtle: 24-1.0073347-turtle.txt
N-Triples: 24-1.0073347-rdf-ntriples.txt
Original Record: 24-1.0073347-source.json
Full Text

Full Text

THE SOCIAL FACILITATION OF BULLYING: A MULTILEVEL ANALYSIS  by Ellen Shumka  B.A., The University of Waterloo, 2007  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (School Psychology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  October 2012  © Ellen Shumka, 2012  Abstract It is well documented that bullying is harmful and relatively common among children and adolescents. Children report understanding that bullying is wrong, yet bullying continues to be a persistent problem in schools. The goal of the present study was to examine whether children’s bullying behaviours were socially facilitated by group norms and beliefs. Children’s justifications and rationalizations for engaging in wrongful behaviour, a phenomenon referred to as moral disengagement (MD), have been linked to bullying behaviour at the individual level. Specifically, children who report engaging in bulling tend to report more MD than those who do not. Only one study to date, however, has examined MD at the group level and results indicated that group levels of MD, over and above individual levels, predicted engagement in bullying. Group level processes, especially group norms supporting aggression and bullying, have also been linked to greater bullying perpetration. The current investigation extended this research by examining how group levels of MD and normative beliefs about deviancy influenced bullying using two unique samples of schoolchildren. The first study examined the influence of two group level variables (MD and normative beliefs about deviancy) on bullying over a school year in a sample of 376 students (surveyed in Grade 5 and 6) from 38 schools in Southern Ontario. The second study examined the influence of group MD on 1128 students across 74 classrooms in Vancouver, British Columbia. Results across both studies did not support the hypothesis that group levels of MD (Study 1 & 2) and normative beliefs about deviancy (Study 1 only) influenced engagement in bullying. The findings suggest that further investigations are required in order to better understand the effects group level MD and normative beliefs on bullying behaviour.  ii  Preface The current study is based on research conducted by Dr. Tracy Vaillancourt and the MacMaster community-university research alliance (MAC-CURA). Approval for the secondary use of these data was obtained from the University of British Columbia’s Behavioural Research Ethics Board (certificate of approval # H12-02435).  iii  Table of Contents Abstract.......................................................................................................................................... ii	
   Preface........................................................................................................................................... iii	
   Table of Contents ......................................................................................................................... iv	
   Chapter 1: Introduction .............................................................................................................. 6	
   Overview............................................................................................................................ 6	
   Problem Statement ............................................................................................................. 7	
   Research Question ............................................................................................................. 8	
   Chapter 2: Literature Review..................................................................................................... 9	
   Overview............................................................................................................................ 9	
   Forms of Bullying .............................................................................................................. 9	
   Psychosocial Correlates of Bullying.................................................................................. 9	
   Social Cognitive Theory .................................................................................................. 11	
   Moral Disengagement............................................................................................... 13	
   Group Socialization Theory............................................................................................. 18	
   Chapter 3: Overview of the Current Investigation................................................................. 25	
   Study 1 ............................................................................................................................. 25	
   Study 2 ............................................................................................................................. 26	
   Chapter 4: Study 1 ..................................................................................................................... 27	
   Method ............................................................................................................................. 27	
   Procedure ......................................................................................................................... 27	
   Participants....................................................................................................................... 28	
   Measures .......................................................................................................................... 29	
    iv  4.5	
   Results.............................................................................................................................. 32	
   Preliminary Analysis................................................................................................. 32	
   Main Analysis and Rationale .................................................................................... 35	
   Assumption Testing ........................................................................................... 36	
   Multilevel Analysis Results ............................................................................... 37	
   Multiple Regression Results .............................................................................. 38	
   Chapter 5: Study 2 ..................................................................................................................... 45	
   Method ............................................................................................................................. 45	
   Participants....................................................................................................................... 45	
   Measures .......................................................................................................................... 46	
   Preliminary Analyses & Assumption Testing.................................................................. 47	
   Results.............................................................................................................................. 48	
   Chapter 6: Discussion ................................................................................................................ 49	
   References.................................................................................................................................... 54	
   Appendix A .................................................................................................................................. 68	
   Appendix B .................................................................................................................................. 69	
   Appendix C .................................................................................................................................. 71	
   Appendix D .................................................................................................................................. 72	
   Appendix E .................................................................................................................................. 73	
    v  Chapter 1: Introduction 1.1  Overview The phenomenon of bullying has been documented for centuries (see Koo, 2007, for a  review). However, only in more recent history has bullying been studied empirically. The results of several studies have established that bullying involvement is associated with serious psychological and social difficulties (Bosworth, Espelage, & Simon, 1999; Card, Isaacs, & Hodges, 2007; Gibb, Horwood, & Fergusson, 2011; Hawker & Boulton, 2000; Swearer, Song, Cary, Eagle, & Mickelson, 2001; Ttofi, Farrington, Lösel, & Loeber, 2011) and thus should not be dismissed as a “normal” part of childhood. The present study explored various social and cognitive factors that perpetuate bullying behaviour, considering the influence of both individual and group processes on behaviour over a school year. Bullying is considered a specific subtype of aggression, where a person intentionally harms another individual, repeatedly, in a relationship that has a power imbalance, either physically or psychologically (Olweus, 1994). Although research in the field of bullying has increased substantially in the past few decades (Hymel & Swearer, 2008), the mechanisms underlying bullying behaviour are not clearly understood. The present study examined one mechanism related to bullying: the moral cognitions of students who bully. In his theory of moral agency, Bandura (1986) first posited the social cognitive mechanisms through which good people become capable of committing horrific and wrongful deeds. The theory helps to explain how everyday people, seemingly moral and good, can behave immorally. Bandura’s concept of moral disengagement (MD, discussed in detail below) describes a process by which people justify their actions when a discrepancy exists between their values and  6  behaviour. The concept of MD is relevant to bullying because some people may consider such behaviour to be immoral (Caravita & Gini, 2012). As will be shown in the review of literature to follow, current research has demonstrated a clear and consistent relationship between MD and bullying (Almeida, Correia, & Marinho, 2010; Gini, 2006; Hymel, Rocke-Henderson & Bonanno, 2005; Menesini et al., 2003). Children who reported engaging in bullying demonstrated greater MD than children who do not bully, or who reported being victimized. Moreover, research has shown that children with high levels of MD become increasingly aggressive over time (Barchia & Bussey, 2010). It has also been found that children who bully tend to associate with other bullies (Salmivalli, Lagerspetz, Björkqvist, Österman, & Kaukiainen, 1996), who likely also display higher levels of MD. Group attitudes and normative beliefs about bullying have been shown to contribute to negative peer behaviour (e.g., Guerra, Williams, & Sadek, 2011). The association of like-minded peers, who consider bullying to be a common occurrence and perhaps acceptable, may facilitate MD and perpetuate bullying. For example, if a child is exposed to peers who endorse phrases such as, “Some kids get bullied because they deserve it, ” and believes bullying to be common among his or her group of friends, s/he may be more likely justify his/her wrongful behaviour in a similar manner. Accordingly, the current study explored how group level beliefs, referred to as “group MD” and “normative beliefs about deviancy”, impact student reports of bullying behaviour over a school year. 1.2  Problem Statement Morally disengaging from harmful behaviour can lead people to behave wrongfully  without feeling guilty. It is important to study moral and immoral behaviour in children, so we can learn how best to intervene in bullying behaviour at a young age. Research on school  7  bullying and MD is a bourgeoning area of inquiry (e.g., Gini, Pozzoli, & Hauser, 2011; Obermann, 2011; Perren, Gutzwiller-Helfenfinger, Malti, & Hymel, 2011; Pornari & Wood, 2010). However, the majority of research on bullying and MD to date has focused on individual rather than group level beliefs and norms. Only one study to date, conducted by Vaillancourt and colleagues (2006), has explored the impact of group level MD on bullying behaviour. In this study, higher levels of MD, at both the group and individual level, predicted more engagement in bullying. The current study extended Vaillancourt et al.’s findings by exploring the impact of both individual and group level MD, as well as group normative beliefs about deviancy, on reported levels of bullying among elementary and middle school age children over a school year. 1.3  Research Question The present investigation addresses the question of whether group (i.e., school, classroom)  levels of MD and normative beliefs about deviancy predict bullying behaviour over and above individual levels of MD and normative beliefs about deviancy.  8  Chapter 2: Literature Review 2.1  Overview This chapter first outlines the various forms of bullying, as well as the psychosocial  correlates of bullying involvement. Next, Bandura’s (1986) social cognitive theory of MD is discussed, as well as its relationship to bullying behaviour. Finally, the social aspects of bullying are considered, including an overview of group socialization theory (Harris, 1995) and the role of normative beliefs in bullying behaviour among students. 2.2  Forms of Bullying Bullying can take a variety of forms. Traditionally, bullying has been described as overt,  physically harmful behaviour perpetrated primarily by boys (Olweus, 1978). However, over the past several decades researchers have distinguished various forms of bullying. Physical (e.g., hitting, kicking and pushing) and verbal bullying (e.g., threats, insults and ridicule) are both considered direct forms of aggression. Indirect, social or relational forms of bullying (Crick & Grotpeter, 1995; Underwood, Scott, Galperin, Bjornstad, & Sexton, 2004) include behaviours aimed at harming a person’s social status or relationships (e.g., exclusion, gossiping, rumor spreading). Lastly, electronic or cyberbullying (e.g., slanderous texts, posting embarrassing pictures) has emerged alongside recent technological advances (Raskauskas, 2007; Slonje & Smith, 2008). In the current study, a composite index of bullying was comprised based on selfreports of all four forms. 2.3  Psychosocial Correlates of Bullying The need to belong is a fundamental motivator for people of all ages to engage in social  interaction and develop relationships with peers (Baumeister & Leary, 1995). It is advantageous for individuals to make good impressions on potential social companions because positive social  9  relationships are essential for mental and physical well-being (Heinrich & Gullone, 2006). In contrast, antisocial behaviours, such as bullying, are associated with detrimental physical and mental health difficulties for all of the individuals involved (e.g., Bosworth et al., 1999; Gibb et al., 2011; Hawker & Boulton, 2000; Swearer et al., 2001). Victims of bullying are at increased risk for psychological distress. Victims are more likely to be depressed (Biggam & Power, 1999; Craig, 1998; Dao et al., 2006; Fekkes, Pijpers, Fredriks, Vogels, & Verloove-Vanhorick, 2006; Hawker & Boulton, 2000; Nansel et al., 2001), report higher levels of anxiety (Biggam & Power, 1999; Dao et al., 2006; Fekkes et al., 2006; Hodges & Perry, 1996), and experience more physical health problems (Greco, Freeman, & Dufton, 2007), compared to those not involved in bullying. Additionally, victims of bullying are at increased risk for post-traumatic stress disorder (PTSD) as well as substance abuse or dependence (Rivers, 2004). Victims are not the only ones at risk, however. While some studies have shown that that children who bully others reported few psychological difficulties (e.g., Juvonen, Graham, & Schuster, 2003), the majority of research findings support the view that engaging in bullying behaviour is associated with psychosocial distress (Gibb et al., 2011). Some of these difficulties include depression (Bosworth et al., 1999; Kaltiala-Heino, Rimpela, Marttunen, Rimpela, & Rantanen, 1999; Slee, 1995; Ttofi, Farrington, Lösel, & Loeber, 2011), anxiety (Duncan, 1999; Swearer, et al., 2001), as well as higher levels of anger and conduct problems (Bosworth et al., 1999; Griffin & Gross, 2004; Loeber & Hay, 1997; Olweus, 1994; Schaeffer, Petras, Ialongo, Poduska, & Kellam, 2003). Bullying occurs in many forms and is problematic for individuals involved. It is also common. Approximately 30% of Grade 6 to 10 students in a large U.S. sample reported “moderate” to “frequent” levels of involvement as either bullies, victims, or both (Nansel et al.,  10  2001). These rates are conservative in comparison to results published by the World Heath Organization (WHO), where 44.6% of Canadian 13-year-olds reported engaging in bullying at least “sometimes” (Krug, Dahlberg, Mercy, Zwi, & Lozano, 2002). These figures rank Canada as 11th in overall reported levels of bullying across 27 countries (Krug et al., 2002). Consistent with the WHO study results, Vaillancourt and colleagues (2010) reported that 48.9% of a large sample of Canadian students (Grades 4 to 12) reported some level of involvement in bullying perpetration. Bullying is a common problem affecting children across the world, with no exceptions in Canada. In summary, bullying occurs in a variety of forms and involvement in bullying is associated with numerous psychosocial difficulties. We know that bullying is common among schoolchildren, yet we know very little about the processes underlying such disruptive behaviour. In the next section, Bandura’s theory of moral disengagement (MD) is outlined, which helps to explain how individuals think, act, and cope with harmful behaviour, such as bullying. 2.4  Social Cognitive Theory Socal cognitive theory, as first conceived by Albert Bandura in his 1986 publication,  Social Foundations of Thought and Action: A Social Cognitive Theory, is based upon social learning theory, which is based on the premise that human behaviour is learned by observing other people and through the rewards and consequences experienced. The premise that social factors influence behaviour is expanded by social cognitive theory to include personal components, such as individual beliefs, motivations, and affect. Social cognitive theory is interactionist, meaning it assumes personal, behavioural, and environmental factors influence each other bidirectionally (Bandura, 1986). This triadic relationship, as applied to moral  11  reasoning, implies that both moral thought, including self-sanctions1, and moral environment, including social sanctions, influence whether one acts morally or immorally. Social cognitive theory also considers personal factors, such as one’s thoughts, biology, feelings, and intentions, as major determinants of immoral or harmful behaviour (Bandura, 1986). Within the domain of personal factors, self-regulation is one system that influences whether one acts morally or not. Self-regulation is an internal control mechanism that manages thoughts, feelings, motivations and actions (Bandura, 1986). Self-regulation operates through three subfunctions: self-monitoring, judgment of conduct, and affective self-reaction (Bandura, 1986). These subfunctions operate together so that people can: (a) recognize their thoughts and motivations to behave in a particular way, (b) judge their intended behaviour in relation to those around them or to their own personal moral code of conduct and, (c) choose whether or not to execute a behaviour, depending on how that makes them feel (Bandura, 1991). According to social cognitive theory, personal factors contributing to immoral behaviour cannot be dissociated from social influence. Bandura (1986) argued that social cues delivered through modeling, persuasion, and instruction influence personal thought and action. Further, depending on one’s age, size, gender, and ethnicity, personal factors may elicit different responses from the social environment. For example, if a young child observes an older, popular child call another child a name while bystanders laugh, he or she may be more likely to think about hurting someone similarly, and eventually do it. The young child may be especially susceptible to social influences if he or she does not have strong personal beliefs or standards that are contrary to social cues being delivered (Bandura, 1991).  1  The term “sanctions” refers to punishment or penalty for violating a moral principle. 12  The concept that moral thoughts and environment influence moral behaviour can be viewed as an extension of traditional theories of moral development (Piaget, 1932/65; Kohlberg, 1969 as cited in Turiel, 2006). Specifically, social cognitive theory maintains that, although some children in middle childhood may have attained “conventional” levels of moral reasoning, they behave in ways that are inconsistent with their moral development. Thinking and acting in ways that are consistent with one’s beliefs or moral standards are more likely to make someone feel good, whereas acting in ways that are contrary to one’s beliefs and standards would cause unease or distress. However, people often act in ways that are inconsistent with their moral standards. Moral disengagement (MD) is a process by which individuals disengage from moral standards, as described in the following section. 2.4.1  Moral Disengagement Bandura (1999, 2001) posits that self-sanctions can be disengaged from immoral conduct  using a variety of mechanisms that provide insight into how people rationalize or justify wrongful behaviour. Specifically, Bandura (1999) distinguishes four MD mechanisms: cognitive restructuring, minimizing one’s agentive role, disregarding or distorting negative consequences, and blaming or dehumanizing the victim. Each of the four MD mechanisms involve eight unique strategies through which individuals can morally disengage from self-censure, as described below. Cognitive restructuring is a mechanism through which individuals may relieve themselves from sanctions associated with culpable behaviour. Three distinct strategies for cognitive restructuring have been identified. Moral justification is a strategy through which one reconstructs immoral behaviour into something that serves a moral or positive purpose. They need not sanction such behaviour—they may even feel proud for committing what they now  13  consider to be a moral or positive act. Within the context of schools, bullies are more likely to justify their behaviour by believing that “bullying is just a normal part of being a kid” or “sometimes it’s okay to bully other people” (Hymel, et al., 2005). Euphemistic labeling, or how one refers to victims can also shape their thought patterns and subsequent actions (Bandura, 1991). Referring to bullying as “rough and tumble play” or thinking about bullying as “just kidding around” are examples of euphemistic labeling. Bandura (1991) refers to cognitive restructuring through justification and euphemistic labeling as the most effective disengagement mechanisms because they not only allow the individual to eliminate the need for self-sanctions, but allow him/her to distort and/or minimize the impact of such behaviour. Advantageous comparison is a third cognitive restructuring strategy through which the individual compares his or her immoral act to more flagrant inhumanities, allowing the perpetrator to view his/her behavior as relatively less harmful. A child who bullies using verbal or indirect tactics may contend, “At least I didn’t hit him. Sam hits people all of the time and I would never do that”. Another MD mechanism involves minimizing one’s responsibility for the act (Bandura, 1991). Displacement of responsibility is a strategy employed to eliminate personal responsibility by displacing it onto an authority figure, thereby removing self-sanctions by placing the responsibility onto another person. For this strategy to be effective, the authority figure must be legitimate and must appear coercive, and the person using the method must assume responsibility for being diligent. For example, children may displace their responsibility for intervening in bullying instances by thinking, “adults at school should be responsible for protecting kids from bullies” (Hymel et al., 2005). Diffusion of responsibility involves the weakening of selfsanctions by placing blame across many individuals. For example, statements like, “nobody else stepped in to stop it” or “the teacher saw it and she didn’t do anything either” is consistent with  14  diffusing the responsibility to intervene in bullying. This specific form of disengagement is based on Bandura, Underwood, and Fromson’s (1975) research revealing that people act more harshly when responsibility for the action is collective, as compared to when they feel personally responsible. Disregarding or distorting consequences is a mechanism that involves minimizing harmful effects of the behaviour and emphasizing the benefits. This can be achieved by discrediting the evidence that any harm was caused. For example, children might view the bullying as a way to “make people tougher” (Hymel et al., 2005). During a focus group conducted by Guerra et al. (2011) one adolescent remarked, “I think if you are having fun it is not bullying, but if you are actually mean it is bullying if it hurts another person”. This type of thinking opens the door for teasing someone “in good fun” and disregarding the harm it may cause the victim. The final two MD strategies involve blaming and dehumanizing the victim, sometimes by ascribing subhuman characteristics to victims, in an effort to blunt empathizing with the person’s feelings, hopes or concerns. For example, thoughts like “some kids get bullied because they deserve it” or “most students who get bullied bring it on themselves” are clear examples of blaming the victim, whereas statements like, “kids who get picked on are losers” reflect efforts to dehumanize the victim. The MD mechanisms and strategies that children and youth use to justify bullying behaviour have been assessed using self-report measures in which students are asked to endorse various statements reflecting different ways of MD, as in Bandura’s (1999; 2002) original MD scale (e.g., Almeida et al., 2010; Gini, 2006; Hymel et al., 2005; Pornari & Wood, 2010). In addition, MD has been assessed using self-reports of moral emotions, with emotions of pride,  15  excitement or indifference reflecting MD and emotions of shame or guilt reflecting moral responsibility or engagement (e.g., Menesini et al., 2003). Across studies, bullying behaviour has been consistently linked to greater MD (Barchia & Bussey, 2011; Gini, 2006; Hymel et al., 2005; Menesini et al., 2003; Obermann, 2011). For example, in a sample of 494 junior secondary students (Grades 8 to 10), Hymel et al. (2005) reported that levels of MD accounted for 38% of the variance in self-reported bullying. Students who reported engaging in bullying reported the lowest levels of MD, and students who admitted that they repeatedly bullied others reported the highest levels of MD. Further, in a study of predominantly Caucasian secondary students in Australia (Grades 7 to 10), Barchia and Bussey (2011) found that students who rated themselves as high on MD also reported more aggressive behaviour over time than children with low MD scores, even after accounting for the stability of aggressive behaviour. These findings are supported by more recent research conducted with a sample of Danish sixth and seventh graders, where both peer and self-reports of bullying were significantly associated with MD (Obermann, 2011). In this study, students who were viewed as bullies (by their peers, themselves, or both) reported significantly higher levels of MD than peers who were not considered bullies (by their peer or themselves). Gini (2006) further explored the relationship between MD and bullying by examining levels of MD across elementary schoolchildren (ages 8 to 11) identified by peers as taking on different roles in bullying incidents. Bullies and victims, bystanders, uninvolved students, as well as students who reinforced or assisted in bullying or who defended those who were victimized were compared. Results indicated that children who bully reported higher levels of MD compared to children who were uninvolved or victims. Further, children who bullied others, and  16  those who reinforced or assisted the bully reported higher levels of MD compared to bystanders who defended the victim. Menesini et al. (2003) investigated the degree to which peer-nominated bullies reported feelings of moral responsibility (guilt, shame) compared to feelings of MD (indifference, pride) in response to hypothetical bullying behaviour. Children who were identified by their peers as bullies were more likely to ascribe feelings of pride (in 2/3 of sample) or indifference (in 1/3 of sample) to a hypothetical bully compared to victims or uninvolved students. Clearly, children who bully are more likely to morally disengage. Looking beyond the individual, the current research explored how group level or contextual factors influenced bullying behaviour over a school year. Bandura (1991) proposed that when various strategies of MD are combined, the effects increase exponentially. Within the framework of social cognitive and social learning theory, the use of various MD strategies by other individuals in the environment should influence one’s own use of MD strategies. Accordingly, one would expect that children who are involved with peers who are more morally disengaged would be more likely to think and act in similar ways. Only one investigation to date by Vaillancourt et al. (2006) has examined the effect of group levels of MD on bullying. In a sample of 16,879 predominantly Caucasian (65%) students in Grades 4-12, from 116 schools, the authors reported that school level MD predicted bullying over and above individual levels of MD. The current study expands on Vaillancourt and colleagues’ investigation by examining the effects of group levels of MD and group normative beliefs about deviant behaviour, over and above individual levels of MD and beliefs about deviancy, on bullying behaviour over a school  17  year. Group socialization theory (described below) guides the hypothesis that peer groups, as well as individual levels of MD influence differences in bullying behaviour. 2.5  Group Socialization Theory Group Socialization Theory (Harris, 1995/2009; 1998) outlines the processes through  which individuals are socialized by groups, with particular attention to the influence of peers in the development of personality. Peer processes and influences may be particularly important considerations in research on bullying, as researchers have long recognized that bullying is a group phenomenon (e.g., Olweus, 1993). Indeed, observational research has documented that peers are present in the vast majority of bullying episodes (Craig & Pepler, 1997), and recent research by Salmivalli, Peets, and Hodges (2010), has shown that peers play a number of critical roles in bullying episodes. Although most research in the area of bullying to date has focused on individual characteristics of children who bully, more recent attention has been placed on the influence of the social context in which bullying takes place. Harris (1998/2009) reviews a large body of social psychological research that documents several processes that operate once individuals are placed into groups. Her Group Socialization Theory suggests that it is through such processes that individual behavior is modified by the group. One of the group processes suggested by Harris that is particularly relevant to the present study is the process of “within group assimilation”. According to Group Socialization Theory, once individuals are in a group, they tend to favour their own group and discriminate against other groups. Over time, the differences between the groups are overemphasized and similarities are underemphasized. At the same time, people who categorize themselves as belonging to the same group tend to become more similar over time, a process that Harris refers to as “within group assimilation”. Once an individual categorizes him/herself as a member of a group, they  18  tend to ascribe to the attitudes, beliefs, rules, and behaviours of that group (Turner, 1987, as cited in Harris, 1998/2009). In fact, there is evidence to demonstrate that the more one identifies with the group, the more he/she is likely to adhere to the norms of the group (Jetten, Postmes, & McAuliffe, 2002; Jetten, Spears, & Manstead, 1996, 1997). Thus, over time, a person is more likely to think, feel, and behave in ways consistent with the group prototype, or norm. Although Harris recognizes that individuals also differentiate within groups, through the development of status hierarchies and the use of social comparison, within-group assimilation nevertheless continues, creating greater similarity among group members over time. The current study focused on whether and how children adopt the beliefs and behaviours of their school peer group, specifically with regard to MD, normative beliefs about deviancy, and bullying. Consistent with the “within group assimilation” processes described by Group Socialization Theory (Harris, 1998/2009), research on bullying has shown that children prefer their own group (Gini, 2006), and also that children who are friends share similar levels of aggressive and antisocial behaviour compared to nonfriends (Hartup, 2005; Haselager, Hartup, Van Lieshot, & Risken-Walraven, 1998). The idea that “birds of a feather flock together” begs the question of whether aggressive children simply select other aggressive children as friends, or whether belonging to an aggressive group actually causes children to become more aggressive. Group Socialization Theory would posit that, once children self-categorize as members of a proaggression group, they would be more likely to conform to the aggressive thoughts, beliefs, and behaviours of that group. Several studies have demonstrated that children who spend more time with other antisocial children show increased antisocial behaviour (Dishion, 1990; Tremblay, Masse, Vitaro, & Dobkin, 1995). Thus belonging to a particular group can affect future  19  behaviour, particularly aggressive behavior. Research on the influence of group norms on individual behavior further supports these arguments. The power of peer group norms has been of interest for decades, with seminal research conducted by Soloman Asch and Musafer Sherif in the 1950s. These researchers demonstrated that when individuals form a group (i.e., eight men sitting in a room, or boys attending summer camp, respectively), certain expectations are formed regarding how group members behave, and group members do tend to conform to those expectations (Asch, 1955; Sherif, Harvey, White, Hood, & Sherif, 1961, as cited in Harris, 1998). More recently, a growing body of research has emerged that supports the role of group norms in aggressive behaviour, with numerous studies demonstrating a link between normative beliefs about aggression and aggressive behaviour (Espelage et al., 2003; Henry et al., 2000; Huesmann & Guerra, 1997). For example, children’s level of aggression has been found to increase when they are in a classroom that views aggressive behaviour as normative (Henry et al., 2000). With regard to bullying behaviour specifically, classroom norms have been found to play a significant role in predicting student behaviour in bullying situations (Salmivalli & Voeten, 2004). Anti-bullying classroom norms were significantly and negatively related to bullying and behaviour that supported or reinforced bullying behaviour (e.g., laughing, watching) in Grade 5 and 6 students. Specifically, low antibullying normative beliefs in the classroom (i.e., bullying is acceptable) were associated with higher reports of bullying and greater reinforcement of bullying. Espelage and colleagues (2003) examined the contextual effects of the peer group on bullying over time using multilevel analyses. Specifically, they explored how bullying and fighting were associated with peer group membership (i.e., “within-group similarity”) in a sample of male (N = 155) and female (N = 171) American Grade 6, 7 and 8 students. The authors  20  reported that the contextual effects explained more variance in bullying than fighting behaviour, and suggested that future research on contextual influences focuses on specific subtypes of aggression. Results of studies exploring contextual effects on bullying per se (not fighting) are outlined next. First, Espelage et al. (2003) used a social network analysis examining peer nominations of friendship to indentify students’ peer group affiliations, or “cliques”. The cliques were then used to group students in subsequent multilevel analyses. Results revealed significant peer-group homogeneity with regards to self-reported bullying, supporting their hypothesis that students affiliated with other students who reported levels of bullying similar to their own. The differences in bullying across peer groups were also examined, and results indicated that the average amount of bullying reported within a student’s peer group in the fall (Time 1) predicted student levels of bullying behaviour in the spring (Time 2), even after controlling for previous bullying involvement (Time 1). The authors also reported that a significant amount of variation in student levels of bullying at Time 2 remained, which could be accounted for by additional predictors. The results from Espelage et al.’s (2003) study support the hypothesis that peer groups influence individual levels of bullying behaviour. However, the results indicated that peer group behaviour (i.e.,bullying) and previous bullying behaviour predicted self-reports of bullying involvement at Time 2, but did not account for all of the variation. The current study extends Espelage et al.’s findings by examining the influence of group level attitudes and beliefs on individual bullying behaviour, after controlling for previous levels of bullying. In the present study, the two group level processes of interest are “group MD” and “normative beliefs about  21  deviancy”. One study to date has examined the effects of group normative beliefs on bullying using both multilevel and qualitative analyses (Guerra et al., 2011), which is described next. Guerra and colleagues (2011) explored individual and contextual predictors in bullying and victimization using a mixed-methods approach. Specifically, the study examined levels of (a) self-esteem, (b) normative beliefs supporting bullying, and (c) school climate in a sample of 2678 elementary (Grade 5), middle (Grade 8) and high school (Grade 11) students from America over a three-year period. Both quantitative and qualitative analyses were conducted. A multilevel analysis examined changes in bullying from Time 1 to Time 2 using student self-reports, and a qualitative analysis examined student experiences with bullying using semi-structured interview data derived from focus groups conducted at Time 1. Results from the multilevel analysis indicated that the strongest predictor of individual bullying was normative beliefs supporting bullying behaviour, although normative beliefs did not predict levels of victimization (Guerra et al., 2011). The results from the qualitative analysis revealed several themes for why students bully, which included the concepts of normative beliefs (e.g., “everybody bullies”), and bullying as entertainment (“bullying is fun”). Quantitative and qualitative results were consistent; a clear link exists between normative beliefs about bullying and changes in bullying behaviour. In addition to underscoring the importance of studying normative beliefs and bullying behaviour, the Guerra et al. (2011) study provided further support for exploring the role of MD in bullying. The responses from children in the focus group revealed a “bullying as entertainment” theme, which encompassed the students’ ideas that bullying could be fun. Conceptualizing harmful behaviours, such as bullying, as a fun activity is an example of the MD mechanism of disregarding or distorting the consequences. Together, the results from Guerra et  22  al.’s (2011) study provide support for further exploring the role of MD and normative beliefs about bullying on student levels of bullying. A gap currently exists in the bullying literature. Research demonstrates that bullying is a social phenomenon, yet there is an insufficient understanding of the influence of MD at the group level on bullying. Although studies have demonstrated links between bullying and MD at the individual level, far less is known about the group level MD processes. Vaillancourt and colleagues (2006) have examined the effects of group MD using a cross-sectional multilevel design and reported that group level MD, over and above individual levels of MD, significantly predicted individual levels of self-reported bullying. Group socialization theory provides a framework for understanding the influence of peers on bullying behaviour and MD. In particular, Group Socialization Theory would suggest that individuals are more likely to bully when they categorize themselves as members of a group who normalizes bullying behaviour and/or engage in more MD. Consistent with these hypotheses, Vaillancourt et al. (2006) reported that higher group levels of MD were associated with more reported bullying, over and above individual levels of MD. Further, two studies have explored the contextual effects that influence bullying using multilevel modeling. Espelage et al. (2003) reported that previous levels of individual and peer-group bullying predict future bullying behaviour, although these behaviours did not explain all of the variance in bullying. Guerra et al. (2011) reported that school normative beliefs predicted bullying behaviour, yet such influences have not been examined for group level MD. In the current study, it was hypothesized that group level beliefs that serve to justify or rationalize bullying affect the degree to which children normalize such negative behaviour and in turn lead to greater bullying. Perceptions that “everyone bullies” or that “bullying frequently  23  occurs at my school” may reflect justification (MD) processes. The present study adds to the current literature by including both MD and group norms about deviancy as individual and group level predictors of bullying in order to examine their unique contribution to bullying behaviour, while controlling for previous bullying involvement.  24  Chapter 3: Overview of the Current Investigation Previous research has demonstrated links between MD and bullying at the individual (Ando, Asakura, & Simons-Morton, 2005; Gini, 2006; Hymel et al., 2005; Menesini et al., 2003) and group level (Vaillancourt et al., 2006). Results from research using multilevel analyses have indicated that bullying behaviour can be predicted by peer-group and individual levels of bullying (Espelage et al., 2003), and normative beliefs about bullying (Guerra et al., 2011). The present investigation examined the influence of group level variables on bullying behaviour in two separate studies. The research design for each study is provided next. 3.1  Study 1 Study 1 investigated school level MD school level MD and normative beliefs about  deviancy on bullying behaviour over a school year, considering the influence of group over and above individual levels of MD and normative beliefs (see Table 1). Specifically, bullying levels at Time 2 (T2) were predicted by individual and school level MD and normative beliefs about deviancy at T2, while controlling for sex of subject and previous bullying behaviour at Time 1 (T1). Given evidence that bullying most often occurs in unsupervised areas, such as the playground/school yard and hallways, rather than classrooms, which are highly supervised (Vaillancourt et al., 2010), schools, rather than classroom, were considered the most appropriate unit of analysis.  25  Table 1 Study 1 Research Design T2 Individual = T1 Individual + Sex of + T2 Individual + T2 Individual Bullying Bullying Subject Moral Normative (Outcome Variable) Disengagement Beliefs about Deviancy + T2 School + T2 School Moral Normative Disengagement Beliefs about Deviancy 3.2  Study 2 Study 2 investigated the influence of classroom level MD, over and above individual  MD, on bullying behaviour in a separate sample of schoolchildren. Classrooms, rather than schools, were analyzed due to the large number of classrooms and small number of schools in the sample. Further, normative beliefs about deviancy were not measured in the larger study from which the data were drawn, and thus, were not included in the analysis. Table 2 Study 2 Research Design Individual Bullying = Individual (Outcome Variable) Bullying  + Sex of Subject  + Individual Moral Disengagement + Classroom Moral Disengagement  26  Chapter 4: Study 1 4.1  Method Data for the current study were drawn from the first two years of the MacMaster Teen  Study (2007/8 and 2009/10), collected as a part of an ongoing research project conducted by Dr. Tracy Vaillancourt, aimed at understanding the longitudinal associations between peer victimization, mental health, and academic achievement (Campbell, Missiuna, & Vaillancourt, 2012). The original project received ethical approval at MacMaster University, and the present analysis received ethical approval for the secondary use of the data from the Behaviour Research Ethics Board from the University of British Columbia. Data for the present study was anonymous, with all identifying information removed. 4.2  Procedure The principals of 50 elementary schools with Grade 5 classrooms from one large public  school district in southern Ontario were randomly contacted to participate in the study. Schools that declined participation (N = 17) were replaced through random selection until 50 participating schools were identified. Of the 50 participating schools, 41 received in-class presentations by trained research assistants who described the study and distributed consent forms to students. In nine schools, principals requested a package containing information about the purpose of the study and consent forms and agreed to provide and discuss the information contained in the package with the Grade 5 classroom teachers and students. In the first year of the study, data collection took place in each of the 50 schools during the spring of 2008. In the second year, students were contacted through their homes to complete the questionnaire online (54%) and as well as through their schools. The majority of the surveys (67%) for the second year were returned by the students in the summer of 2009, and all surveys  27  were received by the spring of 2010. Each year, students completed a self-report questionnaire consisting of questions on bullying, mental health, and academic achievement. A subset of questions from the larger study was analyzed for the present study. Specifically, the present study examined student responses to questions about bullying involvement, MD, and normative beliefs about deviant behaviour. 4.3  Participants Students who attended elementary and middle schools in southern Ontario were invited to  participate in the first two waves of the longitudinal MacMaster Teen Study. The first wave (T1) of data collection took place in 50 schools, when students were in Grade 5. In the second wave (T2), students were in Grade 6 and were spread across 63 schools. Approximately half of the students stayed in the same school they attended in Grade 5, and the other half moved to new schools. The majority of the students identified themselves as European-Canadian (65%), with the next largest group reporting they “did not know” their ethnicity (11%). The remainder of the sample was ethnically diverse, with students identifying as “Other” (3%), African/West-IndianCanadian (3%), South-Asian-Canadian (3%), Middle-Eastern-Canadian (2%), Asian-Canadian (2%), South/Latin American-Canadian (1%), and Native-Canadian (1%). Nine percent of the students had missing data for ethnicity. Across the 50 schools participating at T1, 1922 consent forms were distributed and 1121 (58%) students received parental consent to participate in the research. Subsequently, 34 of these students declined assent. Of the 1087 students who received parent consent and who themselves agreed to participate at T1, 602 (53% female) were surveyed in both the first and second wave (2007/8 and 2009/10, respectively) of the study. In the current sample, however, the number of students who participated at T1 and T2  28  ranged from 2 to 44 students per school, with an average of 12 students at each school. Given the fact that school level variables were of primary interest in the present study, it was important to consider the number of students per school who participated in the study at each time point, as statistical analyses may be biased by small numbers of observations per group. Accordingly, students who attended schools where fewer than 5 students (in total) participated in the study in a given year were excluded. This decision was based on efforts to optimize the number of schools considered, while maintaining a sufficient number of students per school to minimize bias in estimates of the standard errors (Clarke & Wheaton, 2007; Hox, 2002; Maas & Hox, 2004). In total, 69 students were excluded from analyses because they attended a school where fewer than 5 students (in total) participated at either T1 or T2, or both. Additionally, 157 students were omitted because they had missing data for one or more of the variables of interest (T1, bullying; T2, bullying, MD, normative beliefs about deviancy). In total, the final sample consisted of 376 students (54% female), attending 38 schools. 4.4  Measures Bullying Involvement (Olweus Bully/Victim Questionnaire). The Olweus’ (1996)  Bully/Victim Questionnaire is the mostly widely-used measure of peer victimization and bullying in the world. It was used by the World Health Organization in the Health Behaviour in School-Aged Survey (HBSC 1997/98, 2001/02; see Roberts et al., 2009). On the Olweus Questionnaire, students are asked a single question about their bullying behaviour (i.e., “Have you bullied another person?”). A revised version of this measure was used to assess involvement in bullying perpetration in the present study. In addition to responding to one general bullying question, students responded to four additional questions about engagement in each of four forms of bullying: physical, verbal, social, and electronic.  29  Students first read a definition of bullying that distinguished bullying and teasing, and identified the three major features of bullying (i.e., it is ongoing, intentional, and includes a power imbalance). Students then answered each of the five items regarding their own bullying involvement on a 5-point Likert scale (0 = never, 1 = only once or twice, 2 = 2 or 3 times a month, 3 = once a week, 4 = several times a week). A composite measure, computed as the sum of all five bullying items (i.e., a general question on bullying, and questions regarding the four forms of bullying), was used to examine self-reported bullying behaviour, with higher scores indicating more engagement in bullying perpetration (see Appendix A). Internal consistency of the composite measure in the present study was good at both T1 (α = .74) and T2 (α = .71). Group level indices of MD were computed as the average of scores on each index from all participants within a given school, with an average of 12 students participating per school (range = 5 to 44 peers) contributing to these composites. Normative Beliefs about Deviancy. Beliefs legitimizing anti-social behaviours were measured using an adapted version of the Attitudinal Intolerance of Deviance scale (Jessor, Donovan, & Costa, 1989; Judd, Jessor, & Donovan, 1986). On this measure, the perceived “wrongness” of deviant or delinquent-type behaviours (e.g., “To take little things that don’t belong to you.”, “To break something that belongs to another person just to get even.”) were rated on a 4-point scale (0 = not at all wrong, 1 = a little bit wrong, 2 = wrong, 3 = very wrong) (see Appendix B). For the present study, the 11-items were adapted to reflect deviant behaviours more closely related to school activities (e.g., “To cheat on a test.”) rather than home life (e.g., “To lie to your parent about where you have been.”). A composite measure, computed as the sum of students’ responses to 11 items, was used to examine the degree to which students disapprove of behaviours that violate social norms, with higher scores indicating less tolerance  30  for deviant behaviour. The original Attitudinal Intolerance of Deviance scale was reported to have high internal consistency (α = .85 - .90; Clark, Dogan, & Akbar, 2003; Jessor, Van Den Bos, Vanderryn, Costa, & Turbin, 1995). For the present sample, the internal consistency was also high (α = .81). Group level indices of normative beliefs about deviancy were computed as the average of scores on each index from all participants within a given school, with 5 to 44 peers contributing to these composites (M = 12 students per school). Moral Disengagement Towards Bullying Behaviour. Students were asked to report on their level of MD with regards to bullying behaviour by answering five items using a 4-point scale (0 = strongly disagree, 1 = disagree, 2= agree, 3= strongly disagree, see Appendix C). The five MD items for the present study were adapted from the 18-item MD measure used by Hymel et al. (2005). The items that comprised the present MD scale best represent the MD mechanisms of cognitive restructuring (i.e., “Sometimes it’s okay to bully other people.”, “In my group of friends, bullying is okay.”), and blaming or dehumanizing the victim (i.e., “Kids get bullied because they are different.”, “Some kids get bullied because they deserve it.”, “Some kids get bullied because they hurt other kids.”). It is important to note, however, that a subsequent factor analysis conducted by Hymel et al. (2005) demonstrated that most of their original MD items (13 of 18) loaded on a single factor, rather than the four conceptual categories outlined by Bandura (1999, 2002). Two of the items included in the present study (“Kids get bullied because they are different.”, “Some kids get bullied because they hurt other kids.”) were not the items that Hymel et al. (2005) reported as significantly loading onto single factor. A composite measure, computed as the sum of students’ responses to the five items, was used to examine the degree to which students morally disengaged from bullying behaviours, with higher scores indicating more MD towards bullying. For the present sample, the internal consistency for all five items was good (α  31  = .69). Group level indices of MD were computed as the average of scores on each index from all participants within a given school (range = 5 to 44, M = 12 students per school). 4.5 4.5.1  Results Preliminary Analysis Preliminary analyses were conducted to examine the psychometric qualities of the  measures used in the present study. First considered were the composite indices of bullying, computed at both T1 and T2. As indicated by the means and standard deviations for these measures (see Table 2), on average, most students reported bullying a few times a year. Indicators of skewness revealed that responses to bullying were positively and significantly skewed at both Time 1 (z = 25.48, p < .05) and Time 2 (z = 11.35, p < .05), which means that most students endorsed items indicating that they engaged in little to no bullying. Indicators of kurtosis were also significant at T1 (z = 72.11, p < .05) and T2 (z = 7.2, p < .05). In fact, 53% of students at T1 and 55% of students at T2 reported “not bullying at all in the past year”. Despite the heavily skewed response, the present study is most interested in explaining existing variability in bullying behaviour. Thus, it was important to include bullying as a continuous, dependent variable. As indicated in the Methods section, the internal consistency for both T1 and T2 bullying measures were acceptable (Field, 2009; see Table 2). Next, the composites of MD and normative beliefs about deviance were examined for T2 at the individual and school level. For individual MD, the means and standard deviation indicated that, on average, students reported low levels of MD (see Table 2). Indicators of skewness (z =1.48, p > .05) and kurtosis (z = -.68, p > .05) revealed that responses to individual levels of MD were approximately normal. School levels of MD, on average, were also low (see  32  Table 2) and approximately normal according to indicators of skewness (z = -1.47, p >.05) and kurtosis (z = 2.98, p < .05). For normative beliefs about deviancy, the means and standard deviations indicated that, on average, students reported high levels of intolerance towards deviancy (see Table 2). Indicators of skewness (z = -.98, p > .05) revealed that responses to individual levels of normative beliefs about deviancy were approximately normal, with higher than expected levels of kurtosis (z = - 4.39, p < .05). School levels of normative beliefs about deviancy, on average, were also high (see Table 2) and non-normal according to statistically significant indicators of skewness (z = -5.03, p < .05) and kurtosis (z = 6.42, p < .05). However, Q-Q plots revealed that the skewness and kurtosis were influenced by several outlying schools with low levels of normative beliefs about deviancy. The values from these schools fell within the acceptable ranges, and were thus included in further analyses. Table 3 Psychometrics of Measures N  Mean  Standard Deviation  Internal Consistency (a)  Bullying Time 1 Time 2  376 376  1.09 1.06  1.73 1.48  .74 .71  Individual Moral Disengagement  376  4.52  2.57  .69  Individual Normative Beliefs about Deviancy  376  29.99  3.21  .81  School Moral Disengagement  38  4.60  1.02  --  School Normative Beliefs about Deviancy  38  29.83  1.31  --  Next, a series of correlational analyses were conducted to explore the zero-order relationship among individual levels of bullying, MD and normative beliefs about deviancy. The 33  correlations provided information on whether the measures of individual MD and normative beliefs about deviancy were related to bullying in the manner expected, based on previous research. Higher scores on measures of MD and lower scores on measures of normative beliefs about deviancy were expected to be related to greater self-reported engagement in bullying. As shown in Table 3, all measures were significantly correlated in the expected direction. Specifically, higher scores on MD and lower scores on normative beliefs against deviancy were moderately related to higher levels of bullying at both T1 and 2. Additionally, the moderate correlation observed between MD and normative beliefs about deviancy support treating the two variables as related yet conceptually distinct constructs (r = - .48, p < .01). Further, the positive but modest correlation observed between bullying at T1 and T2 confirms the moderate stability of the behaviour over time, with sufficient variability to suggest change over time in at least some cases. Table 4 Correlations between Measures Individual Bullying Time 1 Time 2  Individual Moral Disengagement  Individual Normative Beliefs about Deviancy  Individual Bullying Time 1 Time 2  -.43*  ---  --  --  Individual Moral Disengagement  .34*  .42*  --  --  Individual Normative Beliefs about Deviancy Note: *p < .01, N = 376  -.22*  -.39*  -.48*  --  After examining the relationships among the measures at the individual level (Table 3), the characteristics of the school, as reflected in the group level variables, were examined in regard to their relationship with individual level composites. The relationship between school 34  and individual MD, school and individual normative beliefs against deviancy, and bullying at both time points were examined (see Table 4). All relationships were statistically significant (p < .05), with the exception of the relationship between school MD and school normative beliefs about deviancy with bullying levels at Time 1 (p > .05). Table 5 Correlations between School and Student Composites Bullying Time 1 Time 2  Individual Moral Disengagement  Individual Normative Beliefs about Deviancy  School Moral Disengagement  .07  .18*  .33*  -.19*  School Normative Beliefs about Deviancy  -.02  -.19*  -.14*  .32*  Note: *p < .05, N = 38 4.5.2  Main Analysis and Rationale The data were analyzed using multilevel modeling. A multilevel modeling technique was  selected for the present study because the data involved students nested within schools, and the hypotheses involved group and individual level predictors that are multilevel in nature (Raudenbush & Bryk, 2002). Multilevel analyses account for the nesting or “clustering” of students in schools and allow for the opportunity to use both individual and group level predictors to explain differences in the dependent variable. It is important and necessary to account for these clusters statistically and standard statistical tests (e.g., regression, analysis of variance) rely on the assumption of independence of observations, and thus, do not account for the clustering of data (Hox, 2010). Multilevel analysis accounts for the clustering effect, which helps to reduce Type 1 errors. Many of the research studies that have informed the bullying literature to date have not accounted for the effect of the group in their analyses. The present 35  study adds to the bullying literature by investigating the influence of group level MD and normative beliefs about deviancy, over and above individual levels, on bullying over the course of a school year using multilevel analyses.  Assumption Testing First considered was whether the assumptions of normality and linearity, as required for  multilevel analyses, were met (Hox, 2010). To this end, level 1and level 2 residuals were plotted against their normal scores. A Q-Q plot based on level 1 indicated that the residuals were nonlinear, suggesting non-normality. However, non-normality at level 1 is less problematic in multilevel modeling than at level 2 (Maas & Hox, 2004). The standardized residuals at level 2 produced a line that was approximately diagonal which indicates conformity to normality. Three outliers appeared at the upper right hand corner of the graph. These extreme scores fell within the acceptable range of response options, and thus, were considered valid and retained in the analyses. To check the assumption of linearity, the level 1 and level 2 residuals were plotted against the predicted values. The assumption of linearity was violated at level 1, but the plotted points at level 2 appeared to have no strong structure, and were evenly distributed around zero, which means that the assumption of linearity could be retained for the multilevel model. Another assumption of multilevel models is that the variance of the residuals is the same in all groups (Hox, 2010). The assumption of equal variances was examined using a Chi-square (χ2) test, as outlined by Raudenbush and Bryk (2002). The statistically significant Chi-square (χ2 = 44.36, (1, N = 38), p < .05) observed suggests that the level 1 variances were not homogeneous across all level 2 units. Chi-square tests may also become significant if the number of students per school is small (fewer than 10) and if data are not normally distributed. Thus, although  36  results of preliminary analyses suggested some violation of assumptions for level 1 variables, level 2 variables were deemed adequate. Given the primary focus of the research was on level 2 influences, further analyses were deemed worthy of pursuit. As noted previously (see Tables 3 and 4), correlational analyses were conducted to examine the relationships among variables, and results indicated they were modest in magnitude, in the expected directions, and statistically significant. Additionally, the relationship between school MD and school normative beliefs towards deviancy was moderately and negatively correlated, with higher levels of moral disengagement in a school associated with lower normative beliefs towards deviant behaviour (r = -.49, p < .001). Therefore, there was no evidence of multicolinearity.  Multilevel Analysis Results The first step in a multilevel analysis is to run a fully unconditional (null) model, where  within-group and between-group variability of the outcome variable are estimated. An intraclass correlation (ICC) is calculated to determine the ratio of variability between schools to variability within schools. In the present study, a high ICC would indicate that schools differ in the amount of bullying and thus school level factors can be explored (i.e., a multilevel analysis is supported; Hox, 2010). The unconditional model for the present study, with bullying at T2 as the outcome variable, indicated no statistically significant variation across schools (χ2 = 39.61, (1, N = 38, p > .05). Specifically, the ICC indicated than less than 2% of the variance in bullying was due to school level factors. A unique element of the present study was that previous reports of bullying could be statistically controlled while examining the influence of school level factors. Therefore, a conditional model with bullying at T2 as the outcome, controlling for bullying at T1 was  37  examined despite the lack of significant variation in the null model. Results from the conditional model also indicated no statistically significant variation across schools in T2 bullying, after accounting for T1 bullying (χ2 = 42.34, (2, N = 38, p > .05). In conclusion, the use of multilevel modeling was not supported for the current study.  Multiple Regression Results Given that multilevel modeling was not possible with the present data set, a series of  multiple regression analyses were conducted to examine the influence of moral disengagement and normative beliefs about deviancy, at the individual and school level, on bullying behaviour. Gender was initially included in all of the analyses, but did not predict bullying therefore was removed from subsequent analyses (p > .05). First, the relationship between individual level MD and bullying was explored. Results from a simple regression analysis predicting T2 bullying from T2 moral disengagement revealed that individual level MD in Grade 6 (Step 1) predicted students’ levels of bullying during the same year (β = .42, p < .05), and accounted for 18% of the variance (R2 = .18, F(1,375) = 87.4, p < .05). A subsequent multiple regression analysis verified that the relationship between individual MD and bullying in Grade 6 (Step 2) held even after controlling for Grade 5 (Step 1) levels of bullying (β = .32, p < .05). Together, both individual MD and previously reported bullying explained 27% of the variance in Grade 6 bullying (R2 = .27, F(2,374) = 69.61.87, p < .05). These results (see Table 5) are consistent with previous findings that support the relationship between individual levels of MD and bullying behaviour (Barchia & Bussey, 2011; Gini, 2006; Hymel et al., 2005; Menesini et al., 2003; Obermann, 2011).  38  Table 6 Multiple Regression Analysis Predicting T2 Bulling Behaviour from T1 Bullying and T2 Individual Moral Disengagement Independent Variable B S.E. B β R2 F T1 Bullying  .28  .04  .33*  --  --  T2 Individual Moral Disengagement  .18  .03  .30*  --  --  .27  69.61  Model Summary *p < .05  Next, the relationship between T2 individual level beliefs about deviancy (i.e., the degree to which students believe that delinquent behaviours are wrong), and T2 bullying was explored. Results of a simple regression analysis indicated that student’s beliefs about deviancy in Grade 6 (Step 1) predicted their reported bullying involvement in the same year (β = -.42, p < .05), accounting for 18% of the variance (R2 = .18, F(1,375) = 94.8, p < .05). Specifically, the stronger a student’s belief that deviant behaviour is wrong the less likely they were to report bullying involvement in Grade 6. Further, results of a follow-up regression analysis showed that this relationship (Step 2) held after controlling for Grade 5 (Step 1) bullying (β = -.31, p < .05). Together, both individual beliefs about deviancy and previously reported bullying explained 28% of the variance in Grade 6 bullying (R2 = .28, F(2,374) = 73.17, p < .05). Again, these results (see Table 6) are consistent with the influence of individual normative beliefs about deviancy and bullying.  39  Table 7 Multiple Regression Analysis Predicting T2 Bulling Behaviour from T1 Bullying and T2 Individual Normative Beliefs about Deviancy Independent Variable B S.E. B β R2 F T1 Bullying  .31  .04  .37*  --  --  T2 Individual Normative Beliefs about Deviancy  1.14  .02  -.31*  --  --  --  --  --  .28  73.17  Model Summary *p < .05  Finally, of interest in the present study was whether the combined understanding both an individual’s level of MD and beliefs towards deviancy provided additional information about their Grade 6 bullying. Results of a multiple regression analysis (see Table 7) indicated that student levels of MD and beliefs about deviancy in Grade 6 (Step 2) accounted for 31% of the variance in their bullying involvement for that year, even after controlling for their bullying involvement in Grade 5 (Step 1; R2 = .31, F(3,373) = 56.01, p < .05). Table 8 Multiple Regression Analysis Predicting T2 Bulling Behaviour from T1 Bullying, T2 Individual Moral Disengagement and T2 Individual Normative Beliefs about Deviancy Independent Variable B S.E. B β R2 F T1 Bullying T2 Individual Moral Disengagement T2 Individual Normative Beliefs about Deviancy Model Summary  .27  .04  .32*  --  --  .12  .03  .20*  --  --  -.10  .02  -.23*  --  --  --  --  --  .31  56.01  *p < .05  40  In order to examine the influence of school level factors, two additional multiple regression analyses were conducted. The first examined the contribution of school MD, over and above individual levels of MD, on bullying in Grade 6, controlling for previous levels of bullying. For this analysis, Grade 5 bullying was entered in Step 1 and individual level of MD was entered Step 2, with school level MD entered in Step 3. Results from the analysis revealed that school MD did not significantly predict bullying, over and above individual levels of MD, after controlling for previous levels of bullying (β = .07, p > .05). Table 9 Multiple Regression Analysis Predicting T2 Bulling Behaviour from T2 Individual Moral Disengagement and T2 School Moral Disengagement Independent Variable B S.E. B β R2 T1 Bullying T2 Individual Moral Disengagement T2 School Moral Disengagement Model Summary  F  .29  .04  .34*  --  --  .16  .03  .28*  --  --  .12  .08  .07  --  --  --  --  --  .28  47.33  *p < .05 A second multiple regression analysis examined the influence of school normative beliefs about deviancy (Step 3), over and above individual levels of normative beliefs about bullying (Step 2), on bullying in Grade 6, after controlling for previous levels of bullying (Step 1). As was found for MD, school normative beliefs about deviancy did not significantly predict bullying, over and above individual beliefs towards deviancy, after controlling for previous levels of bullying (β = .03, p > .05). To summarize, the results of these two regression analyses showed that school levels of MD and normative beliefs about deviancy did not predict bullying involvement in Grade 6 after controlling for individual level normative beliefs and MD, as well 41  as previous bullying involvement. These results fail to support the hypotheses of the present study. The results are also contrary to previous reports that group factors influence bullying behaviour (Esplage et al., 2003; Guerra et al., 2011), and specifically that MD at the school level predicts bullying, over and above individual levels of MD (Vaillancourt et al., 2006). Table 10 Multiple Regression Analysis Predicting T2 Bulling Behaviour from T2 Individual Normative Beliefs about Deviancy and T2 School Normative Beliefs about Deviancy Independent Variable B S.E. B β R2 F T1 Bullying T2 Individual Normative Beliefs about Deviancy T2 School Normative Beliefs about Deviancy Model Summary  .31  .04  .36*  --  --  -.15  .02  -.32*  --  --  .04  .06  .03  --  --  --  --  --  .28  48.50  *p < .05 A final set of analyses were conducted to explore the inconsistencies observed between the findings from the present study and those reported by Vaillancourt and colleagues (2006), and whether the differences could be explained by the variations in the analyses conducted. Specifically, Vaillancourt et al. did not control for previous bullying in their analyses, whereas the present study did. Accordingly, a series of multiple regression analyses were conducted examining school level influences of MD and normative beliefs about deviancy at each time point (i.e., Grade 5, Grade 6) without controlling for prior bullying. If school level variables significantly predicted bullying at each time point, over and above individual level factors, then the current study would replicate the findings from Vaillancourt and colleagues (2006) study.  42  Specifically, for these multiple regression analyses (one for T1 data and one for T2 data), individual level MD as well as individual normative beliefs about deviance were entered in Step 1, and school level MD as well a school level normative beliefs were entered in Step 2, in order to predict concurrent reports of bullying involvement. Results indicated that for both T1 and T2, neither school MD (β = .04, p > .05 for T1; β = .09, p >.05 for T2) nor school normative beliefs about deviancy (β = .07, p > .05 for T1; β = .08, p >.05 for T2) significantly predicted T2 bullying, after controlling for individual levels of MD and normative beliefs about deviancy. These results (see Tables 9 and 10) are inconsistent with the current hypothesis and previous reports (Vaillancourt et al., 2006). Potential explanations for the inconsistencies are outlined in the discussion. Table 11 Multiple Regression Analyses Predicting T1 Bulling Behaviour from T1 Individual and T1 School Variables Independent Variable B S.E. B β R2 T1 Individual Moral Disengagement T1 Individual Normative Beliefs about Deviancy T1 School Moral Disengagement T1 School Normative Beliefs about Deviancy Model Summary  F  .28  .04  .41*  --  --  -.10  .03  -.16*  --  --  .08  .11  .04  --  --  .12  .10  .07  --  --  --  --  --  .25  30.24  *p < .05  43  Table 12 Multiple Regression Analyses Predicting T2 Bulling Behaviour from T2 Individual and T2 School Variables Independent Variable B S.E. B β R2 T2 Individual Moral Disengagement T2 Individual Normative Beliefs about Deviancy T2 School Moral Disengagement T2 School Normative Beliefs about Deviancy Model Summary  F  .16  .03  .27*  --  --  -.13  .09  .09*  --  --  .15  .09  .09  --  --  .12  .08  .08  --  --  --  --  --  .23  27.77  *p < .05  44  Chapter 5: Study 2 Given the failure to replicate previous research on the influence of group level MD and normative beliefs about deviancy on bullying in the preceding analyses, a second investigation was conducted using an existing data set to examine whether the lack of statically significant findings of school level variables was unique to the current sample, and thus, could be explained by the methodological limitations of the present study (e.g., number of students per school, measure of MD). In order to investigate this possibility, a multilevel analysis examining the contextual level effects of MD on bullying was examined using a novel sample of schoolchildren in which more extensive data on peer group level MD was available and well-established measure of MD was used. 5.1  Method Data for the current investigation were drawn from the second year of the School Climate  Study (2011), a three-year project conducted by Drs. David Smith, Shelley Hymel, and John LeBlanc, aimed at evaluating the social and academic atmospheres of schools. The data were collected in May and June of 2011 from 5 school districts and 19 elementary schools (69% participation rate) in western Canada. For this study, all participating students completed an extensive survey on school climate, bullying behaviour, bystander beahviour, and MD in their classrooms (90-minute sessions) using paper-and-pencil responses or digital response recorders (“clickers”). Students were assured of the confidentiality of their responses. For the present analyses, only data obtained for MD and bullying were considered. 5.2  Participants The initial sample consisted of 1278 Grade 6 and 7 students (54% girls) from 19 schools  located in southern British Columbia. For the present evaluation, classrooms were used as the  45  unit of analysis to investigate the influence of contextual effects on bullying behaviour. This decision was based on sample size; a total of 74 classroom participated in the study, compared to only 19 schools. Out of the 1278 Grade 6 and 7 students, 73 students belonged to classes were fewer than 10 students participated in the study and 77 were omitted due to missing data, and thus were deleted from further analyses. Again the decision rules were based on optimizing the number of students per classroom while maintaining large enough classes overall to minimize potential bias in the estimates of the standard errors (Maas & Hox, 2004). Therefore, the final sample consisted of 1128 students across 74 classrooms were used for the analyses, with an average of 18 students per classroom (range = 10 to 28). All students received parent consent prior to participation. Student assent for participation was also required. Only students who received parent permission who themselves agreed to participate were included in the present study. The sample was ethnically diverse: 31% Caucasian, 26% South Asian, 20% Asian, 12% mixed, 4% other, 2% Latin American, 2% “I don’t know”, 1% Middle Eastern, 1% Aboriginal, and 1% African/Caribbean. 5.3  Measures Bullying. Bullying involvement was measured using a 5-item measure that was similar to  that used in the MacMaster Teen Study (Study 1). Students first read a definition of bullying that distinguished bullying and teasing, and identified the three major features of bullying (i.e., it is ongoing, intentional, and includes a power imbalance). Students then responded to five items regarding their own involvement in bullying others both generally and with regard to four different types of bullying: physical, verbal, relational and cyber. Responses to each item were made on a 5-point Likert-type scale (1 = never, 2 = only once or a few times, 3 = every month, 4 = every week, 5 = several times a week). A composite measure including all bullying items (i.e.,  46  a general question on bullying, and questions regarding the four forms of bullying) computed as the average of responses to all five items, with higher scores indicating more engagement in bullying perpetration (see Appendix D). Internal consistency for the bullying measure in the present sample was high (α = .80). Moral Disengagement. Student levels of MD were assessed using Bandura’s (1999; 2002) Moral Disengagement Scale, a 32-item measure developed to measure MD (see Appendix E). In contrast to the MD measure used in the preceding analyses, which focused on MD with regard to bullying per se, the items included on the present scale reflected more general moral disengagement tendencies (e.g., “It’s alright to fight and protect your friends.”, “Some people deserve to be treated badly.”, “Kids are not at fault for misbehaving if their parents push them too much.”). Internal consistency for the present sample was high (α = .91). Student responses to the 32 MD were averaged to create an overall individual level composite, with higher scores reflecting greater MD. Group level indices of MD were computed as the average of overall scores on the MD measure obtained from all participants within a classroom. 5.4  Preliminary Analyses & Assumption Testing It was determined that the data met the assumptions of linearity, normality, homogeneity  of variances, and multicolinearity. Results of correlational analyses revealed a statistically significant relationship between gender and all variables of interest (p < .05), thus, was controlled for in the multilevel modeling analysis. In general, boys reported higher levels of moral disengagement (r = -.25, p < .05) and slightly more bullying than girls (r = - .07, p < .05). All other variables were significantly and moderately correlatedin the expected directions.  47  Table 13 Correlations between Classroom and Student Composites Individual Individual Moral Bullying Disengagement  Classroom Moral Disengagement  Gender  -.07*  -.25*  -.07*  Classroom Moral Disengagement  .09*  .36*  --  Note: *p < .05 5.5  Results Calculation of the ICC indicated that 3% of the variation in bullying was due to  classroom level effects (p < .05), which supported the use of a multilevel model. Next, the level 1 conditional model with gender and individual MD was examined. Results revealed that individual MD predicted bullying behaviour (β1 = 0.33, t(71) = 8.3, p <.001), after controlling for gender, which was not significant (β2 = 0.01, t(1028) = 8.3, p > .05). The level 2 conditional model comparing the effects of MD at both the individual and classroom levels on self-reported bullying, controlling for gender, was then examined. Results indicated that MD at the classroom level did not significantly predict bullying behaviour, after controlling for individual levels of MD (µ01 = -0.12, t(70) = -1.56, p > .05). These results are consistent with the previous analyses conducted, and reasons for present results and implications for future research are discussed in the section to follow.  48  Chapter 6: Discussion The current study examined the effects of MD and normative beliefs about deviancy, at the individual and group level, on bullying in schools. Results across multiple analyses and two independent samples did not support the hypothesis that contextual levels of MD (Study 1 and 2) nor normative beliefs about deviancy (Study 1) influenced bullying either concurrently (Study 1 and 2) or over time (Study 1). Specifically, in Study 1, results from the initial multilevel analyses indicated that there was not significant variability in student reports of bullying behaviour between schools to support multilevel analyses. As an alternative, a series of multiple regression analyses were conducted to address the questions of interest. Results from these analyses indicated that individual levels of MD, normative beliefs about deviancy, and previous bullying behaviour together accounted for 31% of the variance in bullying. Neither school MD nor school normative beliefs about bullying significantly predicted bullying behaviour, in any of the subsequent multiple or multilevel regression analyses. In Study 2, using a more extensive but more general measure of MD, and evaluating group level MD within classrooms rather than schools, the same pattern of results emerged. Although individual levels of MD were predictive of concurrent bullying, group levels of MD did not add significantly to this prediction, contrary to expectations. These results fail to replicate findings reported by Vaillancourt and colleagues (2006), which indicated that school levels of MD predicted bullying over and above individual levels of MD. One explanation for these differences is the diversity in the populations examined across studies. Vaillancourt et al. (2006) investigated school levels of moral disengagement across 116 schools, containing students from Grades 4 through 12. The variability in the sample may have provided opportunity for more heterogeneous responses of bullying behaviour across schools,  49  although ICC’s were not reported. In the current study, students in elementary and middle schools (Grade 5/6, Study 1) and elementary classrooms (Grade 6/7, Study 2) were examined. One possibility is that older students (e.g., high school), as considered in the original Vaillancourt et al. research, may be influenced differently by peer attitudes and norms than younger students (Henry et al., 2000), and examining this relationship may elucidate some differences in the results across studies. Several methodological limitations of the present study may have also contributed to the lack of statistically significant findings. In particular, the present study assessed MD and normative beliefs at the contextual level by aggregating means from the responses of individuals who belonged to the same group (i.e., school, class). In both studies the number of peers contributing to the assessment of group level MD (and normative beliefs in Study 1) were limited, leading to questions regarding the validity of the present assessments of group level beliefs and perceptions. This was particularly problematic in Study 1, with an average of 12 students (range of 5-44) contributing to assessments of “school level” moral disengagement and normative beliefs. This problem was less evident in Study 2 in which group level beliefs were assessed at the classroom level, yet even here some classrooms were omitted due to limited numbers of participating peers. It is possible that group level MD was assessed on the basis of a more representative sample of peers in the Vaillancourt et al. research. A related concern involves the way in which group level beliefs and perceptions were assessed. Across both Vaillancourt et al. (2006) and the present study, group level variables were computed as the average of responses across participants in the group. However, actual group perceptions may not be consistent with what individuals perceive to be the group’s beliefs or norms. It may have been more appropriate to ask students about the MD they observed or  50  perceived in their schools/classrooms (e.g., How true is the following statement, “Students in my school think it is okay to bully someone if they are different.”). Asking students about their own perceptions of peer norms directly may provide greater variability in the responses across contexts and individuals, and thus, enable researchers to better detect potential influences of peer MD and normative beliefs about deviancy on bullying. Another methodological consideration for the present investigation concerns the ways in which MD was measured. In Study 1, the first set of analyses examined school factors of MD and normative beliefs towards bullying using multiple regression, where students reported their levels of MD about bullying behaviour. In Study 2, classroom levels of MD were analyzed using multilevel modeling, where students reported on their general levels of MD (Bandura, 1999; 2001) rather than MD with regard to bullying per se. Although no studies to date have examined the relationship between general measures of MD (Bandura, 1999; 2001), and bullying-specific indices of MD, the two measures share similar features. First, both measures were based on Bandura’s social cognitive theory of moral agency (2001) to assess individual levels of moral disengagement. Additionally, the results across studies using either of the measures appears to be consistent: students who reported higher levels of MD also tended to report engaging in more bullying behaviour (e.g., Hymel et al., 2005; Obermann, 2011; Vaillancourt et al., 2006). The consistent findings between Bandura’s MD scale and the MD composite used in the present study provide support for the generalizability of the current findings at the individual level. The results of the current investigation add to the literature on bullying and MD at the individual level in two ways. First, the strong relationship between student MD and bullying at the individual level was supported across diverse samples (i.e., a large sample from British Columbia, a small sample from Ontario), in various contexts (i.e., schools, classrooms), and  51  across between different measures of MD. Second, the strong relationship between MD and bullying at the individual level did not hold when MD was aggregated at the school or classroom level. It is clear from the current study that investigating contextual level effects of MD would benefit from the development of interviews or scales that aim to measure the construct of MD at the level of interest (i.e., group). Tapping student’s “perceptions” of the constructs at the group level, rather than their individual beliefs may be a better predictor of their behaviour, although the challenge with this is shared method variance. In addition to exploring MD at the school or classroom level, future research in the area may benefit from examining how school level factors, such as school climate, are related to individual MD and bullying. For example, it would be interesting to explore how the construct of collective efficacy, or the belief that one’s peer group or community would be willing to provide support and protection in a school (Williams & Guerra, 2011), would influence the relationship between individual MD and bullying using multilevel analyses. The closer the bullying literature comes to understanding the various relationships between and among contextual and individual factors, the better researchers can inform schools and educators on how to effectively intervene in bullying behaviour. Research in the area of MD, normative beliefs about deviancy, and bullying may also benefit from considering the various factors that influence the contextual climate, such as teachers. Teachers are a major component of schools, and are often times the leaders in implementing classroom strategies aimed at reducing bullying. Teachers have reported that they understand the definition of bullying and its harmful effects on students, yet they have also reported not always intervening in bullying behaviour (Mishna, Scarcello, Pepler, & Weiner,  52  2005). Examining teacher levels of MD may be another avenue for future research, and would be especially insightful if combined with other contextual indicators MD. Although more research is required in order to better understand the relationship between social processes, such as group levels of MD and normative beliefs about deviancy, the current research supported the established link between individual MD and normative beliefs, as well as previous levels of bullying, on student levels of bullying behaviour. The lack of statistically significant findings of group level processes in the current study has provided direction for the assessment and design of future studies examining these constructs.  53  References Almeida, A., Correia, I., & Marinho, S. (2010). Moral disengagement, normative beliefs of peer group, and attitudes regarding roles in bullying. Journal of School Violence, 9, 23-36. doi:10.1080/15388220903185639 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. doi:10.1177/0272431605276933 Asch, S. E. (1956). Studies of independence and conformity: A minority of one against a unanimous majority. Psychological Monographs, 70, 1-70. doi:10.1037/h0093718 Bandura, A. (1977). Social Learning Theory. New York, NY: General Learning Press. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1991). Social cognitive theory of moral thought and action. In W.M. Kurtines & J. L. Gerwirtz (Eds.), Handbook of moral behavior and development (Vol. 1, pp. 45-103). Hillsdale, NJ: Erlbaum. Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3, 193-209. doi: 10.1207/s15327957pspr0303_3 Bandura, A. (2002). Selective moral disengagement in the exercise of moral agency. Journal of Moral Education, 31,101-119. doi:10.1080/0305724022014322 Bandura, A., Caprara, G.V., Barbaranelli, C., Pastorelli, C., Regalia, C. (2001). Sociocognitive self-regulatory mechanisms governing transgressive behaviour. Journal of Personality and Social Psychology, 80, 125-135. doi:10.1037/0022-3514.80.1.125  54  Bandura, A., Underwood, B., & Fromson, M. E. (1975). Disinhibition of aggression through diffusion of responsibility and dehumanization of victims. Journal of Personality and Social Psychology, 9, 253-269. Retrieved from http://www.des.emory.edu/mfp/Bandura1975.pdf Barchia, K., & Bussey, K. (2010). Individual and collective social cognitive influences on peer aggressive: Exploring the contribution of aggression efficacy, moral disengagement, and collective efficacy. Aggressive Behavior, 35, 1-14. doi:10.1002/ab.20375 Biggam, F. H. & Power, K. G. (1999). Social problem-solving skills and psychological distress among incarcerated young offenders: The issue of bullying and victimization. Cognitive Therapy and Research, 23, 307-326. doi:10.1023/A:1018795632067 Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529. Retrieved from http://persweb.wabash.edu/facstaff/hortonr/articles%20for%20class/ baumeister%20and%20leary.pdf Bosworth, K., Espelage, D. L., & Simon, T. R. (1999). Factors associated with bullying behavior in middle school students. The Journal of Early Adolescence, 19, 341-362. doi:10.1177/0272431699019003003 Campbell, W. H., Missuina, C., & Vaillancourt, T. (2012). Peer victimization and depression in children with and without motor coordination difficulties. Psychology in the Schools, 0, 1-14. doi:10.1002/pits.21600 Caravita, S. & Gini, G. (2012). Morality in Bullying: What is right and what is wrong? Education.com, Special Issue on Bullying (www.education.com).  55  Card, N. A., Isaacs, J., & Hodges, E. V. E. (2007). Correlates of school victimization: Recommendations for prevention and intervention. In J. E. Zins, M. J. Elias, & C. A. Maher (Eds.), Bullying, Victimization, and Peer Harassment: A Handbook of Prevention and Intervention (pp. 339-368). New York: Haworth Press. Clark, R., Dogan, R. R., & Akbar, N. J. (2003). Youth and parental correlates of externalizing symptoms, adaptive functioning, and academic performance: An exploratory study in preadolescent blacks. doi:10.10170S095457940606007X Clarke, P., & Wheaton, B. (2007). Addressing data sparseness in contextual population research using cluster analysis to create synthetic neighborhoods. Sociological Methods & Research, 35, 311-351. doi: 10.1177/0049124106292362 Craig, W. (1998). The relationship among bullying, victimization, depression, anxiety, and aggression in elementary school children. Personality and Individual Differences, 24, 123-130. doi:10.1016/S0191-8869(97)00145-1 Craig, W. M., & Pepler, D. J. (1997). Observations of bullying and victimization in the school yard. Canadian Journal of School Psychology, 13, 41-59. doi:10.1177/082957359801300205 Crick, N. R., & Grotpeter, J. K. (1995). Relational aggression, gender, and social- psychological adjustment. Child Development, 66, 710-722. doi:10.2307/1131945 Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297-334. doi:10.1007/BF02310555 Dao, T. K., Kerbs, J. J., Rollin, S. A., Potts, I., Gutierrez, R., … & Prevatt, F. (2006). The association between bullying dynamics and psychological distress. Journal of Adolescent Health, 39, 277-282. doi:10.1016/j.jadohealth.2005.11.001  56  Dickens, C. (1839). Oliver Twist. London, UK: R. Bentley. Dishion, T. J. (1990). The peer context of troublesome child and adolescent behavior. In P. Leone (Ed.), Understanding troubled and troublesome youth (pp. 128– 153). Newbury Park, CA: Sage. Dodge, K. A. (1991). The structure and function of reactive and proactive aggression. In D. J. Pepler & K. H. Rubin (Eds.), The Development and Treatment of Childhood Aggression (pp. 201-216). Hillsdale, NJ: Erlbaum. Duncan, R. D. (1999). Peer and sibling aggression: An investigation of intra-and extra-familial bullying. Journal of Interpersonal Violence, 14, 871-886. doi:10.1177/088626099014008005 Espelage, D. L., Holt, M. K., & Henkel, R. R. (2003). Examination of peer-group contextual effects on aggression during early adolescence. Child Development, 74, 205–220. doi:10.1111/1467-8624.00531 Fekkes, M., Pijpers, F. I. M., Fredriks, A. M., Vogels, T., & Verloove-Vanhorick, S. P. (2006). Do bullied children get ill, or do ill children get bullied? A prospective cohort study on the relationship between bullying and health-related symptoms. Pediatrics, 117, 15681574. doi:10.1542/peds.2005-0187 Garandeau, C. F., & Cillessen, A. H. N. (2006). From indirect aggression to invisible aggression: A conceptual view on bullying and peer group manipulation. Aggression and Violent Behavior, 11, 612– 625. doi:10.1016/j.avb.2005.08.005 Gibb, S. J., Horwood, J. L., & Fergusson, D. M. (2011). Bullying, victimization/perpetration, in childhood and later adjustment: findings from a 30 year longitudinal study. Journal of Aggression, Conflict and Peace Research, 3, 82-88. doi: 10.1108/17596591111132891  57  Gini, G. (2006). Social cognition and moral cognition in bullying: What’s wrong? Aggressive Behaviour, 32, 528-539. doi:10.1002/ab.20153 Gini, G., Pozzoli, T., & Hauser, M. (2011). Bullies have enhanced moral competence to judge relative to victims, but lack moral compassion. Personality and Individual Differences, 50, 603-608. doi: 10.1016/j.paid.2010.12.002 Greco, L. A., Freeman, K. E., & Dufton, L. (2007). Overt and relational victimization among children with frequent abdominal pain: Links to social skills, academic functioning, and health service use. Journal of Pediatric Psychology, 32, 319–329. doi:10.1093/jpepsy/jsl016 Griffin, R. S., & Gross, A. M. (2004). Childhood bullying: Current empirical findings and future directions for research. Aggression and Violent Behavior, 9, 379–400. doi:10.1016/S1359-1789(03)00033-8 Guerra, N. K., Williams, K. R., & Sadek, S. (2011). Understanding bullying and victimization during childhood and adolescence: A mixed method study. Child Development, 82, 295310. doi: http://dx.doi.org/10.1111/j.1467-8624.2010.01556.x Harris, J. R. (1995). Where is the child's environment? A group socialization theory of development. Psychological Review, 102, 458-489. doi:10.1111/j.14678624.2010.01556.x Harris, J. R. (1998/2009). The nurture assumption: Why children turn out the way they do. New York: The Free Press. Hartup, W. W. (2005). The development of aggression: Where do we stand? Developmental Origins of Aggression (pp. 3-22). New York: Guilford Press.  58  Haselager, G. J. T., Hartup, W. W., Van Lieshot, C. F. M., & Risken-Walraven, J. M. A. (1998). Similarities between friends and nonfriends in middle childhood. Child Development, 69, 1198-1208. doi:10.2307/1132369 Hawker, D. S. J., & Boulton, M. J. (2000). Twenty years’ research on peer victimization and psychosocial maladjustment: A meta-analytic review of cross-sectional studies. Journal of child Psychology and Psychiatry, 41, 441-455. doi: http://dx.doi.org/10.1111/14697610.00629 Heinrich, L. M., & Gullone, E. (2006). The clinical significance of loneliness: A literature review. Clinical Psychology Review, 26, 695-718. doi:10.1016/j.cpr.2006.04.002 Henry, D., Guerra, N. G., Huesmann, L. R., Tolan, P. H., VanAcker, R., & Eron, L. D. (2000). Normative influences on aggression in urban elementary school classrooms. American Journal of Community Psychology, 28, 59–81. doi:10.1023/A:1005142429725 Henry, D. B., Farrell, A. D., Schoeny, M. E., Tolan, P. H., & Dymnicki, A. B. (2011). Influence of school-level variables on aggression and associated attitudes of middle school students. Journal of School Psychology, 48, 481-503. doi:10.1016/j.jsp.2011.04.007 Hodges, E., & Perry, D. (1996). Victims of peer abuse: An overview. Journal of Emotional and Behavioral Problems, 5, 23-28. Hox, J. (2010). Multilevel analysis: Techniques and observations. UK: Routeledge. Huesmann, L. R., & Guerra, N. G. (1997). Children's normative beliefs about aggression and aggressive behavior. Journal of Personality and Social Psychology, 72, 408–419. doi:10.1037/0022-3514.72.2.408 Hymel, S., Rocke Henderson, N. & Bonanno, R. (2005). Moral disengagement: A framework for understanding bullying among adolescents. Special issue of the Journal Of Social  59  Sciences on “Peer Victimization in Schools: An International Perspective”, edited by Drs. O. Aluede, A.G. McEachern & M.C. Kenny, Special Issue No. 8, 1-11. Retrieved from http://www.krepublishers.com/06-Special%20Volume-Journal/JSS-00Special%20Volumes/JSS-SI-08-Peer-Vicit-Schools-Web/JSS-SI-08-01-001-011-HymelS/JSS-SI-08-01-001-011-Hymel-S-Text.pdf. Hymel, S. & Swearer, S. (2008). Bullying: An age-old problem that needs new solutions. Education.com, Special Issue on Bullying (www.education.com). Jessor, R., Donovan, J. E., & Costa, F. M. (1989). Attitudinal intolerance of deviance, friends modeling problem behaviors, and expectations of academic competence scales. Unpublished instruments, University of Colorado, Institute of Behavioral Science. Jessor, R., Van Den Bos, J., Vanderryn, J., Costa, F. M., & Turbin, M. S. (1995). Protective factors in adolescent problem behavior: Moderator effects and developmental change. Developmental Psychology, 31, 923-933. Retrieved from http://www.colorado.edu/ibs/jessor/pubs/1995_Jessor_VanDenBos_Vanderryn_etal_DE VPSYCH_ProblemBehaviorModeratorEffects.pdf Jetten, J., Spears, R., & Manstead, A. S. R. (1996). Intergroup norms and intergroup discrimination: Distinctive self-categorization and social identity effects. Journal of Personality and Social Psychology, 71, 1222-1233. doi:10.1037/0022-3514.71.6.1222 Jetten, J., Spears, R., & Manstead, A. S. R. (1997). Strength of identification and intergroup differentiation: the influence of group norms. European Journal of Social Psychology, 27, 603-609. doi:10.1002/(SICI)1099-0992(199709/10)27:5<603::AIDEJSP816>3.0.CO;2-B Jetten, J., Postmes, T., & McAuliffe, B. J. (2002). ‘We’re all individuals’: Group norms of  60  individualism and collectivism, levels of identification and identity threat. European Journal of Social Psychology, 32, 189-207. doi:10.1002/ejsp.65 Judd, C. M., Jessor, R., & Donovan, J. E. (1986). Structural equation models and personality research. Journal of Personality, 54, 149-198. doi: 10.1111/j.1467-6494.1986.tb00392.x Juvonen, J., Graham S., & Schuster, M. (2003). Bullying among young adolescents: The strong, weak, and troubled. Pediatrics, 112, 1231-1237. doi:10.1542/peds.112.6.1231 Kaltiala-Heino, R., Rimpela, M., Marttunen, M., Rimpela, A., & Rantanen, P. (1999) Bullying, depression, and suicidal ideation in Finnish adolescents: School survey. British Medical Journal, 319, 348-351. Retrieved from http://www.bmj.com/cgi/reprint/319/7206/348 Kim, Y. S., Leventhal, B. L., Koh, Y. J., Hubbard, A., & Boyce, W. T. (2006). School bullying and youth violence: Causes or consequences of psychological behavior? Archives of General Psychiatry, 63, 1035-1041. doi:10.1001/archpsyc.63.9.1035 Kindermann, T. A. (1993). Natural peer groups as contexts for individual development: The case of children’s motivation in school. Developmental Psychology, 29, 970-977. doi:10.1037/0012-1649.29.6.970 Koo, H. (2007). A time line of the evolution of school bullying in differing social contexts. Asia Pacific Education Review, 8, 107-116. Retrieved from https://webspace.utexas.edu/lab3346/School%20Bullying/Koo2007/Koreabullyinghistory 2007.pdf Krug, E. G., Dahlberg, L., Mercy, J. A., Zwi, A. B., & Lozano, R. (2002) World Report on Violence and Health. World Health Organization.  61  Lagerspetz, K., Björkqvist, K., & Peltonen, T. (1988). Is indirect aggression typical of females? Gender differences in aggressiveness in 11- to 12-year-old children. Aggressive Behavior, 14, 403-414. doi:10.1002/1098-2337(1988)14:6<403::AID-AB2480140602>3.0.CO;2-D Loeber, R. & Hay, D. (1997). Key issues in the development of aggression and violence from childhood to early adulthood. Annual Review of Psychology, 48, 371-410. doi:10.1146/annurev.psych.48.1.371 Maas, C. J. & Hox, J. J. (2004). Robustness issues in multilevel regression analysis. Statistica Neerlandica, 58, 127-137. Menesini, E., Sanchez, V., Fonzi, A., Ortega, R., Costabile, A., & Feudo, G. L. (2003). Moral emotions and bullying: A cross-national comparison of differences between bullies, victims and outsiders. Aggressive Behavior, 29, 515-530. doi:10.1002/ab.10060 Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W., Simons-Morton, B., & Scheidt, P. (2001). Bullying behaviors among US youth: prevalence and association with psychosocial adjustment. Journal of the American Medical Association, 285, 2094–2100. doi:10.1001/jama.285.16.2094 Nishina, A., & Juvonen, J. (2005). Daily reports of witnessing and experiencing peer harassment in middle school. Child Development, 76, 435-450. doi:10.1111/j.14678624.2005.00855.x Obermann, M. (2011). Moral disengagement in self-reported and peer-nominated school bullying. Aggressive Behaviour, 37, 133-144. doi: 10.1002/ab.20378 Olweus, D. (1978). Aggression in the schools: Bullies and whipping boys. Washington D.C.: Hemisphere Press: Wiley. Olweus, D. (1994). Annotation: Bullying at school: Basic facts and effects of a school-based  62  intervention program. Journal of Child Psychology and Psychiatry, 35, 1171-1190. doi:10.1111/j.1469-7610.1994.tb01229.x Olweus D. (1996). The Revised Olweus Bully/Victim Questionnaire. Mimeo. Bergen, Norway: Research Center for Health Promotion (HEMIL Center), University of Bergen. Pellegrini, A.D., & Long, J. (2002). A longitudinal study of bullying, dominance, and victimization during the transition from primary to secondary school. British Journal of Developmental Psychology, 20, 259–280. doi:10.1348/026151002166442 Pepler, D., Craig, W., Connolly, J., Yuile, A., & McMaster, L. (2006). A developmental perspective on bullying. Aggressive Behavior, 32, 376-384. doi:10.1002/ab.20136 Pepler, D., Jiang, D., Craig, W., & Connolly, J. (2008). Developmental trajectories of bullying and associated factors. Child Development, 79, 325-338. doi:10.1111/j.14678624.2007.01128.x Perkins, H. W., Craig, D. W., & Perkins, J. M. (2011). Using social norms to reduce bullying: A research intervention among adolescents in five middle schools. Group Processes Intergroup Relations, 14, 703-722. doi:10.1177/1368430210398004 Perren, S., Gutzwiller-Helfenfinger, E., Malti, T., & Hymel, S. (2011). Moral reasoning andemotion attributions of adolescent bullies, victims, and bully-victims. British Journal of Developmental Psychology. Published online: 14 September 2011. doi: 10.1111/j.2044-835X.2011.02059.x Pornari, C. D. & Wood, J. (2010). Peer and cyber-aggression in secondary school students: The role of moral disengagement, hostile attribution bias and outcome expectancies. Aggressive Behaviour, 36, 81-94. doi: 10.1002/ab.20336	
    63  Raskauskas, J. (2007). Involvement in traditional and electronic bullying among adolescents. Developmental Psychology, 43, 564-575. doi:10.1037/0012-1649.43.3.564 Raudenbush, S. W. & Bryk, A. S. (2002). Hierarchical linear models. Applications and data analysis methods, Second Edition. Thousand Oaks, CA: Sage Publications. Rivers, I. (2004). Recollections of bullying at school and their long term implications for lesbians, gay men, and bisexuals. Crisis: The Journal of Crisis Intervention and  Suicide  Prevention, 25, 169–175. doi:10.1027/0227-5910.25.4.169 Roberts, C., Freeman, J., Samdal, O., Schnohr, C. W., de Looze, M. E., Nic Gabhainn, S…, & the International HBSC Study Group (2009). The Health Behaviour in School-aged Children (HBSC) study: methodological developments and current tensions. International Journal of Public Health, 54, 140-150. doi: 10.1007/s00038-009-5405-9 Rodkin, P. C., Farmer, T. W., Pearl, R., & Van Acker, R. (2000). Heterogeneity of popular boys: Antisocial and prosocial configurations. Developmental Psychology, 36, 14-24. doi:10.1037/0012-1649.36.1.14 Salmivalli, C. (2010). Bullying and the peer group: A review. Aggression and Violent Behaviour, 15, 112-120. doi:10.1016/j.avb.2009.08.007 Salmivalli, C., Lagerspetz, K., Bjorkqvist, K., Osterman, K., & Kaukiainen, A. (1996). Bullying as a group process: Participant roles in their relations to social status within the group. Aggressive Behavior, 22, 1-15. doi:10.1002/(SICI)1098-2337(1996)22:1<1::AIDAB1>3.0.CO;2-T Salmivalli, C., Peets, K., & Hodges, E. V. E. (2010). Bullying. Chapter in P. K. Smith & C. H. Hart (Eds), The Wiley-Blackwell Handbook of Childhood Social Development, 2nd Edition (pp. 510-528). Oxford, UK: Wiley-Blackwell. doi:10.1002/9781444390933.ch27  64  Salmivalli, C., & Voeten, M. (2004). Connections between attitudes, group norms, and behaviors associated with bullying in schools. International Journal of Behavioral Development, 28, 246-258. doi:10.1080/01650250344000488 Schaeffer, C. M., Petras, H., Ialongo, N., Poduska, J., & Kellam, S. (2003). Modeling growth in boys’ aggressive behavior across elementary school: Links to later criminal involvement, conduct disorder, and antisocial personality disorder. Developmental Psychology, 39, 1020–1035. doi:10.1037/0012-1649.39.6.1020 Siegler, R., DeLoache, J., & Eisenberg, N. (2010). Moral Development. How Children Develop, Third Edition (544-581). New York: Worth. Slee, R. T. (1995). Peer victimization and its relationship to depression among Australian primary school students. Personality and Individual Differences, 18, 57-62. doi:10.1016/0191-8869(94)00114-8 Slonje, R. & Smith, P. K. (2008). Cyberbullying: Another main type of bullying? Scandinavian Journal of Psychology, 49, 147-154. doi:10.1111/j.1467-9450.2007.00611.x Sutton, J., Smith, P. K., & Swettenham, J. (1999). Social cognition and bullying: Social inadequacy or skilled manipulation? British Journal of Developmental Psychology, 17, 435-450. doi:10.1348/026151099165384 Swearer, S. M., Song, S.Y., Cary, P. T., Eagle, J. W., Mickelson, W. T. (2001). Psychosocial correlates in bullying and victimization: The relationship between depression, anxiety, and bully/victim status. Journal of Emotional Abuse, 2, 95-121. doi:10.1300/J135v02n02_07  65  Tremblay, R. E., Masse, L. C., Vitaro, F., & Dobkin, P. L. (1995). The impact of friends’ deviant behavior on early onset of delinquency: Longitudinal data from 6 to 13 years of age. Development and Psychopathology, 7, 649– 667. doi:10.1017/S0954579400006763 Ttofi, M. M., Farrington, D. P., Lösel, F., & Loeber, R. (2011). Do the victims of school bullies tend to become depressed later in life? A systematic review and meta-analysis of longitudinal studies. Journal of Aggression, Conflict and Peace Research, 3, 63 – 73. doi: 10.1108/17596591111132873 Turiel, E. (2006). The development of morality. In W. Damon & R. M. Lerner (Eds.), Handbook of child psychology, 6th Ed., Vol. 3: N. Eisenberg (Ed.), Social, emotional, and personality development (pp. 789-857). New York: Wiley. Underwood, M., Scott, B., Galperin, M., Bjornstad, G., & Sexton, A. (2004). An observational study of social aggression under varied conditions: Gender and developmental differences. Child Development, 75, 1538-1555. doi:10.1111/j.1467-8624.2004.00756.x Vaillancourt, T., Hymel, S., & McDougall, P. (2003). Bullying is power: Implications for schoolbased intervention strategies, Journal of Applied School Psychology, 19, 157-176. doi:10.1300/J008v19n02_10 Vaillancourt, T., Hymel, S., Duku, E., Krygsman, A., Cunningham, L., Davis, C., … & Cunningham, C. (2006). Beyond the dyad: An analysis of the impact of group attitudes and behavior on bullying. Symposium presented at the biennial meeting for the International Society for the Study of Behavioral Development, Melbourne, Australia. Vaillancourt, T., Brittain, H., Bennett, L., Arnocky, S., McDougall, P., Hymel, S…, & Cunningham, L. (2010). Places to avoid: Population-based study of student reports of  66  unsafe and high bullying areas at school. Canadian Journal of School Psychology, 25, 40-54. doi: 10.1177/0829573509358686 Mishna, F., Scarcello, I., Pepler, D., & Weiner, J. (2005). Teachers’ understanding of bullying. Canadian Journal of Education, 28, 718-738. doi: 10.2307/4126452 Williams, K. R., & Guerra, N. G. (2011). Perceptions of collective efficacy and bullying perpetration in schools. School Problems, 58, 126-143. doi: 10.1525/sp.2011.58.1.126  67  Appendix A MacMaster Teen Study: Bullying Items  Since the start of the school year (September)… 1. How often have you taken part in bullying another student? Not at all  Only a few times this year  Every month  Every week  Many times a week            2. How often have you taken part in physically bullying others by hitting, kicking, shoving, etc.? Not at all  Only a few times this year  Every month  Every week  Many times a week            3. How often have you taken part in verbally bullying others by insults, put downs, or threats? Not at all  Only a few times this year  Every month  Every week  Many times a week            4. How often have you taken part in bullying others by exclusion, rumors, or getting others not to like someone? Not at all  Only a few times this year  Every month  Every week  Many times a week            5. How often have you taken part in bullying others using the computer, text messages, or email messages pictures to threaten someone or make them look bad? Not at all  Only a few times this year  Every month  Every week  Many times a week            68  Appendix B  MacMaster Teen Study: Normative Beliefs about Deviancy Items How wrong do you think it is to do these things? 1. To take little things that don’t belong to you. Very wrong  Wrong  A little bit wrong  Not at all wrong          2. To give your teacher a fake excuse for being absent (away). Very wrong  Wrong  A little bit wrong  Not at all wrong          3. To bother people in a movie theatre even if you have been asked to stop. Very wrong  Wrong  A little bit wrong  Not at all wrong          4. To borrow $5 or so from a friend without really expecting to pay it back. Very wrong  Wrong  A little bit wrong  Not at all wrong          Very wrong  Wrong  A little bit wrong  Not at all wrong          5. To cheat on a test.  6. To skip school without a good excuse. Very wrong  Wrong  A little bit wrong  Not at all wrong         69  7. To get into fist fights with other people. Very wrong  Wrong  A little bit wrong  Not at all wrong          8. To break something that belongs to another person just to get even. Very wrong  Wrong  A little bit wrong  Not at all wrong          9. To break into a place that is locked just to look around. Very wrong  Wrong  A little bit wrong  Not at all wrong          10. To damage public or private property that does not belong to you just for fun. Very wrong  Wrong  A little bit wrong  Not at all wrong          11. To threaten a teacher because you were angry about something at school. Very wrong  Wrong  A little bit wrong  Not at all wrong          70  Appendix C MacMaster Teen Study: Moral Disengagement Items The following sentences ask you about your beliefs and attitudes about bullying. For each sentence, circle the answer to show how much you agree or disagree with the sentence.  1. Sometimes it’s okay to bully other people. Strongly Disagree  Disagree  Agree  Strongly Agree          2. In my group of friends, bullying is okay. Strongly Disagree  Disagree  Agree  Strongly Agree          3. Kids get bullied because they are different. Strongly Disagree  Disagree  Agree  Strongly Agree          4. Some kids get bullied because they deserve it. Strongly Disagree  Disagree  Agree  Strongly Agree          5. Some kids get bullied because they hurt other kids. Strongly Disagree  Disagree  Agree  Strongly Agree          71  Appendix D School Climate Study: Bullying Items  72  Appendix E School Climate Study: Moral Disengagement Items  73  74  75  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items