Violence and Victims

Violence and Victims Understanding Ecological Factors Associated with Bullying across the Elementary to Middle School Transition in the United States --Manuscript Draft-Manuscript Number:

VV-D-14-00046R1

Full Title:

Understanding Ecological Factors Associated with Bullying across the Elementary to Middle School Transition in the United States

Article Type:

Original Research

Keywords:

bullying; cluster analysis; early adolescence; middle school; transition

Corresponding Author:

Jun Sung Hong, Ph.D. Wayne State University Detroit, MI UNITED STATES

Corresponding Author Secondary Information: Corresponding Author's Institution:

Wayne State University

Corresponding Author's Secondary Institution: First Author:

Dorothy L Espelage, Ph.D.

First Author Secondary Information: Order of Authors:

Dorothy L Espelage, Ph.D. Jun Sung Hong, Ph.D. Mrinalini Rao, Ph.D. Robert Thornberg, Ph.D.

Order of Authors Secondary Information: Abstract:

This study examines socio-demographic characteristics and social-environmental factors associated with bullying during the elementary-to-middle school transition from a sample of 5th students (n = 300) in three elementary schools at Time 1. Of these, 237 participated at Time 2 as 6th grade students. Using cluster analyses, we found groups of students who reported no increase in bullying, some decrease in bullying, and some increase in bullying. Students who reported increases in bullying also reported decreases in school belongingness and teacher affiliation, and increases in teacher dissatisfaction. Students who reported decreases in bullying also reported decreases in victimization. These findings suggest that changes across the transition in students' relations to school and their teachers are predictive of changes in bullying.

Author Comments:

April 28, 2014 Roland D. Maiuro, Ph.D. Editor-in-Chief Violence and Victims Dear Dr. Maiuro: Thank you very much for giving us the opportunity to revise and resubmit our manuscript entitled, "Social-Ecological Factors Associated with Bullying across the Elementary-Middle School Transition" to Violence and Victims. All of the reviewers' suggestions were addressed and the changes in the manuscript were highlighted. We responded to each of the reviewers' suggestions (please refer to Response to Reviewers). I would like to re-submit our manuscript for publication to Violence and Victims. Although Dr. Dorothy Espelage is the first author and I am the second author, I am submitting this manuscript on behalf of the co-authors. Enclosed are all the requested documents, including the revised manuscript. I can Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

assure you that this manuscript has not been submitted elsewhere. If you have any questions or concerns, please do not hesitate to contact me. I can be reached by email at [email protected] or by phone at 313-577-9367. Thank you for your consideration. Sincerely, Jun Sung Hong, Ph.D. Assistant Professor School of Social Work Wayne State University 4756 Cass Avenue Detroit, MI 48202

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

Title Page with Author Information

Understanding Ecological Factors Associated with Bullying across the Elementary to Middle School Transition in the United States Dorothy L. Espelage University of Illinois, Urbana-Champaign Jun Sung Hong Wayne State University Mrinalini A. Rao Yale University Robert Thornberg Linköping University

Dorothy L. Espelage, Ph.D., Professor, Phone: 217-333-9139, email: [email protected], address: Department of Educational Psychology, University of Illinois, 220A Education, 1310 S. Sixth Street, Champaign, IL 61820-6925, USA; Jun Sung Hong, Ph.D. (corresponding author), Phone: 313-577-9367, Fax: 313-577-8770, email: [email protected], address: School of Social Work, Wayne State University, 4756 Cass Avenue, Detroit, MI 48202, USA; Mrinalini A. Rao, Ph.D., Postdoctoral Research Associate, Phone: 217-722-8004, email: [email protected], Center for Emotional Intelligence, Yale University, 340 Edwards Street, New Haven, CT 06520, USA; Robert Thronberg, Ph.D., Associate Professor, Phone: 46-13 282118, Fax: +46 13 282145, email: [email protected], address: Department of Behavioural Sciences and Learning, Linköping University, SE-58183, Linköping, Sweden DATE RE-SUBMITTED: April 28, 2014

Manuscript - must NOT contain any author information Click here to download Manuscript - must NOT contain any author information: Manuscript (Apr 28 2014).docx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Elementary to Middle School 1

Running head: BULLYING ACROSS THE ELEMENTARY TO MIDDLE

Understanding Ecological Factors Associated with Bullying across the Elementary to Middle School Transition in the United States

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


Abstract This study examines socio-demographic characteristics and social-environmental factors associated with bullying during the elementary-to-middle school transition from a sample of 5th students (n = 300) in three elementary schools at Time 1. Of these, 237 participated at Time 2 as 6th grade students. Using cluster analyses, we found groups of students who reported no increase in bullying, some decrease in bullying, and some increase in bullying. Students who reported increases in bullying also reported decreases in school belongingness and teacher affiliation, and increases in teacher dissatisfaction. Students who reported decreases in bullying also reported decreases in victimization. These findings suggest that changes across the transition in students’ relations to school and their teachers are predictive of changes in bullying. Keywords: bullying; cluster analysis; early adolescence; middle school; transition

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


Introduction Early adolescence is a period in which youth explore their new social roles and pursue social status, as they make a transition from elementary to middle school. Youth at this stage are exposed to a new and unfamiliar environment, with larger classrooms in a larger building, where they interact with unfamiliar peers (Bukowski, Sippola, & Newcomb, 2000). As a result, early adolescents are vulnerable to bullying (Nansel, Haynie, & Simons-Morton, 2003) as perpetrators, victims, or witnesses. In 2010, the Department of Education and the Centers for Disease Control and Prevention collaborated to develop a uniform research definition. This group defined bullying as follows: Bullying is any unwanted aggressive behavior(s) by another youth or group of youths who are not siblings or current dating partners that involves an observed or perceived power imbalance and is repeated multiple times or is highly likely to be repeated. Bullying may inflict harm or distress on the targeted youth including physical, psychological, social, or educational harm. (Gladden et al., 2014, p. 7) These behaviors include verbal and physical aggression that ranges in severity from making threats, spreading rumors, and social exclusion, to physical attacks causing injury. In this current study, bullying was defined as verbal and relational forms of perpetration. Studies on bullying during the elementary-to-middle school transition have primarily focused on individual characteristics and peer level factors drawing from dominance theory (McDougall & Hymel, 1998; Pellegrini, 2002; Pellegrini & Bartini, 2000; Pellegrini & Long, 2002). Little is known about the social-environmental factors that are associated with bullying during this transition. To fill this research gap, we build on the extant literature by examining how socio-demographic characteristics and social-environmental factors (i.e., relationships with parents, peers and

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


teachers, and sense of school belonging) predict changes in bullying during the elementary-tomiddle school transition using the ecological systems framework. Ecological Systems Framework Bronfenbrenner’s (1979) ecological systems framework has been applied to the research on bullying (Barboza et al., 2009; Espelage, 2012). In the area of school bullying and peer victimization, this framework has often been called a social-ecological framework and focuses on understanding how individual characteristics of children interact with environmental contexts or systems to promote or inhibit perpetration and victimization (Espelage, 2012; Espelage & Swearer, 2011; Hong & Espelage, 2012). This article focuses on the structures or locations where children have direct contact, which are referred to as the microsystem. Microsystem includes family, peers, and schools. The interaction between components of the microsystem is referred to as the mesosystem. Examples of a mesosystem applicable to this study are the interrelations between students and teachers, and the extent to which students relationships with parents are associated with their level of bullying perpetration at school. Socio-Demographics Researchers have examined the influence of socio-demographic variables, particularly sex differences in school transitions. Many studies report that boys in general are more likely to engage in physical bullying than girls (Espelage, Low, Rao, Hong, & Little, 2013; Nansel et al., 2001; Varjas, Henrich, & Meyers, 2009). During the 1990s, much research supported the notion that girls are socialized to exercise more relational forms of aggression or social bullying, while boys engage in multiple forms of aggression (Neal, 2007). Despite these findings, several studies have failed to document significant sex differences in relational aggression or social forms of bullying (Card, Stuckey, Sawalani, & Little, 2008; Crick, Casas, & Mosher, 1997). In addition to

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


sex, race/ethnicity has been another major focus of research, and higher frequency of bullying perpetration and victimization among African American students has been reported (Belgrave, 2009; Koo, Peguero, & Shekarkhar, 2012; Wang, Iannotti, & Nansel, 2009). When African American youth report more bullying perpetration (Carlyle & Steinman, 2007; Espelage, Basile, & Hamburger, 2012; Low & Espelage, 2012), these studies have yielded small effect sizes. Thus, the research on both sex and race/ethnicity differences in reports of bullying perpetration are inconsistent and limited. Thus, we include these variables in our models. Caregivers According to social learning theory and social interaction learning theory in particular, one can assert that maladaptive and aggressive social interactions with peers originates in conflictual family dynamics (Patterson, 1982; Patterson, Dishion, & Bank, 1984). Thus, it is not surprising that family conflict has been longitudinally linked to bullying perpetration among middle school youth (Espelage, Low, Rao, Hong, & Little, 2013). In contrast, when bullying victims have warm relationships with their families, they have more positive outcomes, both emotionally and behaviorally (Bowes, Maughan, Caspi, Moffitt, & Arseneault, 2010; Holt & Espelage, 2007). Thus, we hypothesize that students who report greater trust and communication with their caregivers will report fewer increases in bullying across the elementary-middle school transition than students who report alienation and low trust and communication with their caregivers. Peers Given the ecological systems framework that individual characteristics of adolescents interact with group-level factors, many researchers have examined how peer groups contribute to bullying (Espelage, Holt, & Henkel, 2003; Rodkin & Hodges, 2003). Peer relations in middle

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


school differ from those in elementary school, with interactions becoming more frequent. Bullying increases during the transition from elementary to middle school as students become more engaged in positioning themselves and others in a social status or dominance structure (Pellegrini, 2002; Pellegrini & Bartini, 2000; Pellegrini & Long, 2002). Bullying is also viewed as less negative as students enter early adolescence, and those who affiliate with peers who bully others are likely to engage in this behavior (Cook, Williams, Guerra, Kim, & Sadek, 2010). Research also indicates that youth who frequently engage in bullying can be socially accepted and popular or socially rejected and withdrawn (Farmer et al., 2010). Thus, we hypothesize that changes in bullying perpetration will be associated with both concomitant changes in social acceptance and withdrawn behaviors across the elementary-middle school transition. Teacher-Student Relationships The elementary to middle school transition can leave early adolescents feeling vulnerable because of a major discrepancy between the needs of the students and the availability of teacher support (Becker & Luthar, 2002). Middle school teachers are perceived as less caring and supportive than those in elementary schools (Burchinal et al., 2008) due to less structure and supervision in middle schools (Vaillancourt et al., in press). However, students who perceived their teachers as supportive and involved are more likely to do well in school, and less likely to display behavior problems such as bullying (Meehan, Hughes, & Cavell, 2003). Thus, we hypothesize that students will report an increase in bullying perpetration if they report increases in dissatisfaction with their teacher and decreases in affiliation with a teacher across the elementary-middle school transition. School Belongingness

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


Students’ sense of school belonging declines during the middle school years (Anderman, 2003), and consequently, bullying increases from elementary to middle school (Pellegrini, 2002; Pellegrini & Long, 2002). Moreover, as bullying increases, school connectedness tends to decline, and students who feel disconnected from their school engage in more disruptive behavior, such as bullying (Young, 2004). Thus, we hypothesize that youth who report increases in bullying perpetration will report decreases in school sense of belonging as they transition to middle school. Present Study Using the ecological systems theory as a guiding framework, this study investigates how students might vary in their experiences in bullying over the elementary-to-middle school transition while examining a broad range of social-environmental factors. More specifically, we extend the extant research to include caregivers, teachers, and school environment factors associated with bullying during this transition. Although there are consistent mean-level findings that suggest bullying increases during the elementary-middle school transition, it is likely that examining social-environmental factors may illuminate different trajectories with associations to individual and social-environmental variables. We hypothesize that some students report increases and some report decreases in bullying. We also hypothesize that increases in bullying can contribute to increases in victimization and psychosocial issues, and decreases in adult support, social competence, and school sense of belonging. Method Participants Three hundred 5th grade students from three elementary schools in a Midwestern state participated in the study at Time 1 (T1; M=10.83 years of age, SD=.52). Of these, 158 (53%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


were males and 142 (47%) were females. Racial/ethnic composition included 132 (44%) European Americans, 39 (13%) African Americans, 61 (20%) Hispanic/Latino Americans, and 53 (18%) Asian Americans. Students receiving free/reduced lunch ranged from 25-40% across the three schools. Two hundred and thirty-seven out of 300 students from T1 participated in the study at Time 2 (T2; M=11.32 years, SD=.53). We were unable to locate 63 students because they had moved and/or transferred to another school district. The 237 students completed the transition from elementary school to middle school and were enrolled in the 6th grade. There were 116 (49%) males and 121 (51%) females with the following racial/ethnic distribution: 106 (44%) European Americans, 28 (12%) African Americans, 52 (22%) Hispanic/Latino Americans, and 42 (18%) Asian Americans. Six students identified themselves as biracial and three students as American Indian. Procedure Three elementary schools agreed to participate in the study in a district where all of the youth then attend the same middle school. The University Institutional Review Board and the school district approved a waiver of active parental consent procedure. We sent all parents of 5th grade students the information forms; they were asked to return the form only if they did not want their son/daughter to participate. Response rates ranged from 95 to 97% among the three schools. Students gave their assent by signing the survey coversheet. T1 data were collected as students were near the end of 5th grade and before they completed the middle school transition. T2 data were collected after students had been in middle school for three months, allowing them adequate time to adjust to the new classroom environment and school culture. At T1 and T2, participants completed the study survey during a

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


45-minute free period. Appropriate measures were taken to maintain confidentiality by ensuring that the participants were sitting far enough from one another. Data Analysis Plan First, descriptive analyses and correlational analyses were conducted across all study variables. Second, cluster analysis was used with bullying perpetration at Time 1 and at Time 2 as the input variables to determine groups of youth who reported no change, increases, or decreases in bullying perpetration across the elementary-middle school transition. Third, when a cluster solution was determined, these clusters were evaluated as a function of sex or race/ethnicity. Finally, cluster differences on students self-reports of their family, peer, teacher, and school experiences are presented. Measures Socio-Demographic Scale Participants completed a socio-demographic questionnaire that included questions about their sex, race/ethnicity, and age. Bullying and Victimization Scales Self-reported bullying. We used the nine-item University of Illinois Bully Scale (Espelage & Holt, 2001) to assess the frequency of teasing, name-calling, social exclusion, and rumor spreading at school. Participants were asked how often in the past 30 days they did the following to other students at school: teased other students, upset other students for the fun of it, excluded others from their group of friends, helped harass other students, or threatened to hit or hurt another student. Response options include: “Never”, “1 or 2 times”, “3 or 4 times”, “5 or 6 times”, and “7 or more times.” The construct validity of this scale has been supported via exploratory and confirmatory factor analysis (Espelage & Holt, 2001). The scale correlated

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


moderately with the Youth Self-Report Aggression Scale (r = .65; Achenbach, 1991), which suggests that it was somewhat unique from general aggression. Concurrent validity of this scale was established with significant correlations with peer nominations of bullying perpetration (Espelage et al., 2003). More specifically, students who reported the highest level of bullying perpetration on the scale received significantly more bullying nominations (M = 3.50, SD = 6.50) from their peers than students who did not self-report high levels of bullying perpetration (M = .98; SD = 1.10; Espelage et al., 2003). This scale is also distinct from pure aggression in factor analyses (Espelage, Low, et al., 2013). Acceptable estimates of the scale’s internal consistency were found for the current sample (α=.90 for T1, α=.85 for T2). Self-reported victimization. We used the four-item University of Illinois Victimization Scale (Espelage & Holt, 2001) to assess self-reported victimization over a 30-day period (e.g., “Other students called me names,” “Other students made fun of me,” and “I got hit and pushed by other students”). Response options include “Never,” “1 or 2 times,” “3 or 4 times,” “5 or 6 times,” and “7 or more times.” Higher scores indicate more self-reported victimization and scores converged with peer nominations. Acceptable estimates of the scale’s internal consistency were found for the current sample (α=.88 for T1, α=.87 for T2). Self-Perception Scales The Self-Perception Profile for Children (SPPC; Harter, 1985) provides a self-report selfassessment of 8- to 13-year-old students. The current investigation only included items corresponding to the global self-worth and social acceptance subscales, as these domains are shown to be important in the elementary-middle school transition (Robinson et al., 1995). A sample item from self-worth subscale is “Some kids are often unhappy with themselves BUT other kids are pretty pleased with themselves.” A sample item from the Social Acceptance

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


Subscale is “Some kids find it hard to make friends BUT other kids find it’s pretty easy to make friends.” For each item, the participant was first asked to decide whether he or she is more like the kids on the left side or more like the kids on the right side. Next, students were asked to indicate whether that is only “Sort of True” or “Really True” for him or her. Acceptable estimates of the scales’ internal consistency were found for the sample (Global Self-Work: α=.74 for T1, α=.75 for T2; α=.77 for T1, α=.78 for T2; Social Acceptance: α=.74 for T1, α=.75 for T2). Adult Support Scales We used the People in My Life scale (PIML; Cook et al., 1995) to assess participants’ relationships with adults. PIML is an adaptation of the Inventory of Parent and Peer Attachment (IPPA; Armsden & Greenberg, 1987), and was created to assess 10- to 12-year-old participants’ attachments to their parents and teachers. Relationship with caregivers. PIML (Cook et al., 1995) was used, which included three subscales (Trust, Communication, and Alienation). Trust subscale contains ten items (e.g., “I trust my parents”); Communication subscale consists of five items (e.g., “My parents listen to what I have to say”); and Alienation subscale includes five items (e.g., “I feel scared in my home”). Participants were asked to think about the main person who takes care of them and raises them (e.g., mother, father, grandparent, stepmother, stepfather), and finish the sentence, “The main person who takes care of and raises me is my _____.” Then, items from the subscales were reworded to be inclusive of other possible caregivers. Participants responded to items using a 4-point scale (1=almost never or never true; 4=almost always or always true). Acceptable estimates of internal consistency for the Trust (α=.84 for T1, α=.88 for T2), Communication

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


(α=.72 for T1, α=.70 for T2), and Alienation (α=.65 for T1, α=.70 for T2) subscales were found here. Relationship with teachers. We also used PIML to measure participants’ relationships with their teachers. Sample items from the two subscales are: “My teachers are proud of the things I do” (Affiliation with Teachers Subscale, 8 items) and “I get upset easily with my teachers” (Dissatisfaction with Teachers Subscale, 3 items). We asked participants to identify the elementary (or middle) school teacher to whom they feel the closest, and answer each item with that teacher in mind. Participants responded to items using a 4-point scale (1=almost never or never true; 4=almost always or always true). We found acceptable estimates of internal consistency for the sample for both Affiliation with Teachers subscale (α=.90 for T1, α=.89 for T2) and Dissatisfaction with Teachers subscale (α=.76 for T1, α=.79 for T2). School Belonging Scale The Psychological Sense of School Membership (PSSM; Goodenow, 1993) was used to assess students’ sense of belonging or psychological membership in their school. PSSM comprises 18 items (e.g., “Other students in this school take my opinions seriously”). Participants responded to items using a 5-point scale (1=not at all true to 5=completely true). Higher scores reflect a stronger sense of school belonging. Acceptable estimates of the scale’s internal consistency were found for the current sample (α=.84 for T1, α=.87 for T2). Internalized Symptoms Scales We used the Youth Self-Report (YSR, Achenbach, 1991) to assess participants’ problematic thoughts, behaviors, and feelings. Three of the eight subscales (i.e., Withdrawn, Somatic Complaints, and Anxious/Depressed Subscales) were included in this investigation because items on these subscales include internalizing behaviors of the YSR using factor analytic

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


data of 11- to 18-year-old samples (Achenbach, 1991). Further, these symptoms are found to be common among students during the elementary-middle school transition (Seidman et al., 1994). The Withdrawn Subscale consists of five items (e.g., “I would rather be alone than with others.”); the Somatic Complaints Subscale includes nine items (e.g., “I feel dizzy.”); and the Anxious/Depressed Subscale comprises 17 items (e.g., “I feel lonely.”). Participants were instructed, “For each item that describes you now or within the past 6 months, how true is each statement, ranging from 0 (Not true) to 2 (Very true). Higher scores reflect a greater number of problematic thoughts, behaviors, and/or feelings. Acceptable estimates of the scale’s internal consistency were found for Withdrawn (α=.64 for T1, α=.65 for T2), Somatic Complaints (α=.70 for T1, α=.75 for T2), and Anxious/Depressed (α=.86 for T1, α=.87 for T2). Results Means, standard deviations, and bivariate correlations among the variables within each time point are presented in Tables 1 and 2. The results of correlational analyses yielded a number of significant associations. Bullying (T1) was associated with lower feelings of school belonging, higher rates of negative teacher relationships, and greater internalized symptoms, including depression/anxiety, withdrawn behaviors, and somatic complaints. Further, students who reported higher levels of Bullying (T1) also reported less caregiver trust. Similar patterns emerged at T2, with higher levels of bullying associated with less caregiver support (e.g., less trust, more alienation), negative experiences at school (e.g., less teacher affiliation, more teacher dissatisfaction, less school sense of belonging), and higher rates of depression/anxiety. Interestingly, bullying was only associated with negative Global Self-Worth (T1). Finally, it is important to note that associations did not emerge between self-reported social acceptance and

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


bullying, which suggests that students who bully tend to believe that they are as accepted by their peers as students who are uninvolved in bullying. Cluster Analysis In order to assess different bully perpetration trajectories across the transition, we conducted k-means cluster analysis to create bully-change subtypes. Items from UIBS at both time points were included in these analyses. First, we used Ward’s (1963) algorithm to derive cluster solutions. This method minimizes the variance within clusters at each stage of grouping. Comparative studies have found that Ward’s method is one of the more effective cluster-analytic approaches (Borgen & Barnett, 1987). However, Ward’s method has also been criticized because of its tendency to produce results that are heavily influenced by profile elevation (Aldenderfer & Blashfield, 1984), and yield clusters with relatively equivalent numbers of observations (Hair & Black, 2000). To account for those factors, we also used the complete linkage method. This method combines cases with the smallest maximum distance at each stage of the agglomeration (Borgen & Barnett, 1987). Results of cluster analyses utilizing both of these methods suggested that a five-cluster solution was appropriate for the data. Following calculation of the cluster solution via these hierarchical methods, the data were then re-analyzed using k-means iterative partitioning, a nonhierarchical clustering method. Milligan (1980) suggested that “k-means clustering” is an appropriate follow-up analysis to hierarchical clustering techniques. Similarly, Kinder, Curtiss, and Kalichman (1991) recommend a two-stage approach, in which both hierarchical and nonhierarchical methods are used. Moving from a four-cluster to five-cluster solution yielded a substantial drop in SSwithin/Nclusters; thus, a five-cluster solution was selected.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65


Initial Cluster Descriptions Cluster 1 (C1), defined as the “Stable Low” cluster, included the majority of the participants (N=146; 61.6%: See Figure 1). These students had both the lowest mean bullying scale score at T1 (M=1.27, SD=0.23) and T2 (M=1.24, SD=0.28). Cluster 2 (C2), defined as “Decreasing,” included participants (N=44; 18.6%) whose bullying decreased over the transition, with moderate levels of Bullying (T1) (M=2.32, SD=0.53) and decreased levels of Bullying (T2) (M=1.61, SD=0.31). The third cluster (C3), labeled as “Moderately Increasing,” included 25 (10.5%) participants who had a moderate bullying scale score at T1 (M=2.35, SD=0.67) and a higher score at T2 (M=2.65, SD=0.26). A paired-sample t-test indicated that this increase was significant (p